From me at heyjay.com  Mon Jun  1 10:14:02 2009
From: me at heyjay.com (Jay Strauss)
Date: Mon, 1 Jun 2009 12:14:02 -0500
Subject: [Chicago-talk] New system architecture
Message-ID: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>

Hi all, need some advice.

I want to build a new system.  My bro and I have a small
equity/options trading application.  We cobbled it together in Perl
and Oracle years ago.  Things have broken, some of the stuff we want
to change.  Ultimately we want to start over.

I'm just trying to decide on what architecture to use going forward.
I know this is a pretty wide question.

Most (all) of the processes will run on a single box.  I don't really
have an idea of which presentation route we'll go (full rich client
maybe using Java/Swing, maybe a Java Script client using GWT, maybe
plain HTML).  The client will probably request functionality from the
server.

I was thinking we'd do the whole API for asking for quotes, requesting
analytics, inserting transactins...  via HTTP (SOAP).

Can anyone give me their thoughts on their lessons learned, maybe
routes to try, stuff I should be thinking of...

(again I know this is pretty open ended, but I'm interested in what
others have done with success)

Thanks
Jay

From michael at potter.name  Mon Jun  1 10:35:22 2009
From: michael at potter.name (Michael Potter)
Date: Mon, 1 Jun 2009 12:35:22 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
Message-ID: <2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>

Jay,

Here is my 2 cents:  If you are intending to replace Perl with Java then
consider Groovy instead of Java.  Groovy has a lot of the benefits of Perl
(Dynamic Language), but runs on the JVM and can interface with class files
that were created with javac.

Let us know what you decide.
-- 
Michael Potter

On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:

> Hi all, need some advice.
>
> I want to build a new system.  My bro and I have a small
> equity/options trading application.  We cobbled it together in Perl
> and Oracle years ago.  Things have broken, some of the stuff we want
> to change.  Ultimately we want to start over.
>
> I'm just trying to decide on what architecture to use going forward.
> I know this is a pretty wide question.
>
> Most (all) of the processes will run on a single box.  I don't really
> have an idea of which presentation route we'll go (full rich client
> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
> plain HTML).  The client will probably request functionality from the
> server.
>
> I was thinking we'd do the whole API for asking for quotes, requesting
> analytics, inserting transactins...  via HTTP (SOAP).
>
> Can anyone give me their thoughts on their lessons learned, maybe
> routes to try, stuff I should be thinking of...
>
> (again I know this is pretty open ended, but I'm interested in what
> others have done with success)
>
> Thanks
> Jay
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090601/183a327d/attachment.html>

From joel.limardo at forwardphase.com  Mon Jun  1 10:37:52 2009
From: joel.limardo at forwardphase.com (Joel Limardo)
Date: Mon, 1 Jun 2009 12:37:52 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
Message-ID: <001301c9e2df$b1dd1410$15973c30$@limardo@forwardphase.com>

If this is an academic exercise please disregard the remainder of this
e-mail.

I don't suggest you write much more code unless you cannot get the prototype
system that you currently have up and running. If you want to make money
(which I suspect that you do) then your next set of requirements ought to
come from the people who are going to pay you for use/access to your system.
I suggest that you spend some time documenting the system you have now; fix
its bugs; make some presentations and other materials and then take it out
for a test drive with real or strongly representative users. Since there are
many existing trading applications out there I suggest you focus on why
yours shows promise and/or is different. From this collect usage data and
complaints, classify them from High to Low (where "high" means "sure, I
would pay for this damn software if only it did X") and then ask the
question you just posed again armed with real metrics like the number of
users your application will support; the amount of data it will need; how
much data processing you expect it to perform; cost for operation; etc. 

There's an old Chinese proverb I go by which goes something like this --
when all doors are open, all doors are closed. I translate that to mean when
all requirements are open you are likely to start spinning your wheels.  I
go to actual users and ask them what they want, which typically eliminates
the problem entirely.

That's my two cents.  

-----Original Message-----
From: chicago-talk-bounces+joel.limardo=forwardphase.com at pm.org
[mailto:chicago-talk-bounces+joel.limardo=forwardphase.com at pm.org] On Behalf
Of Jay Strauss
Sent: Monday, June 01, 2009 12:14 PM
To: Chicago.pm chatter
Subject: [Chicago-talk] New system architecture

Hi all, need some advice.

I want to build a new system.  My bro and I have a small
equity/options trading application.  We cobbled it together in Perl
and Oracle years ago.  Things have broken, some of the stuff we want
to change.  Ultimately we want to start over.

I'm just trying to decide on what architecture to use going forward.
I know this is a pretty wide question.

Most (all) of the processes will run on a single box.  I don't really
have an idea of which presentation route we'll go (full rich client
maybe using Java/Swing, maybe a Java Script client using GWT, maybe
plain HTML).  The client will probably request functionality from the
server.

I was thinking we'd do the whole API for asking for quotes, requesting
analytics, inserting transactins...  via HTTP (SOAP).

Can anyone give me their thoughts on their lessons learned, maybe
routes to try, stuff I should be thinking of...

(again I know this is pretty open ended, but I'm interested in what
others have done with success)

Thanks
Jay
_______________________________________________
Chicago-talk mailing list
Chicago-talk at pm.org
http://mail.pm.org/mailman/listinfo/chicago-talk


From me at heyjay.com  Mon Jun  1 11:24:37 2009
From: me at heyjay.com (Jay Strauss)
Date: Mon, 1 Jun 2009 13:24:37 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <5236007079323364169@unknownmsgid>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<5236007079323364169@unknownmsgid>
Message-ID: <39eaccc10906011124h649c2bc5tb8fd6bea33fa7ce0@mail.gmail.com>

Well, the users are really just me and my brother, so at least we can
probably get access to them :)

Yep there are a lot of trading apps out there, but most seem to do
more stuff like technical indicators that you can build some sort of
if-then-else on a bunch of indicators to give a trade signal. We are
doing something a bit different, and have our money management
routines.

Most of our problems come from having switched brokerages (and thus
our data gathering is now different), switching some of our accounting
(making it simpler), and just not maintaining some of our stuff.

Why fix when we can spend twice as much time re-writing :)



On Mon, Jun 1, 2009 at 12:37 PM, Joel Limardo
<joel.limardo at forwardphase.com> wrote:
> If this is an academic exercise please disregard the remainder of this
> e-mail.
>
> I don't suggest you write much more code unless you cannot get the prototype
> system that you currently have up and running. If you want to make money
> (which I suspect that you do) then your next set of requirements ought to
> come from the people who are going to pay you for use/access to your system.
> I suggest that you spend some time documenting the system you have now; fix
> its bugs; make some presentations and other materials and then take it out
> for a test drive with real or strongly representative users. Since there are
> many existing trading applications out there I suggest you focus on why
> yours shows promise and/or is different. From this collect usage data and
> complaints, classify them from High to Low (where "high" means "sure, I
> would pay for this damn software if only it did X") and then ask the
> question you just posed again armed with real metrics like the number of
> users your application will support; the amount of data it will need; how
> much data processing you expect it to perform; cost for operation; etc.
>
> There's an old Chinese proverb I go by which goes something like this --
> when all doors are open, all doors are closed. I translate that to mean when
> all requirements are open you are likely to start spinning your wheels. ?I
> go to actual users and ask them what they want, which typically eliminates
> the problem entirely.
>
> That's my two cents.
>
> -----Original Message-----
> From: chicago-talk-bounces+joel.limardo=forwardphase.com at pm.org
> [mailto:chicago-talk-bounces+joel.limardo=forwardphase.com at pm.org] On Behalf
> Of Jay Strauss
> Sent: Monday, June 01, 2009 12:14 PM
> To: Chicago.pm chatter
> Subject: [Chicago-talk] New system architecture
>
> Hi all, need some advice.
>
> I want to build a new system. ?My bro and I have a small
> equity/options trading application. ?We cobbled it together in Perl
> and Oracle years ago. ?Things have broken, some of the stuff we want
> to change. ?Ultimately we want to start over.
>
> I'm just trying to decide on what architecture to use going forward.
> I know this is a pretty wide question.
>
> Most (all) of the processes will run on a single box. ?I don't really
> have an idea of which presentation route we'll go (full rich client
> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
> plain HTML). ?The client will probably request functionality from the
> server.
>
> I was thinking we'd do the whole API for asking for quotes, requesting
> analytics, inserting transactins... ?via HTTP (SOAP).
>
> Can anyone give me their thoughts on their lessons learned, maybe
> routes to try, stuff I should be thinking of...
>
> (again I know this is pretty open ended, but I'm interested in what
> others have done with success)
>
> Thanks
> Jay
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From merlyn at stonehenge.com  Mon Jun  1 11:25:07 2009
From: merlyn at stonehenge.com (Randal L. Schwartz)
Date: Mon, 01 Jun 2009 11:25:07 -0700
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	(Jay Strauss's message of "Mon, 1 Jun 2009 12:14:02 -0500")
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
Message-ID: <86y6sbx14s.fsf@blue.stonehenge.com>

>>>>> "Jay" == Jay Strauss <me at heyjay.com> writes:

Jay> I want to build a new system.  My bro and I have a small
Jay> equity/options trading application.  We cobbled it together in Perl
Jay> and Oracle years ago.  Things have broken, some of the stuff we want
Jay> to change.  Ultimately we want to start over.

Smalltalk/Seaside, using GemStone/S for the storage.  Once you learn
how to avoid ORMs, you'll never look back. :)

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn at stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

From me at heyjay.com  Mon Jun  1 11:25:37 2009
From: me at heyjay.com (Jay Strauss)
Date: Mon, 1 Jun 2009 13:25:37 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
Message-ID: <39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>

I don't think we've set upon Java or not.  But I don't feel there is a
good Perl rich GUI alternative (I know about wxperl, and am not in
love)

On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name> wrote:
> Jay,
>
> Here is my 2 cents:? If you are intending to replace Perl with Java then
> consider Groovy instead of Java.? Groovy has a lot of the benefits of Perl
> (Dynamic Language), but runs on the JVM and can interface with class files
> that were created with javac.
>
> Let us know what you decide.
> --
> Michael Potter
>
> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>
>> Hi all, need some advice.
>>
>> I want to build a new system. ?My bro and I have a small
>> equity/options trading application. ?We cobbled it together in Perl
>> and Oracle years ago. ?Things have broken, some of the stuff we want
>> to change. ?Ultimately we want to start over.
>>
>> I'm just trying to decide on what architecture to use going forward.
>> I know this is a pretty wide question.
>>
>> Most (all) of the processes will run on a single box. ?I don't really
>> have an idea of which presentation route we'll go (full rich client
>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>> plain HTML). ?The client will probably request functionality from the
>> server.
>>
>> I was thinking we'd do the whole API for asking for quotes, requesting
>> analytics, inserting transactins... ?via HTTP (SOAP).
>>
>> Can anyone give me their thoughts on their lessons learned, maybe
>> routes to try, stuff I should be thinking of...
>>
>> (again I know this is pretty open ended, but I'm interested in what
>> others have done with success)
>>
>> Thanks
>> Jay
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>
>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From jason at hostedlabs.com  Mon Jun  1 11:54:53 2009
From: jason at hostedlabs.com (Jason Rexilius)
Date: Mon, 01 Jun 2009 13:54:53 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
Message-ID: <4A2423FD.9030408@hostedlabs.com>

Sorry to chime in here late in the conversation.  I was an architect at 
JP morgan for FX and fixed income web-based trading apps and some 
desktop apps so have a bit of experience in related fields.


For fat client, I would suggest C++ if your desktop app needs good 
performance and Qt if your app needs good cross platform capability.

If you do not need that level of performance I would jump all the way to 
the other end of the spectrum and look at Flex/Flash for fat-client-like 
capability or really well engineered AJAX interactivity (I built 
web-based FX trading apps using java, perl and javascript/HTML back in 
the day that moved very, very fast).

Swing sux big time and you can get pretty close to its performance with 
well engineered flash client, particularly if the server component is 
local network.


For server side, java is one choice, I think perl is still a strong 
candidate depending on _how_ you code.  Python is also a strong 
candidate.  If you are really looking to rock it hard, the best might be 
Erlang. But really it comes down to what language you are comfortable 
with.  A good perl coder can build a faster app than a bad Java coder..

Me personally, unless you were doing heavy real-time trading based on 
market timing, I would build client in HTML/JS and server side with LAMP 
stack (perl, python or PHP) with shared mem or other high performance 
IPC mechanism to whatever your trading data feed provider is (reuters, 
bloomberg, etc.).  If you needed to do transforms or manips on data feed 
I'd right that in C/C++ or erlang.  If youre just talking to a DB, then 
the basic principles of a good handler that wraps all access to DBs and 
good schema design should really be enough. ORMs are often more baggage 
than help but maybe use a simple code generator to do light ORM class 
building or something..




Jay Strauss wrote:
> I don't think we've set upon Java or not.  But I don't feel there is a
> good Perl rich GUI alternative (I know about wxperl, and am not in
> love)
> 
> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name> wrote:
>> Jay,
>>
>> Here is my 2 cents:  If you are intending to replace Perl with Java then
>> consider Groovy instead of Java.  Groovy has a lot of the benefits of Perl
>> (Dynamic Language), but runs on the JVM and can interface with class files
>> that were created with javac.
>>
>> Let us know what you decide.
>> --
>> Michael Potter
>>
>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>> Hi all, need some advice.
>>>
>>> I want to build a new system.  My bro and I have a small
>>> equity/options trading application.  We cobbled it together in Perl
>>> and Oracle years ago.  Things have broken, some of the stuff we want
>>> to change.  Ultimately we want to start over.
>>>
>>> I'm just trying to decide on what architecture to use going forward.
>>> I know this is a pretty wide question.
>>>
>>> Most (all) of the processes will run on a single box.  I don't really
>>> have an idea of which presentation route we'll go (full rich client
>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>>> plain HTML).  The client will probably request functionality from the
>>> server.
>>>
>>> I was thinking we'd do the whole API for asking for quotes, requesting
>>> analytics, inserting transactins...  via HTTP (SOAP).
>>>
>>> Can anyone give me their thoughts on their lessons learned, maybe
>>> routes to try, stuff I should be thinking of...
>>>
>>> (again I know this is pretty open ended, but I'm interested in what
>>> others have done with success)
>>>
>>> Thanks
>>> Jay
>>> _______________________________________________
>>> Chicago-talk mailing list
>>> Chicago-talk at pm.org
>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk

From daveviner at pobox.com  Mon Jun  1 12:27:59 2009
From: daveviner at pobox.com (Dave Viner)
Date: Mon, 1 Jun 2009 12:27:59 -0700
Subject: [Chicago-talk] New system architecture
In-Reply-To: <4A2423FD.9030408@hostedlabs.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
Message-ID: <9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>

I would tend to agree with Jason.  Personally, I don't like most
desktop apps.  I like to use my browser.  So, AJAX is perfect.  Just
architect your web services to respond in JSON, and use jQuery to
handle the communication.  This setup allows you to grow - if you and
your brother hire other people, or want to license the software - and
it allows you to make changes to the appropriate component (front-end
versus service layer vs database).  (Note that many system would allow
for similar levels of encapsulation.  I'm just pointing out that a
well-designed AJAX system allows the same flexibility as a
well-designed desktop app or flash app or whatever else.)

Another alternative to consider is AIR.  You can make a pretty simple
desktop app using just HTML and JS.  Depending on your comfort with
using Adobe as the main execution platform, it's an interesting
approach that also buys you cross-platform compatibility.  Seesmic
Desktop is a full AIR app which has some pretty cool features.

HTH
Dave Viner


On Mon, Jun 1, 2009 at 11:54 AM, Jason Rexilius <jason at hostedlabs.com> wrote:
> Sorry to chime in here late in the conversation. ?I was an architect at JP
> morgan for FX and fixed income web-based trading apps and some desktop apps
> so have a bit of experience in related fields.
>
>
> For fat client, I would suggest C++ if your desktop app needs good
> performance and Qt if your app needs good cross platform capability.
>
> If you do not need that level of performance I would jump all the way to the
> other end of the spectrum and look at Flex/Flash for fat-client-like
> capability or really well engineered AJAX interactivity (I built web-based
> FX trading apps using java, perl and javascript/HTML back in the day that
> moved very, very fast).
>
> Swing sux big time and you can get pretty close to its performance with well
> engineered flash client, particularly if the server component is local
> network.
>
>
> For server side, java is one choice, I think perl is still a strong
> candidate depending on _how_ you code. ?Python is also a strong candidate.
> ?If you are really looking to rock it hard, the best might be Erlang. But
> really it comes down to what language you are comfortable with. ?A good perl
> coder can build a faster app than a bad Java coder..
>
> Me personally, unless you were doing heavy real-time trading based on market
> timing, I would build client in HTML/JS and server side with LAMP stack
> (perl, python or PHP) with shared mem or other high performance IPC
> mechanism to whatever your trading data feed provider is (reuters,
> bloomberg, etc.). ?If you needed to do transforms or manips on data feed I'd
> right that in C/C++ or erlang. ?If youre just talking to a DB, then the
> basic principles of a good handler that wraps all access to DBs and good
> schema design should really be enough. ORMs are often more baggage than help
> but maybe use a simple code generator to do light ORM class building or
> something..
>
>
>
>
> Jay Strauss wrote:
>>
>> I don't think we've set upon Java or not. ?But I don't feel there is a
>> good Perl rich GUI alternative (I know about wxperl, and am not in
>> love)
>>
>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
>> wrote:
>>>
>>> Jay,
>>>
>>> Here is my 2 cents: ?If you are intending to replace Perl with Java then
>>> consider Groovy instead of Java. ?Groovy has a lot of the benefits of
>>> Perl
>>> (Dynamic Language), but runs on the JVM and can interface with class
>>> files
>>> that were created with javac.
>>>
>>> Let us know what you decide.
>>> --
>>> Michael Potter
>>>
>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>>>
>>>> Hi all, need some advice.
>>>>
>>>> I want to build a new system. ?My bro and I have a small
>>>> equity/options trading application. ?We cobbled it together in Perl
>>>> and Oracle years ago. ?Things have broken, some of the stuff we want
>>>> to change. ?Ultimately we want to start over.
>>>>
>>>> I'm just trying to decide on what architecture to use going forward.
>>>> I know this is a pretty wide question.
>>>>
>>>> Most (all) of the processes will run on a single box. ?I don't really
>>>> have an idea of which presentation route we'll go (full rich client
>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>>>> plain HTML). ?The client will probably request functionality from the
>>>> server.
>>>>
>>>> I was thinking we'd do the whole API for asking for quotes, requesting
>>>> analytics, inserting transactins... ?via HTTP (SOAP).
>>>>
>>>> Can anyone give me their thoughts on their lessons learned, maybe
>>>> routes to try, stuff I should be thinking of...
>>>>
>>>> (again I know this is pretty open ended, but I'm interested in what
>>>> others have done with success)
>>>>
>>>> Thanks
>>>> Jay
>>>> _______________________________________________
>>>> Chicago-talk mailing list
>>>> Chicago-talk at pm.org
>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>>> _______________________________________________
>>> Chicago-talk mailing list
>>> Chicago-talk at pm.org
>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From me at heyjay.com  Tue Jun  2 07:28:38 2009
From: me at heyjay.com (Jay Strauss)
Date: Tue, 2 Jun 2009 09:28:38 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <4A2423FD.9030408@hostedlabs.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
Message-ID: <39eaccc10906020728i2c8e98c1yfc0547cd59327189@mail.gmail.com>

Hi, the client side of things isn't very complex.  My client is really
just display of our indicators and trade suggestions.  We are not
doing high frequency trades or trades that require lighting fast
execution.  HTML is probably fine.

I'm more concerned with the server side of things.  It seems like
there are some many different architectures these days I don't know
where to begin.  I could do everything in Perl, just having modules
talk to each other or I could go all the way to the other end like
some sort of spring/struts or SOAP kinda thing.

Furthermore, while I'm much more comfortable in Perl, It seems like
Java is the direction to go if I ever want to put my app on a web host
or some other app server.

I'm just sort of overwhelmed these days.


On Mon, Jun 1, 2009 at 1:54 PM, Jason Rexilius <jason at hostedlabs.com> wrote:
> Sorry to chime in here late in the conversation. ?I was an architect at JP
> morgan for FX and fixed income web-based trading apps and some desktop apps
> so have a bit of experience in related fields.
>
>
> For fat client, I would suggest C++ if your desktop app needs good
> performance and Qt if your app needs good cross platform capability.
>
> If you do not need that level of performance I would jump all the way to the
> other end of the spectrum and look at Flex/Flash for fat-client-like
> capability or really well engineered AJAX interactivity (I built web-based
> FX trading apps using java, perl and javascript/HTML back in the day that
> moved very, very fast).
>
> Swing sux big time and you can get pretty close to its performance with well
> engineered flash client, particularly if the server component is local
> network.
>
>
> For server side, java is one choice, I think perl is still a strong
> candidate depending on _how_ you code. ?Python is also a strong candidate.
> ?If you are really looking to rock it hard, the best might be Erlang. But
> really it comes down to what language you are comfortable with. ?A good perl
> coder can build a faster app than a bad Java coder..
>
> Me personally, unless you were doing heavy real-time trading based on market
> timing, I would build client in HTML/JS and server side with LAMP stack
> (perl, python or PHP) with shared mem or other high performance IPC
> mechanism to whatever your trading data feed provider is (reuters,
> bloomberg, etc.). ?If you needed to do transforms or manips on data feed I'd
> right that in C/C++ or erlang. ?If youre just talking to a DB, then the
> basic principles of a good handler that wraps all access to DBs and good
> schema design should really be enough. ORMs are often more baggage than help
> but maybe use a simple code generator to do light ORM class building or
> something..
>
>
>
>
> Jay Strauss wrote:
>>
>> I don't think we've set upon Java or not. ?But I don't feel there is a
>> good Perl rich GUI alternative (I know about wxperl, and am not in
>> love)
>>
>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
>> wrote:
>>>
>>> Jay,
>>>
>>> Here is my 2 cents: ?If you are intending to replace Perl with Java then
>>> consider Groovy instead of Java. ?Groovy has a lot of the benefits of
>>> Perl
>>> (Dynamic Language), but runs on the JVM and can interface with class
>>> files
>>> that were created with javac.
>>>
>>> Let us know what you decide.
>>> --
>>> Michael Potter
>>>
>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>>>
>>>> Hi all, need some advice.
>>>>
>>>> I want to build a new system. ?My bro and I have a small
>>>> equity/options trading application. ?We cobbled it together in Perl
>>>> and Oracle years ago. ?Things have broken, some of the stuff we want
>>>> to change. ?Ultimately we want to start over.
>>>>
>>>> I'm just trying to decide on what architecture to use going forward.
>>>> I know this is a pretty wide question.
>>>>
>>>> Most (all) of the processes will run on a single box. ?I don't really
>>>> have an idea of which presentation route we'll go (full rich client
>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>>>> plain HTML). ?The client will probably request functionality from the
>>>> server.
>>>>
>>>> I was thinking we'd do the whole API for asking for quotes, requesting
>>>> analytics, inserting transactins... ?via HTTP (SOAP).
>>>>
>>>> Can anyone give me their thoughts on their lessons learned, maybe
>>>> routes to try, stuff I should be thinking of...
>>>>
>>>> (again I know this is pretty open ended, but I'm interested in what
>>>> others have done with success)
>>>>
>>>> Thanks
>>>> Jay
>>>> _______________________________________________
>>>> Chicago-talk mailing list
>>>> Chicago-talk at pm.org
>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>>> _______________________________________________
>>> Chicago-talk mailing list
>>> Chicago-talk at pm.org
>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From me at heyjay.com  Tue Jun  2 07:47:00 2009
From: me at heyjay.com (Jay Strauss)
Date: Tue, 2 Jun 2009 09:47:00 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
Message-ID: <39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>

I like the idea of a java script (AJAX) client, although I don't know
javascript, which just means more learning curve.

You say use JSON.  I've never used it (reading about it now), do you
typically build really atomic pieces of code and communicate via JSON,
even between processes on the same host?

Where do the API of individual components end where you pass data in
native format, and the JSON begin?

Thanks
Jay

On Mon, Jun 1, 2009 at 2:27 PM, Dave Viner <daveviner at pobox.com> wrote:
> I would tend to agree with Jason. ?Personally, I don't like most
> desktop apps. ?I like to use my browser. ?So, AJAX is perfect. ?Just
> architect your web services to respond in JSON, and use jQuery to
> handle the communication. ?This setup allows you to grow - if you and
> your brother hire other people, or want to license the software - and
> it allows you to make changes to the appropriate component (front-end
> versus service layer vs database). ?(Note that many system would allow
> for similar levels of encapsulation. ?I'm just pointing out that a
> well-designed AJAX system allows the same flexibility as a
> well-designed desktop app or flash app or whatever else.)
>
> Another alternative to consider is AIR. ?You can make a pretty simple
> desktop app using just HTML and JS. ?Depending on your comfort with
> using Adobe as the main execution platform, it's an interesting
> approach that also buys you cross-platform compatibility. ?Seesmic
> Desktop is a full AIR app which has some pretty cool features.
>
> HTH
> Dave Viner
>
>
> On Mon, Jun 1, 2009 at 11:54 AM, Jason Rexilius <jason at hostedlabs.com> wrote:
>> Sorry to chime in here late in the conversation. ?I was an architect at JP
>> morgan for FX and fixed income web-based trading apps and some desktop apps
>> so have a bit of experience in related fields.
>>
>>
>> For fat client, I would suggest C++ if your desktop app needs good
>> performance and Qt if your app needs good cross platform capability.
>>
>> If you do not need that level of performance I would jump all the way to the
>> other end of the spectrum and look at Flex/Flash for fat-client-like
>> capability or really well engineered AJAX interactivity (I built web-based
>> FX trading apps using java, perl and javascript/HTML back in the day that
>> moved very, very fast).
>>
>> Swing sux big time and you can get pretty close to its performance with well
>> engineered flash client, particularly if the server component is local
>> network.
>>
>>
>> For server side, java is one choice, I think perl is still a strong
>> candidate depending on _how_ you code. ?Python is also a strong candidate.
>> ?If you are really looking to rock it hard, the best might be Erlang. But
>> really it comes down to what language you are comfortable with. ?A good perl
>> coder can build a faster app than a bad Java coder..
>>
>> Me personally, unless you were doing heavy real-time trading based on market
>> timing, I would build client in HTML/JS and server side with LAMP stack
>> (perl, python or PHP) with shared mem or other high performance IPC
>> mechanism to whatever your trading data feed provider is (reuters,
>> bloomberg, etc.). ?If you needed to do transforms or manips on data feed I'd
>> right that in C/C++ or erlang. ?If youre just talking to a DB, then the
>> basic principles of a good handler that wraps all access to DBs and good
>> schema design should really be enough. ORMs are often more baggage than help
>> but maybe use a simple code generator to do light ORM class building or
>> something..
>>
>>
>>
>>
>> Jay Strauss wrote:
>>>
>>> I don't think we've set upon Java or not. ?But I don't feel there is a
>>> good Perl rich GUI alternative (I know about wxperl, and am not in
>>> love)
>>>
>>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
>>> wrote:
>>>>
>>>> Jay,
>>>>
>>>> Here is my 2 cents: ?If you are intending to replace Perl with Java then
>>>> consider Groovy instead of Java. ?Groovy has a lot of the benefits of
>>>> Perl
>>>> (Dynamic Language), but runs on the JVM and can interface with class
>>>> files
>>>> that were created with javac.
>>>>
>>>> Let us know what you decide.
>>>> --
>>>> Michael Potter
>>>>
>>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>>>>
>>>>> Hi all, need some advice.
>>>>>
>>>>> I want to build a new system. ?My bro and I have a small
>>>>> equity/options trading application. ?We cobbled it together in Perl
>>>>> and Oracle years ago. ?Things have broken, some of the stuff we want
>>>>> to change. ?Ultimately we want to start over.
>>>>>
>>>>> I'm just trying to decide on what architecture to use going forward.
>>>>> I know this is a pretty wide question.
>>>>>
>>>>> Most (all) of the processes will run on a single box. ?I don't really
>>>>> have an idea of which presentation route we'll go (full rich client
>>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>>>>> plain HTML). ?The client will probably request functionality from the
>>>>> server.
>>>>>
>>>>> I was thinking we'd do the whole API for asking for quotes, requesting
>>>>> analytics, inserting transactins... ?via HTTP (SOAP).
>>>>>
>>>>> Can anyone give me their thoughts on their lessons learned, maybe
>>>>> routes to try, stuff I should be thinking of...
>>>>>
>>>>> (again I know this is pretty open ended, but I'm interested in what
>>>>> others have done with success)
>>>>>
>>>>> Thanks
>>>>> Jay
>>>>> _______________________________________________
>>>>> Chicago-talk mailing list
>>>>> Chicago-talk at pm.org
>>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>>
>>>> _______________________________________________
>>>> Chicago-talk mailing list
>>>> Chicago-talk at pm.org
>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>>
>>> _______________________________________________
>>> Chicago-talk mailing list
>>> Chicago-talk at pm.org
>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From daveviner at pobox.com  Tue Jun  2 09:03:50 2009
From: daveviner at pobox.com (Dave Viner)
Date: Tue, 2 Jun 2009 09:03:50 -0700
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
	<39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
Message-ID: <9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>

Hi Jay,

JSON is a transport mechanism.  Your web service returns some
information.  Traditionally, it would be something like XML.  JSON is
just easier when the code invoking the web service is Javascript.
That's all.

The more challenging part is figuring out what web services you really
need.  Usually, you start with either REST or SOAP as the primary
definition of your architecture.  Personally, I always start with REST
because it's more in tune with the general architecture of the web as
a whole.  (This is a bit of a philosophical debate, but you can read
more about that elsewhere.)  You can develop a REST-afarian system by
thinking through the "objects" or "resources" of your system.  To use
a language metaphor, in a REST architecture, the resources are the
nouns - the things on which you act.  The verbs are limited to GET,
PUT, POST, and DELETE.  GET returns the "representation of the
resource".  DELETE removes the resource.  PUT (usually) creates the
resource, and POST updates an existing resource.

You can get a decent grounding in REST architecture design by reading
just a few articles:
"How to create a REST Protocol" -
http://www.xml.com/pub/a/2004/12/01/restful-web.html - this is one of
the first solid walkthroughs of how to make your own protocol.  (Note
that for the purposes of this discussion, "protocol" and "web service"
are roughly synonymous.)  Read this 3 page article first.

"PUT or POST" -
http://jcalcote.wordpress.com/2008/10/16/put-or-post-the-rest-of-the-story/
- this article focuses on the distinction between PUT and POST.  Both
create or modify existing resources, so people frequently get confused
as to why and when to use each.


Those should give you a pretty solid understanding of how to get going
in the REST world.

Here are some thoughts on your specific questions:
"do you typically build really atomic pieces of code and communicate
via JSON, even between processes on the same host?"
- yes.  You define what your resources are.  Then you define what each
REST verb means for each resource.  The response from your server will
be in JSON.  Location of the host doesn't matter.  All communication
between components is HTTP.  This has the rather spectacular benefit
of allowing you to scale - when it's architected correctly.

"Where do the API of individual components end where you pass data in
native format, and the JSON begin?"
- This is hard to answer without knowing more of your process.  I'd
suggest starting with the definition of your resources, then the
verbs.  That should help you define what the services are, and what
actions you "take" or "effect" on them.  JSON is just the way to send
information in response to the service request.  You could substitute
XML for JSON and the resulting system would be identical.  Or you
could use plain text of specified formats (e.g., csv, tsv, or simple
new-line delimited keywords) and it's still the same.

HTH
Dave Viner



On Tue, Jun 2, 2009 at 7:47 AM, Jay Strauss <me at heyjay.com> wrote:
> I like the idea of a java script (AJAX) client, although I don't know
> javascript, which just means more learning curve.
>
> You say use JSON. ?I've never used it (reading about it now), do you
> typically build really atomic pieces of code and communicate via JSON,
> even between processes on the same host?
>
> Where do the API of individual components end where you pass data in
> native format, and the JSON begin?
>
> Thanks
> Jay
>
> On Mon, Jun 1, 2009 at 2:27 PM, Dave Viner <daveviner at pobox.com> wrote:
>> I would tend to agree with Jason. ?Personally, I don't like most
>> desktop apps. ?I like to use my browser. ?So, AJAX is perfect. ?Just
>> architect your web services to respond in JSON, and use jQuery to
>> handle the communication. ?This setup allows you to grow - if you and
>> your brother hire other people, or want to license the software - and
>> it allows you to make changes to the appropriate component (front-end
>> versus service layer vs database). ?(Note that many system would allow
>> for similar levels of encapsulation. ?I'm just pointing out that a
>> well-designed AJAX system allows the same flexibility as a
>> well-designed desktop app or flash app or whatever else.)
>>
>> Another alternative to consider is AIR. ?You can make a pretty simple
>> desktop app using just HTML and JS. ?Depending on your comfort with
>> using Adobe as the main execution platform, it's an interesting
>> approach that also buys you cross-platform compatibility. ?Seesmic
>> Desktop is a full AIR app which has some pretty cool features.
>>
>> HTH
>> Dave Viner
>>
>>
>> On Mon, Jun 1, 2009 at 11:54 AM, Jason Rexilius <jason at hostedlabs.com> wrote:
>>> Sorry to chime in here late in the conversation. ?I was an architect at JP
>>> morgan for FX and fixed income web-based trading apps and some desktop apps
>>> so have a bit of experience in related fields.
>>>
>>>
>>> For fat client, I would suggest C++ if your desktop app needs good
>>> performance and Qt if your app needs good cross platform capability.
>>>
>>> If you do not need that level of performance I would jump all the way to the
>>> other end of the spectrum and look at Flex/Flash for fat-client-like
>>> capability or really well engineered AJAX interactivity (I built web-based
>>> FX trading apps using java, perl and javascript/HTML back in the day that
>>> moved very, very fast).
>>>
>>> Swing sux big time and you can get pretty close to its performance with well
>>> engineered flash client, particularly if the server component is local
>>> network.
>>>
>>>
>>> For server side, java is one choice, I think perl is still a strong
>>> candidate depending on _how_ you code. ?Python is also a strong candidate.
>>> ?If you are really looking to rock it hard, the best might be Erlang. But
>>> really it comes down to what language you are comfortable with. ?A good perl
>>> coder can build a faster app than a bad Java coder..
>>>
>>> Me personally, unless you were doing heavy real-time trading based on market
>>> timing, I would build client in HTML/JS and server side with LAMP stack
>>> (perl, python or PHP) with shared mem or other high performance IPC
>>> mechanism to whatever your trading data feed provider is (reuters,
>>> bloomberg, etc.). ?If you needed to do transforms or manips on data feed I'd
>>> right that in C/C++ or erlang. ?If youre just talking to a DB, then the
>>> basic principles of a good handler that wraps all access to DBs and good
>>> schema design should really be enough. ORMs are often more baggage than help
>>> but maybe use a simple code generator to do light ORM class building or
>>> something..
>>>
>>>
>>>
>>>
>>> Jay Strauss wrote:
>>>>
>>>> I don't think we've set upon Java or not. ?But I don't feel there is a
>>>> good Perl rich GUI alternative (I know about wxperl, and am not in
>>>> love)
>>>>
>>>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
>>>> wrote:
>>>>>
>>>>> Jay,
>>>>>
>>>>> Here is my 2 cents: ?If you are intending to replace Perl with Java then
>>>>> consider Groovy instead of Java. ?Groovy has a lot of the benefits of
>>>>> Perl
>>>>> (Dynamic Language), but runs on the JVM and can interface with class
>>>>> files
>>>>> that were created with javac.
>>>>>
>>>>> Let us know what you decide.
>>>>> --
>>>>> Michael Potter
>>>>>
>>>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>>>>>>
>>>>>> Hi all, need some advice.
>>>>>>
>>>>>> I want to build a new system. ?My bro and I have a small
>>>>>> equity/options trading application. ?We cobbled it together in Perl
>>>>>> and Oracle years ago. ?Things have broken, some of the stuff we want
>>>>>> to change. ?Ultimately we want to start over.
>>>>>>
>>>>>> I'm just trying to decide on what architecture to use going forward.
>>>>>> I know this is a pretty wide question.
>>>>>>
>>>>>> Most (all) of the processes will run on a single box. ?I don't really
>>>>>> have an idea of which presentation route we'll go (full rich client
>>>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>>>>>> plain HTML). ?The client will probably request functionality from the
>>>>>> server.
>>>>>>
>>>>>> I was thinking we'd do the whole API for asking for quotes, requesting
>>>>>> analytics, inserting transactins... ?via HTTP (SOAP).
>>>>>>
>>>>>> Can anyone give me their thoughts on their lessons learned, maybe
>>>>>> routes to try, stuff I should be thinking of...
>>>>>>
>>>>>> (again I know this is pretty open ended, but I'm interested in what
>>>>>> others have done with success)
>>>>>>
>>>>>> Thanks
>>>>>> Jay
>>>>>> _______________________________________________
>>>>>> Chicago-talk mailing list
>>>>>> Chicago-talk at pm.org
>>>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>>>
>>>>> _______________________________________________
>>>>> Chicago-talk mailing list
>>>>> Chicago-talk at pm.org
>>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>>>
>>>> _______________________________________________
>>>> Chicago-talk mailing list
>>>> Chicago-talk at pm.org
>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>>> _______________________________________________
>>> Chicago-talk mailing list
>>> Chicago-talk at pm.org
>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>>
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From me at heyjay.com  Tue Jun  2 09:52:08 2009
From: me at heyjay.com (Jay Strauss)
Date: Tue, 2 Jun 2009 11:52:08 -0500
Subject: [Chicago-talk] New system architecture
In-Reply-To: <9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
	<39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
	<9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
Message-ID: <39eaccc10906020952i3bfeb8c3nff9e0fd3fc2b88cf@mail.gmail.com>

Hi, Dave,

Thanks for the time you put into the response.  I'm going to need to read
some more on this stuff.  Its seems there are so many different / competing
ideas and frameworks, its sort of impossible to know which way to go.
Struts / Springs / JSP / some templating engine...

I'll start by working up my exact list of tasks and maybe objects I expect.
Maybe things will become more obvious.

On Tue, Jun 2, 2009 at 11:03 AM, Dave Viner <daveviner at pobox.com> wrote:

> Hi Jay,
>
> JSON is a transport mechanism.  Your web service returns some
> information.  Traditionally, it would be something like XML.  JSON is
> just easier when the code invoking the web service is Javascript.
> That's all.
>
> The more challenging part is figuring out what web services you really
> need.  Usually, you start with either REST or SOAP as the primary
> definition of your architecture.  Personally, I always start with REST
> because it's more in tune with the general architecture of the web as
> a whole.  (This is a bit of a philosophical debate, but you can read
> more about that elsewhere.)  You can develop a REST-afarian system by
> thinking through the "objects" or "resources" of your system.  To use
> a language metaphor, in a REST architecture, the resources are the
> nouns - the things on which you act.  The verbs are limited to GET,
> PUT, POST, and DELETE.  GET returns the "representation of the
> resource".  DELETE removes the resource.  PUT (usually) creates the
> resource, and POST updates an existing resource.
>
> You can get a decent grounding in REST architecture design by reading
> just a few articles:
> "How to create a REST Protocol" -
> http://www.xml.com/pub/a/2004/12/01/restful-web.html - this is one of
> the first solid walkthroughs of how to make your own protocol.  (Note
> that for the purposes of this discussion, "protocol" and "web service"
> are roughly synonymous.)  Read this 3 page article first.
>
> "PUT or POST" -
> http://jcalcote.wordpress.com/2008/10/16/put-or-post-the-rest-of-the-story/
> - this article focuses on the distinction between PUT and POST.  Both
> create or modify existing resources, so people frequently get confused
> as to why and when to use each.
>
>
> Those should give you a pretty solid understanding of how to get going
> in the REST world.
>
> Here are some thoughts on your specific questions:
> "do you typically build really atomic pieces of code and communicate
> via JSON, even between processes on the same host?"
> - yes.  You define what your resources are.  Then you define what each
> REST verb means for each resource.  The response from your server will
> be in JSON.  Location of the host doesn't matter.  All communication
> between components is HTTP.  This has the rather spectacular benefit
> of allowing you to scale - when it's architected correctly.
>
> "Where do the API of individual components end where you pass data in
> native format, and the JSON begin?"
> - This is hard to answer without knowing more of your process.  I'd
> suggest starting with the definition of your resources, then the
> verbs.  That should help you define what the services are, and what
> actions you "take" or "effect" on them.  JSON is just the way to send
> information in response to the service request.  You could substitute
> XML for JSON and the resulting system would be identical.  Or you
> could use plain text of specified formats (e.g., csv, tsv, or simple
> new-line delimited keywords) and it's still the same.
>
> HTH
> Dave Viner
>
>
>
> On Tue, Jun 2, 2009 at 7:47 AM, Jay Strauss <me at heyjay.com> wrote:
> > I like the idea of a java script (AJAX) client, although I don't know
> > javascript, which just means more learning curve.
> >
> > You say use JSON.  I've never used it (reading about it now), do you
> > typically build really atomic pieces of code and communicate via JSON,
> > even between processes on the same host?
> >
> > Where do the API of individual components end where you pass data in
> > native format, and the JSON begin?
> >
> > Thanks
> > Jay
> >
> > On Mon, Jun 1, 2009 at 2:27 PM, Dave Viner <daveviner at pobox.com> wrote:
> >> I would tend to agree with Jason.  Personally, I don't like most
> >> desktop apps.  I like to use my browser.  So, AJAX is perfect.  Just
> >> architect your web services to respond in JSON, and use jQuery to
> >> handle the communication.  This setup allows you to grow - if you and
> >> your brother hire other people, or want to license the software - and
> >> it allows you to make changes to the appropriate component (front-end
> >> versus service layer vs database).  (Note that many system would allow
> >> for similar levels of encapsulation.  I'm just pointing out that a
> >> well-designed AJAX system allows the same flexibility as a
> >> well-designed desktop app or flash app or whatever else.)
> >>
> >> Another alternative to consider is AIR.  You can make a pretty simple
> >> desktop app using just HTML and JS.  Depending on your comfort with
> >> using Adobe as the main execution platform, it's an interesting
> >> approach that also buys you cross-platform compatibility.  Seesmic
> >> Desktop is a full AIR app which has some pretty cool features.
> >>
> >> HTH
> >> Dave Viner
> >>
> >>
> >> On Mon, Jun 1, 2009 at 11:54 AM, Jason Rexilius <jason at hostedlabs.com>
> wrote:
> >>> Sorry to chime in here late in the conversation.  I was an architect at
> JP
> >>> morgan for FX and fixed income web-based trading apps and some desktop
> apps
> >>> so have a bit of experience in related fields.
> >>>
> >>>
> >>> For fat client, I would suggest C++ if your desktop app needs good
> >>> performance and Qt if your app needs good cross platform capability.
> >>>
> >>> If you do not need that level of performance I would jump all the way
> to the
> >>> other end of the spectrum and look at Flex/Flash for fat-client-like
> >>> capability or really well engineered AJAX interactivity (I built
> web-based
> >>> FX trading apps using java, perl and javascript/HTML back in the day
> that
> >>> moved very, very fast).
> >>>
> >>> Swing sux big time and you can get pretty close to its performance with
> well
> >>> engineered flash client, particularly if the server component is local
> >>> network.
> >>>
> >>>
> >>> For server side, java is one choice, I think perl is still a strong
> >>> candidate depending on _how_ you code.  Python is also a strong
> candidate.
> >>>  If you are really looking to rock it hard, the best might be Erlang.
> But
> >>> really it comes down to what language you are comfortable with.  A good
> perl
> >>> coder can build a faster app than a bad Java coder..
> >>>
> >>> Me personally, unless you were doing heavy real-time trading based on
> market
> >>> timing, I would build client in HTML/JS and server side with LAMP stack
> >>> (perl, python or PHP) with shared mem or other high performance IPC
> >>> mechanism to whatever your trading data feed provider is (reuters,
> >>> bloomberg, etc.).  If you needed to do transforms or manips on data
> feed I'd
> >>> right that in C/C++ or erlang.  If youre just talking to a DB, then the
> >>> basic principles of a good handler that wraps all access to DBs and
> good
> >>> schema design should really be enough. ORMs are often more baggage than
> help
> >>> but maybe use a simple code generator to do light ORM class building or
> >>> something..
> >>>
> >>>
> >>>
> >>>
> >>> Jay Strauss wrote:
> >>>>
> >>>> I don't think we've set upon Java or not.  But I don't feel there is a
> >>>> good Perl rich GUI alternative (I know about wxperl, and am not in
> >>>> love)
> >>>>
> >>>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
> >>>> wrote:
> >>>>>
> >>>>> Jay,
> >>>>>
> >>>>> Here is my 2 cents:  If you are intending to replace Perl with Java
> then
> >>>>> consider Groovy instead of Java.  Groovy has a lot of the benefits of
> >>>>> Perl
> >>>>> (Dynamic Language), but runs on the JVM and can interface with class
> >>>>> files
> >>>>> that were created with javac.
> >>>>>
> >>>>> Let us know what you decide.
> >>>>> --
> >>>>> Michael Potter
> >>>>>
> >>>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
> >>>>>>
> >>>>>> Hi all, need some advice.
> >>>>>>
> >>>>>> I want to build a new system.  My bro and I have a small
> >>>>>> equity/options trading application.  We cobbled it together in Perl
> >>>>>> and Oracle years ago.  Things have broken, some of the stuff we want
> >>>>>> to change.  Ultimately we want to start over.
> >>>>>>
> >>>>>> I'm just trying to decide on what architecture to use going forward.
> >>>>>> I know this is a pretty wide question.
> >>>>>>
> >>>>>> Most (all) of the processes will run on a single box.  I don't
> really
> >>>>>> have an idea of which presentation route we'll go (full rich client
> >>>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
> >>>>>> plain HTML).  The client will probably request functionality from
> the
> >>>>>> server.
> >>>>>>
> >>>>>> I was thinking we'd do the whole API for asking for quotes,
> requesting
> >>>>>> analytics, inserting transactins...  via HTTP (SOAP).
> >>>>>>
> >>>>>> Can anyone give me their thoughts on their lessons learned, maybe
> >>>>>> routes to try, stuff I should be thinking of...
> >>>>>>
> >>>>>> (again I know this is pretty open ended, but I'm interested in what
> >>>>>> others have done with success)
> >>>>>>
> >>>>>> Thanks
> >>>>>> Jay
> >>>>>> _______________________________________________
> >>>>>> Chicago-talk mailing list
> >>>>>> Chicago-talk at pm.org
> >>>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
> >>>>>
> >>>>> _______________________________________________
> >>>>> Chicago-talk mailing list
> >>>>> Chicago-talk at pm.org
> >>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
> >>>>>
> >>>> _______________________________________________
> >>>> Chicago-talk mailing list
> >>>> Chicago-talk at pm.org
> >>>> http://mail.pm.org/mailman/listinfo/chicago-talk
> >>>
> >>> _______________________________________________
> >>> Chicago-talk mailing list
> >>> Chicago-talk at pm.org
> >>> http://mail.pm.org/mailman/listinfo/chicago-talk
> >>>
> >> _______________________________________________
> >> Chicago-talk mailing list
> >> Chicago-talk at pm.org
> >> http://mail.pm.org/mailman/listinfo/chicago-talk
> >>
> > _______________________________________________
> > Chicago-talk mailing list
> > Chicago-talk at pm.org
> > http://mail.pm.org/mailman/listinfo/chicago-talk
> >
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090602/79d95669/attachment.html>

From daveviner at pobox.com  Tue Jun  2 10:23:00 2009
From: daveviner at pobox.com (Dave Viner)
Date: Tue, 2 Jun 2009 10:23:00 -0700
Subject: [Chicago-talk] New system architecture
In-Reply-To: <39eaccc10906020952i3bfeb8c3nff9e0fd3fc2b88cf@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
	<39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
	<9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
	<39eaccc10906020952i3bfeb8c3nff9e0fd3fc2b88cf@mail.gmail.com>
Message-ID: <9a39b88d0906021023t7c332dd6g6234a39329586383@mail.gmail.com>

No problem.  If you need help, feel free to ping me off-list or
contact me via my consulting group http://www.vinertech.com.

Good luck!
Dave Viner

On Tue, Jun 2, 2009 at 9:52 AM, Jay Strauss <me at heyjay.com> wrote:
> Hi, Dave,
>
> Thanks for the time you put into the response.? I'm going to need to read
> some more on this stuff.? Its seems there are so many different / competing
> ideas and frameworks, its sort of impossible to know which way to go.
> Struts / Springs / JSP / some templating engine...
>
> I'll start by working up my exact list of tasks and maybe objects I expect.
> Maybe things will become more obvious.
>
> On Tue, Jun 2, 2009 at 11:03 AM, Dave Viner <daveviner at pobox.com> wrote:
>>
>> Hi Jay,
>>
>> JSON is a transport mechanism. ?Your web service returns some
>> information. ?Traditionally, it would be something like XML. ?JSON is
>> just easier when the code invoking the web service is Javascript.
>> That's all.
>>
>> The more challenging part is figuring out what web services you really
>> need. ?Usually, you start with either REST or SOAP as the primary
>> definition of your architecture. ?Personally, I always start with REST
>> because it's more in tune with the general architecture of the web as
>> a whole. ?(This is a bit of a philosophical debate, but you can read
>> more about that elsewhere.) ?You can develop a REST-afarian system by
>> thinking through the "objects" or "resources" of your system. ?To use
>> a language metaphor, in a REST architecture, the resources are the
>> nouns - the things on which you act. ?The verbs are limited to GET,
>> PUT, POST, and DELETE. ?GET returns the "representation of the
>> resource". ?DELETE removes the resource. ?PUT (usually) creates the
>> resource, and POST updates an existing resource.
>>
>> You can get a decent grounding in REST architecture design by reading
>> just a few articles:
>> "How to create a REST Protocol" -
>> http://www.xml.com/pub/a/2004/12/01/restful-web.html - this is one of
>> the first solid walkthroughs of how to make your own protocol. ?(Note
>> that for the purposes of this discussion, "protocol" and "web service"
>> are roughly synonymous.) ?Read this 3 page article first.
>>
>> "PUT or POST" -
>>
>> http://jcalcote.wordpress.com/2008/10/16/put-or-post-the-rest-of-the-story/
>> - this article focuses on the distinction between PUT and POST. ?Both
>> create or modify existing resources, so people frequently get confused
>> as to why and when to use each.
>>
>>
>> Those should give you a pretty solid understanding of how to get going
>> in the REST world.
>>
>> Here are some thoughts on your specific questions:
>> "do you typically build really atomic pieces of code and communicate
>> via JSON, even between processes on the same host?"
>> - yes. ?You define what your resources are. ?Then you define what each
>> REST verb means for each resource. ?The response from your server will
>> be in JSON. ?Location of the host doesn't matter. ?All communication
>> between components is HTTP. ?This has the rather spectacular benefit
>> of allowing you to scale - when it's architected correctly.
>>
>> "Where do the API of individual components end where you pass data in
>> native format, and the JSON begin?"
>> - This is hard to answer without knowing more of your process. ?I'd
>> suggest starting with the definition of your resources, then the
>> verbs. ?That should help you define what the services are, and what
>> actions you "take" or "effect" on them. ?JSON is just the way to send
>> information in response to the service request. ?You could substitute
>> XML for JSON and the resulting system would be identical. ?Or you
>> could use plain text of specified formats (e.g., csv, tsv, or simple
>> new-line delimited keywords) and it's still the same.
>>
>> HTH
>> Dave Viner
>>
>>
>>
>> On Tue, Jun 2, 2009 at 7:47 AM, Jay Strauss <me at heyjay.com> wrote:
>> > I like the idea of a java script (AJAX) client, although I don't know
>> > javascript, which just means more learning curve.
>> >
>> > You say use JSON. ?I've never used it (reading about it now), do you
>> > typically build really atomic pieces of code and communicate via JSON,
>> > even between processes on the same host?
>> >
>> > Where do the API of individual components end where you pass data in
>> > native format, and the JSON begin?
>> >
>> > Thanks
>> > Jay
>> >
>> > On Mon, Jun 1, 2009 at 2:27 PM, Dave Viner <daveviner at pobox.com> wrote:
>> >> I would tend to agree with Jason. ?Personally, I don't like most
>> >> desktop apps. ?I like to use my browser. ?So, AJAX is perfect. ?Just
>> >> architect your web services to respond in JSON, and use jQuery to
>> >> handle the communication. ?This setup allows you to grow - if you and
>> >> your brother hire other people, or want to license the software - and
>> >> it allows you to make changes to the appropriate component (front-end
>> >> versus service layer vs database). ?(Note that many system would allow
>> >> for similar levels of encapsulation. ?I'm just pointing out that a
>> >> well-designed AJAX system allows the same flexibility as a
>> >> well-designed desktop app or flash app or whatever else.)
>> >>
>> >> Another alternative to consider is AIR. ?You can make a pretty simple
>> >> desktop app using just HTML and JS. ?Depending on your comfort with
>> >> using Adobe as the main execution platform, it's an interesting
>> >> approach that also buys you cross-platform compatibility. ?Seesmic
>> >> Desktop is a full AIR app which has some pretty cool features.
>> >>
>> >> HTH
>> >> Dave Viner
>> >>
>> >>
>> >> On Mon, Jun 1, 2009 at 11:54 AM, Jason Rexilius <jason at hostedlabs.com>
>> >> wrote:
>> >>> Sorry to chime in here late in the conversation. ?I was an architect
>> >>> at JP
>> >>> morgan for FX and fixed income web-based trading apps and some desktop
>> >>> apps
>> >>> so have a bit of experience in related fields.
>> >>>
>> >>>
>> >>> For fat client, I would suggest C++ if your desktop app needs good
>> >>> performance and Qt if your app needs good cross platform capability.
>> >>>
>> >>> If you do not need that level of performance I would jump all the way
>> >>> to the
>> >>> other end of the spectrum and look at Flex/Flash for fat-client-like
>> >>> capability or really well engineered AJAX interactivity (I built
>> >>> web-based
>> >>> FX trading apps using java, perl and javascript/HTML back in the day
>> >>> that
>> >>> moved very, very fast).
>> >>>
>> >>> Swing sux big time and you can get pretty close to its performance
>> >>> with well
>> >>> engineered flash client, particularly if the server component is local
>> >>> network.
>> >>>
>> >>>
>> >>> For server side, java is one choice, I think perl is still a strong
>> >>> candidate depending on _how_ you code. ?Python is also a strong
>> >>> candidate.
>> >>> ?If you are really looking to rock it hard, the best might be Erlang.
>> >>> But
>> >>> really it comes down to what language you are comfortable with. ?A
>> >>> good perl
>> >>> coder can build a faster app than a bad Java coder..
>> >>>
>> >>> Me personally, unless you were doing heavy real-time trading based on
>> >>> market
>> >>> timing, I would build client in HTML/JS and server side with LAMP
>> >>> stack
>> >>> (perl, python or PHP) with shared mem or other high performance IPC
>> >>> mechanism to whatever your trading data feed provider is (reuters,
>> >>> bloomberg, etc.). ?If you needed to do transforms or manips on data
>> >>> feed I'd
>> >>> right that in C/C++ or erlang. ?If youre just talking to a DB, then
>> >>> the
>> >>> basic principles of a good handler that wraps all access to DBs and
>> >>> good
>> >>> schema design should really be enough. ORMs are often more baggage
>> >>> than help
>> >>> but maybe use a simple code generator to do light ORM class building
>> >>> or
>> >>> something..
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> Jay Strauss wrote:
>> >>>>
>> >>>> I don't think we've set upon Java or not. ?But I don't feel there is
>> >>>> a
>> >>>> good Perl rich GUI alternative (I know about wxperl, and am not in
>> >>>> love)
>> >>>>
>> >>>> On Mon, Jun 1, 2009 at 12:35 PM, Michael Potter <michael at potter.name>
>> >>>> wrote:
>> >>>>>
>> >>>>> Jay,
>> >>>>>
>> >>>>> Here is my 2 cents: ?If you are intending to replace Perl with Java
>> >>>>> then
>> >>>>> consider Groovy instead of Java. ?Groovy has a lot of the benefits
>> >>>>> of
>> >>>>> Perl
>> >>>>> (Dynamic Language), but runs on the JVM and can interface with class
>> >>>>> files
>> >>>>> that were created with javac.
>> >>>>>
>> >>>>> Let us know what you decide.
>> >>>>> --
>> >>>>> Michael Potter
>> >>>>>
>> >>>>> On Mon, Jun 1, 2009 at 12:14 PM, Jay Strauss <me at heyjay.com> wrote:
>> >>>>>>
>> >>>>>> Hi all, need some advice.
>> >>>>>>
>> >>>>>> I want to build a new system. ?My bro and I have a small
>> >>>>>> equity/options trading application. ?We cobbled it together in Perl
>> >>>>>> and Oracle years ago. ?Things have broken, some of the stuff we
>> >>>>>> want
>> >>>>>> to change. ?Ultimately we want to start over.
>> >>>>>>
>> >>>>>> I'm just trying to decide on what architecture to use going
>> >>>>>> forward.
>> >>>>>> I know this is a pretty wide question.
>> >>>>>>
>> >>>>>> Most (all) of the processes will run on a single box. ?I don't
>> >>>>>> really
>> >>>>>> have an idea of which presentation route we'll go (full rich client
>> >>>>>> maybe using Java/Swing, maybe a Java Script client using GWT, maybe
>> >>>>>> plain HTML). ?The client will probably request functionality from
>> >>>>>> the
>> >>>>>> server.
>> >>>>>>
>> >>>>>> I was thinking we'd do the whole API for asking for quotes,
>> >>>>>> requesting
>> >>>>>> analytics, inserting transactins... ?via HTTP (SOAP).
>> >>>>>>
>> >>>>>> Can anyone give me their thoughts on their lessons learned, maybe
>> >>>>>> routes to try, stuff I should be thinking of...
>> >>>>>>
>> >>>>>> (again I know this is pretty open ended, but I'm interested in what
>> >>>>>> others have done with success)
>> >>>>>>
>> >>>>>> Thanks
>> >>>>>> Jay
>> >>>>>> _______________________________________________
>> >>>>>> Chicago-talk mailing list
>> >>>>>> Chicago-talk at pm.org
>> >>>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>> >>>>>
>> >>>>> _______________________________________________
>> >>>>> Chicago-talk mailing list
>> >>>>> Chicago-talk at pm.org
>> >>>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>> >>>>>
>> >>>> _______________________________________________
>> >>>> Chicago-talk mailing list
>> >>>> Chicago-talk at pm.org
>> >>>> http://mail.pm.org/mailman/listinfo/chicago-talk
>> >>>
>> >>> _______________________________________________
>> >>> Chicago-talk mailing list
>> >>> Chicago-talk at pm.org
>> >>> http://mail.pm.org/mailman/listinfo/chicago-talk
>> >>>
>> >> _______________________________________________
>> >> Chicago-talk mailing list
>> >> Chicago-talk at pm.org
>> >> http://mail.pm.org/mailman/listinfo/chicago-talk
>> >>
>> > _______________________________________________
>> > Chicago-talk mailing list
>> > Chicago-talk at pm.org
>> > http://mail.pm.org/mailman/listinfo/chicago-talk
>> >
>> _______________________________________________
>> Chicago-talk mailing list
>> Chicago-talk at pm.org
>> http://mail.pm.org/mailman/listinfo/chicago-talk
>
>
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

From fire at dls.net  Thu Jun 18 05:03:57 2009
From: fire at dls.net (Bradley Slavik)
Date: Thu, 18 Jun 2009 07:03:57 -0500
Subject: [Chicago-talk] XML DTD validation
In-Reply-To: <9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
	<39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
	<9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
Message-ID: <DE500826-7C2F-4ADE-9193-81F0F4D0D64D@dls.net>


Dear Chicago PerlMongers,

I have been working with XML::Checker::Parser (part of libxml-enno)
and found that many of the tests are broken, and it probably could
use to be upgraded to using Test::More. I have fixed most of it, but
maybe I should be expending my efforts on a different perl package?

I know that Andy has worked extensively with XML so maybe he has
some opinion on how to validate XML DTDs, and whether it is worth
any effort at all. (I could be wasting my time)

Also I don't want to offend the author, and I have never really done
clean-up on a CPAN package before. I was hoping that someone
could give me pointers on how to do that correctly too.

Also, I just sat down and learned the perl Test packages today, so
I will probably have some questions about what is the preferred mode
to use these. It seems that this coder rolled his own, and in most cases
I can replace the test calls with Test::More ones, but I did have a  
concern
about lines 223 and 249 of t/chk_batch.t. It has this curious line:

     local *XML::DOM::warning = \&append_str;

which produces this error during testing:

Subroutine XML::DOM::warning redefined at t/chk_batch.t line 223.
Subroutine XML::DOM::warning redefined at t/chk_batch.t line 249.

It is pretty lame, and there is probably a better way to do it, but  
if it
is better to override the warning routine? Just let it get called?

Here is the subroutine (so you don't have to look it up)

sub append_str
{
     $error_str .= shift() . "\n";
}

Probably there is an accepted idiom to use for this purpose when
writing tests, and I would rather go with the flow than create my own
wheel.

Thanks for your assistance.


Bradley Slavik
fire at dls.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090618/ce0a6476/attachment.html>

From fire at dls.net  Thu Jun 18 05:04:12 2009
From: fire at dls.net (Bradley Slavik)
Date: Thu, 18 Jun 2009 07:04:12 -0500
Subject: [Chicago-talk] XML DTD validation
In-Reply-To: <9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
References: <39eaccc10906011014n138593b1ne81deab3a53f5e07@mail.gmail.com>
	<2379dacc0906011035y62853dfexe054481894a5c83b@mail.gmail.com>
	<39eaccc10906011125g2522102clddf5729c1c8821b3@mail.gmail.com>
	<4A2423FD.9030408@hostedlabs.com>
	<9a39b88d0906011227v4db8aac6rb43881c96edf5819@mail.gmail.com>
	<39eaccc10906020747g3fa63247gaf4e47f0f79b9112@mail.gmail.com>
	<9a39b88d0906020903p57cfb23dof896878fda50edd@mail.gmail.com>
Message-ID: <A45762FD-9186-485F-91F4-D3468A185572@dls.net>


Dear Chicago PerlMongers,

I have been working with XML::Checker::Parser (part of libxml-enno)
and found that many of the tests are broken, and it probably could
use to be upgraded to using Test::More. I have fixed most of it, but
maybe I should be expending my efforts on a different perl package?

I know that Andy has worked extensively with XML so maybe he has
some opinion on how to validate XML DTDs, and whether it is worth
any effort at all. (I could be wasting my time)

Also I don't want to offend the author, and I have never really done
clean-up on a CPAN package before. I was hoping that someone
could give me pointers on how to do that correctly too.

Also, I just sat down and learned the perl Test packages today, so
I will probably have some questions about what is the preferred mode
to use these. It seems that this coder rolled his own, and in most cases
I can replace the test calls with Test::More ones, but I did have a  
concern
about lines 223 and 249 of t/chk_batch.t. It has this curious line:

     local *XML::DOM::warning = \&append_str;

which produces this error during testing:

Subroutine XML::DOM::warning redefined at t/chk_batch.t line 223.
Subroutine XML::DOM::warning redefined at t/chk_batch.t line 249.

It is pretty lame, and there is probably a better way to do it, but  
if it
is better to override the warning routine? Just let it get called?

Here is the subroutine (so you don't have to look it up)

sub append_str
{
     $error_str .= shift() . "\n";
}

Probably there is an accepted idiom to use for this purpose when
writing tests, and I would rather go with the flow than create my own
wheel.

Thanks for your assistance.


Bradley Slavik
fire at dls.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090618/95f67049/attachment-0001.html>

From fire at dls.net  Fri Jun 19 09:15:58 2009
From: fire at dls.net (Bradley Slavik)
Date: Fri, 19 Jun 2009 11:15:58 -0500
Subject: [Chicago-talk] XML DTD validation
Message-ID: <41436.1245428158@dls.net>

Dear Perl XML list,

I have been working with XML::Checker::Parser (part of libxml-enno)
and found that many of the tests are broken, and it probably could
use to be upgraded to using Test::More. I have fixed most of it, but
maybe I should be expending my efforts on a different perl package?

Is there a better package in Perl to validate XML files against DTDs or Schemas?
I currently only have DTDs for these XML files.

It was interesting to see this project as well: http://vtd-xml.sourceforge.net/
but I did not see how to use it for validating DTDs or Schemas.

This project also looked promising: http://www.textuality.com/Lark/

And there is always the faithful libxml and Xerces to fall back on.

I just want to get a little feedback, so if I do this, it is done
right and benefits the perl/xml community as a whole, not just me.

It would be nice if someone who has submitted CPAN modules previously
could look over the code before I submit it. I feel confident there are some
nice perl idioms for packages that might make the package more standard.

I just learned the perl Test packages today. I think if the libxml-enno
is still in favor then I should clean up the test suite and go over the
code. Here is an example of a simple question that I just don't know the
answer to:

Lines 223 and 249 of t/chk_batch.t:

    local *XML::DOM::warning = \&append_str;

which produces this error during testing:

Subroutine XML::DOM::warning redefined at t/chk_batch.t line 223.
Subroutine XML::DOM::warning redefined at t/chk_batch.t line 249.

It is pretty lame, and there is probably a better way to do it, but if it
is better to overwrite it, how to avoid the warning? I now know that I
can use no warnings to get around this, but is this the BEST perl practice?

Here is the subroutine (so you don't have to look it up)

sub append_str
{
    $error_str .= shift() . "\n";
}

Probably there is an accepted idiom to use for this purpose when
writing tests, and I would rather go with the flow than create my own
wheel. I am not trying to get someone else to do the work here, just
hoping to get a modicum of support.

Thanks for your assistance.


Bradley Slavik
fire at dls.net



From darren.young at chicagobooth.edu  Tue Jun 23 13:12:48 2009
From: darren.young at chicagobooth.edu (Young, Darren)
Date: Tue, 23 Jun 2009 15:12:48 -0500
Subject: [Chicago-talk] Array::Diff
Message-ID: <07A371D457B501478C1DB3C3DE8372D703FE48C8@GSBHEX2V.gsb.uchicago.edu>

Anyone used Array::Diff before? I have the following but am getting back
ref's for the arrays:

        # determine additions and removals from the list
        @curmembers = sort(@curmembers);
        @newmembers = sort(@newmembers);
        my $diff = Array::Diff->diff( \@curmembers, \@newmembers );
        my @addmembers = $diff->added;
        my @delmembers = $diff->deleted;

        logmsg("Adds: " . scalar(@addmembers));
        logmsg("Dels: " . scalar(@delmembers));

        foreach my $add (@addmembers) {
            logmsg("ADD MEMBER: $add");
        }       
        foreach my $del (@delmembers) {
            logmsg("DEL MEMBER: $del");
        }

Results:
[2009-06-23 15:05:05] main(): Adds: 1
[2009-06-23 15:05:05] main(): Dels: 1
[2009-06-23 15:05:05] main(): ADD MEMBER: ARRAY(0x90db898)
[2009-06-23 15:05:05] main(): DEL MEMBER: ARRAY(0x9107658)

>From the docs I was assuming the return was an array of diff's.

From lee at laylward.com  Tue Jun 23 14:47:10 2009
From: lee at laylward.com (Lee Aylward)
Date: Tue, 23 Jun 2009 17:47:10 -0400
Subject: [Chicago-talk] Array::Diff
In-Reply-To: <07A371D457B501478C1DB3C3DE8372D703FE48C8@GSBHEX2V.gsb.uchicago.edu>
References: <07A371D457B501478C1DB3C3DE8372D703FE48C8@GSBHEX2V.gsb.uchicago.edu>
Message-ID: <06AFEA96-DAA3-4ECD-85D2-3E28DAB521ED@laylward.com>

I just checked the docs on search.cpan.org and it looks like ->added()  
and ->deleted() both return array refs.

e.g. (from the synposis)
$diff->added # [ 'd' ];
$diff->deleted # [ 'a' ];
-- 
Lee

On Jun 23, 2009, at 4:12 PM, Young, Darren wrote:

> Anyone used Array::Diff before? I have the following but am getting  
> back
> ref's for the arrays:
>
>        # determine additions and removals from the list
>        @curmembers = sort(@curmembers);
>        @newmembers = sort(@newmembers);
>        my $diff = Array::Diff->diff( \@curmembers, \@newmembers );
>        my @addmembers = $diff->added;
>        my @delmembers = $diff->deleted;
>
>        logmsg("Adds: " . scalar(@addmembers));
>        logmsg("Dels: " . scalar(@delmembers));
>
>        foreach my $add (@addmembers) {
>            logmsg("ADD MEMBER: $add");
>        }
>        foreach my $del (@delmembers) {
>            logmsg("DEL MEMBER: $del");
>        }
>
> Results:
> [2009-06-23 15:05:05] main(): Adds: 1
> [2009-06-23 15:05:05] main(): Dels: 1
> [2009-06-23 15:05:05] main(): ADD MEMBER: ARRAY(0x90db898)
> [2009-06-23 15:05:05] main(): DEL MEMBER: ARRAY(0x9107658)
>
>> From the docs I was assuming the return was an array of diff's.
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>


From Darren.Young at chicagobooth.edu  Tue Jun 23 15:08:53 2009
From: Darren.Young at chicagobooth.edu (Young, Darren)
Date: Tue, 23 Jun 2009 17:08:53 -0500
Subject: [Chicago-talk] Array::Diff
In-Reply-To: <06AFEA96-DAA3-4ECD-85D2-3E28DAB521ED@laylward.com>
References: <07A371D457B501478C1DB3C3DE8372D703FE48C8@GSBHEX2V.gsb.uchicago.edu>
	<06AFEA96-DAA3-4ECD-85D2-3E28DAB521ED@laylward.com>
Message-ID: <07A371D457B501478C1DB3C3DE8372D703FE49E7@GSBHEX2V.gsb.uchicago.edu>

Yea, description doesn't really say that.

> -----Original Message-----
> From: chicago-talk-bounces+darren.young=chicagobooth.edu at pm.org
> [mailto:chicago-talk-bounces+darren.young=chicagobooth.edu at pm.org] On
> Behalf Of Lee Aylward
> Sent: Tuesday, June 23, 2009 4:47 PM
> To: Chicago.pm chatter
> Subject: Re: [Chicago-talk] Array::Diff
> 
> I just checked the docs on search.cpan.org and it looks like ->added()
> and ->deleted() both return array refs.
> 
> e.g. (from the synposis)
> $diff->added # [ 'd' ];
> $diff->deleted # [ 'a' ];
> --
> Lee
> 
> On Jun 23, 2009, at 4:12 PM, Young, Darren wrote:
> 
> > Anyone used Array::Diff before? I have the following but am getting
> > back
> > ref's for the arrays:
> >
> >        # determine additions and removals from the list
> >        @curmembers = sort(@curmembers);
> >        @newmembers = sort(@newmembers);
> >        my $diff = Array::Diff->diff( \@curmembers, \@newmembers );
> >        my @addmembers = $diff->added;
> >        my @delmembers = $diff->deleted;
> >
> >        logmsg("Adds: " . scalar(@addmembers));
> >        logmsg("Dels: " . scalar(@delmembers));
> >
> >        foreach my $add (@addmembers) {
> >            logmsg("ADD MEMBER: $add");
> >        }
> >        foreach my $del (@delmembers) {
> >            logmsg("DEL MEMBER: $del");
> >        }
> >
> > Results:
> > [2009-06-23 15:05:05] main(): Adds: 1
> > [2009-06-23 15:05:05] main(): Dels: 1
> > [2009-06-23 15:05:05] main(): ADD MEMBER: ARRAY(0x90db898)
> > [2009-06-23 15:05:05] main(): DEL MEMBER: ARRAY(0x9107658)
> >
> >> From the docs I was assuming the return was an array of diff's.
> > _______________________________________________
> > Chicago-talk mailing list
> > Chicago-talk at pm.org
> > http://mail.pm.org/mailman/listinfo/chicago-talk
> >
> 
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk

From joshua.mcadams at gmail.com  Wed Jun 24 08:35:58 2009
From: joshua.mcadams at gmail.com (Joshua)
Date: Wed, 24 Jun 2009 11:35:58 -0400
Subject: [Chicago-talk] YAPC Houston t-shirts
Message-ID: <49d805d70906240835v121fd4ceqb91fcdffce4cd1c6@mail.gmail.com>

There is a huge pile of YAPC::Houston t-shirts here at
YAPC::Pittsburgh.  If anyone wants one, please let me know your size
and I'll do my best to bring some shirts back... no guarantees though
:)

From hwigoda at mindspring.com  Wed Jun 24 08:49:37 2009
From: hwigoda at mindspring.com (Hal Wigoda)
Date: Wed, 24 Jun 2009 10:49:37 -0500
Subject: [Chicago-talk] YAPC Houston t-shirts
In-Reply-To: <49d805d70906240835v121fd4ceqb91fcdffce4cd1c6@mail.gmail.com>
References: <49d805d70906240835v121fd4ceqb91fcdffce4cd1c6@mail.gmail.com>
Message-ID: <549B99C0-145F-4C75-91DF-9ADCA3F86315@mindspring.com>

The biggest you have 4XL of possible.
Hal

Sent from my iPhone

On Jun 24, 2009, at 10:35 AM, Joshua <joshua.mcadams at gmail.com> wrote:

> There is a huge pile of YAPC::Houston t-shirts here at
> YAPC::Pittsburgh.  If anyone wants one, please let me know your size
> and I'll do my best to bring some shirts back... no guarantees though
> :)
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk

From merlyn at stonehenge.com  Wed Jun 24 09:19:59 2009
From: merlyn at stonehenge.com (Randal L. Schwartz)
Date: Wed, 24 Jun 2009 09:19:59 -0700
Subject: [Chicago-talk] YAPC Houston t-shirts
In-Reply-To: <549B99C0-145F-4C75-91DF-9ADCA3F86315@mindspring.com> (Hal
	Wigoda's message of "Wed, 24 Jun 2009 10:49:37 -0500")
References: <49d805d70906240835v121fd4ceqb91fcdffce4cd1c6@mail.gmail.com>
	<549B99C0-145F-4C75-91DF-9ADCA3F86315@mindspring.com>
Message-ID: <86skhpzjrk.fsf@blue.stonehenge.com>

>>>>> "Hal" == Hal Wigoda <hwigoda at mindspring.com> writes:

Hal> The biggest you have 4XL of possible.

Aka - "Programmer Size". :)

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn at stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

From kent at c2group.net  Wed Jun 24 14:50:03 2009
From: kent at c2group.net (Kent Cowgill)
Date: Wed, 24 Jun 2009 17:50:03 -0400
Subject: [Chicago-talk] YAPC Houston t-shirts
In-Reply-To: <86skhpzjrk.fsf@blue.stonehenge.com>
References: <49d805d70906240835v121fd4ceqb91fcdffce4cd1c6@mail.gmail.com>
	<549B99C0-145F-4C75-91DF-9ADCA3F86315@mindspring.com>
	<86skhpzjrk.fsf@blue.stonehenge.com>
Message-ID: <75E30FA3-0612-44C8-BC8B-0ACE02AB7554@c2group.net>


On Jun 24, 2009, at 12:19 PM, Randal L. Schwartz wrote:

>>>>>> "Hal" == Hal Wigoda <hwigoda at mindspring.com> writes:
>
> Hal> The biggest you have 4XL of possible.
>
> Aka - "Programmer Size". :)


Erm.  Speak for yourself?

Kent Cowgill                                     kent at c2group.net

http://kentcowgill.org/blog           http://youtube.com/kcowgill
http://kentcowgill.org/photos   http://flickr.com/people/kcowgill




From fire at dls.net  Wed Jun 24 15:02:12 2009
From: fire at dls.net (Bradley Slavik)
Date: Wed, 24 Jun 2009 17:02:12 -0500
Subject: [Chicago-talk] YAPC Houston t-shirts
Message-ID: <15581.1245880932@dls.net>

On Wed 09/06/24 16:50 , Kent Cowgill kent at c2group.net sent:
> 
> On Jun 24, 2009, at 12:19 PM, Randal L. Schwartz wrote:
> 
> >>>>>> "Hal" == Hal
> Wigoda  a at mindspring.com> writes:>
> > Hal> The biggest you have 4XL of
> possible.>
> > Aka - "Programmer Size". :)
> 
> 
> Erm.  Speak for yourself?
> 
> Kent Cowgill                                     kent at c2group
> .net
> http://kentcowgill.org/blog          
> http://youtube.com/kcowgillhttp://kentcowgill.org/photos  
> http://flickr.com/people/kcowgill
> 
> 
> _______________________________________________
> Chicago-talk mailing list
> Chicago-t
> alk at pm.orghttp://mail.pm.org/mailman/listinfo/chicago-talk

Sorry that you can't fit in "programmer size".

I use mine as a survival tent in the wilderness of Barsoom.


Bradley

From richard at rushlogistics.com  Sat Jun 27 13:24:30 2009
From: richard at rushlogistics.com (Richard Reina)
Date: Sat, 27 Jun 2009 16:24:30 -0400 (EDT)
Subject: [Chicago-talk] =?utf-8?q?Database_or_Code=3F?=
Message-ID: <20090627202430.22E2424E9@alexander.xo.com>

I have to create a summary report out of a database that contains data about flight departures.

Example:

Chicago, IL Phoenix, AZ United

Each report has to summarize an airline's quantity of unique monthly flights  so that in the end I get:

Continental Airlines:
Chicago, IL to New York 21
New York, NY to Dallas 17
Houston, TX to Dallas 45
Chicago, IL to Phoenix 5

I can't find an SQL query that allows me to parse the data in such a way so I was wondering if I could get some general opinions on what would be the easiest way to do this in perl?

From imranjj at gmail.com  Sat Jun 27 13:55:55 2009
From: imranjj at gmail.com (imran javaid)
Date: Sat, 27 Jun 2009 15:55:55 -0500
Subject: [Chicago-talk] Database or Code?
In-Reply-To: <20090627202430.22E2424E9@alexander.xo.com>
References: <20090627202430.22E2424E9@alexander.xo.com>
Message-ID: <c54a332c0906271355n7a25ce75v4adc27ee0754b8ff@mail.gmail.com>

There are many ways to slice and dice this. One is to build a hash of hashes:
$data->{$airline}->{$origin}->{$destination}++



On 6/27/09, Richard Reina <richard at rushlogistics.com> wrote:
> I have to create a summary report out of a database that contains data about
> flight departures.
>
> Example:
>
> Chicago, IL Phoenix, AZ United
>
> Each report has to summarize an airline's quantity of unique monthly flights
>  so that in the end I get:
>
> Continental Airlines:
> Chicago, IL to New York 21
> New York, NY to Dallas 17
> Houston, TX to Dallas 45
> Chicago, IL to Phoenix 5
>
> I can't find an SQL query that allows me to parse the data in such a way so
> I was wondering if I could get some general opinions on what would be the
> easiest way to do this in perl?
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

-- 
Sent from my mobile device

From warren.lindsey at gmail.com  Sat Jun 27 15:37:04 2009
From: warren.lindsey at gmail.com (Warren Lindsey)
Date: Sat, 27 Jun 2009 17:37:04 -0500
Subject: [Chicago-talk] Database or Code?
Message-ID: <4a469f52.1ac1f10a.2a55.3f36@mx.google.com>

Select depart, arrive, count(*) from flights group by depart, arrive;

-----Original Message-----
From: imran javaid <imranjj at gmail.com>
Sent: Saturday, June 27, 2009 3:55 PM
To: Chicago.pm chatter <chicago-talk at pm.org>
Subject: Re: [Chicago-talk] Database or Code?

There are many ways to slice and dice this. One is to build a hash of hashes:
$data->{$airline}->{$origin}->{$destination}++



On 6/27/09, Richard Reina <richard at rushlogistics.com> wrote:
> I have to create a summary report out of a database that contains data about
> flight departures.
>
> Example:
>
> Chicago, IL Phoenix, AZ United
>
> Each report has to summarize an airline's quantity of unique monthly flights
>  so that in the end I get:
>
> Continental Airlines:
> Chicago, IL to New York 21
> New York, NY to Dallas 17
> Houston, TX to Dallas 45
> Chicago, IL to Phoenix 5
>
> I can't find an SQL query that allows me to parse the data in such a way so
> I was wondering if I could get some general opinions on what would be the
> easiest way to do this in perl?
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

-- 
Sent from my mobile device
_______________________________________________
Chicago-talk mailing list
Chicago-talk at pm.org
http://mail.pm.org/mailman/listinfo/chicago-talk


From warren.lindsey at gmail.com  Sat Jun 27 15:41:08 2009
From: warren.lindsey at gmail.com (Warren Lindsey)
Date: Sat, 27 Jun 2009 17:41:08 -0500
Subject: [Chicago-talk] Database or Code?
Message-ID: <4a46a046.48c3f10a.13a7.3c27@mx.google.com>

Whoops, missed by carrier.

Select airline, depart, arrive, count(*) from flights group by airline, depart, arrive;

-----Original Message-----
From: imran javaid <imranjj at gmail.com>
Sent: Saturday, June 27, 2009 3:55 PM
To: Chicago.pm chatter <chicago-talk at pm.org>
Subject: Re: [Chicago-talk] Database or Code?

There are many ways to slice and dice this. One is to build a hash of hashes:
$data->{$airline}->{$origin}->{$destination}++



On 6/27/09, Richard Reina <richard at rushlogistics.com> wrote:
> I have to create a summary report out of a database that contains data about
> flight departures.
>
> Example:
>
> Chicago, IL Phoenix, AZ United
>
> Each report has to summarize an airline's quantity of unique monthly flights
>  so that in the end I get:
>
> Continental Airlines:
> Chicago, IL to New York 21
> New York, NY to Dallas 17
> Houston, TX to Dallas 45
> Chicago, IL to Phoenix 5
>
> I can't find an SQL query that allows me to parse the data in such a way so
> I was wondering if I could get some general opinions on what would be the
> easiest way to do this in perl?
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

-- 
Sent from my mobile device
_______________________________________________
Chicago-talk mailing list
Chicago-talk at pm.org
http://mail.pm.org/mailman/listinfo/chicago-talk


From richard at rushlogistics.com  Sat Jun 27 16:20:55 2009
From: richard at rushlogistics.com (richard at rushlogistics.com)
Date: Sat, 27 Jun 2009 23:20:55 +0000
Subject: [Chicago-talk] Database or Code?
In-Reply-To: <4a46a046.48c3f10a.13a7.3c27@mx.google.com>
References: <4a46a046.48c3f10a.13a7.3c27@mx.google.com>
Message-ID: <1047315676-1246144771-cardhu_decombobulator_blackberry.rim.net-1146764949-@bxe1188.bisx.prod.on.blackberry>

Wow!  I'm not at the office to try this but it looks very promising! Hopefully it will work with MySQL.

Thank you so much. I will let you know how it goes.
Sent via my BlackBerry. Ignore all the typos.

-----Original Message-----
From: Warren Lindsey <warren.lindsey at gmail.com>

Date: Sat, 27 Jun 2009 17:41:08 
To: Chicago pm chatter<chicago-talk at pm.org>
Subject: Re: [Chicago-talk] Database or Code?


Whoops, missed by carrier.

Select airline, depart, arrive, count(*) from flights group by airline, depart, arrive;

-----Original Message-----
From: imran javaid <imranjj at gmail.com>
Sent: Saturday, June 27, 2009 3:55 PM
To: Chicago.pm chatter <chicago-talk at pm.org>
Subject: Re: [Chicago-talk] Database or Code?

There are many ways to slice and dice this. One is to build a hash of hashes:
$data->{$airline}->{$origin}->{$destination}++



On 6/27/09, Richard Reina <richard at rushlogistics.com> wrote:
> I have to create a summary report out of a database that contains data about
> flight departures.
>
> Example:
>
> Chicago, IL Phoenix, AZ United
>
> Each report has to summarize an airline's quantity of unique monthly flights
>  so that in the end I get:
>
> Continental Airlines:
> Chicago, IL to New York 21
> New York, NY to Dallas 17
> Houston, TX to Dallas 45
> Chicago, IL to Phoenix 5
>
> I can't find an SQL query that allows me to parse the data in such a way so
> I was wondering if I could get some general opinions on what would be the
> easiest way to do this in perl?
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
>

-- 
Sent from my mobile device
_______________________________________________
Chicago-talk mailing list
Chicago-talk at pm.org
http://mail.pm.org/mailman/listinfo/chicago-talk

_______________________________________________
Chicago-talk mailing list
Chicago-talk at pm.org
http://mail.pm.org/mailman/listinfo/chicago-talk

From tzz at lifelogs.com  Sun Jun 28 06:15:37 2009
From: tzz at lifelogs.com (Ted Zlatanov)
Date: Sun, 28 Jun 2009 08:15:37 -0500
Subject: [Chicago-talk] Database or Code?
In-Reply-To: <4a46a046.48c3f10a.13a7.3c27@mx.google.com> (Warren Lindsey's
	message of "Sat, 27 Jun 2009 17:41:08 -0500")
References: <4a46a046.48c3f10a.13a7.3c27@mx.google.com>
Message-ID: <m2iqigfqiu.fsf@lifelogs.com>

On Sat, 27 Jun 2009 17:41:08 -0500 Warren Lindsey <warren.lindsey at gmail.com> wrote: 

WL> Whoops, missed by carrier.
WL> Select airline, depart, arrive, count(*) from flights group by airline, depart, arrive;

If you don't need high performance, and you'd rather compose SQL queries
through Perl objects (ignoring the MySQL vs. Postgres vs. Oracle
vs. Sybase vs. others SQL incompatibilities), I found Rose::DB::Object
to be very good.  There's a few other Perl ORMs out there, a recent
review is at http://blog.afoolishmanifesto.com/archives/822

If you want performance (you issue lots of queries or you have lots of
objects), and of course if you're doing more advanced queries, RDBO will
probably not be right for you.

Ted

From andy at petdance.com  Sun Jun 28 12:20:13 2009
From: andy at petdance.com (Andy Lester)
Date: Sun, 28 Jun 2009 14:20:13 -0500
Subject: [Chicago-talk] What Perl stuff is everyone doing?
Message-ID: <6D262F32-338D-4C14-B684-8D243EF403BD@petdance.com>


What kinds of Perl things is everyone doing these days?

YAPC::NA just concluded out in Pittsburgh.  Anyone have reports from  
the field to share?  I know Josh McAdams and Kent Cowgill and Jason  
Crome all went.

The Movable Type fork looks pretty cool as a concept, although I  
haven't looked at the underlying code.  See Mark Stosberg's article at http://mark.stosberg.com/blog/2009/06/movable-type-fork-is-an-opportunity-to-harness-cpan.html 
  .  He's in Indiana, so I think we should count him as honorary  
Chicago.pm.

xoxo,
Andy


--
Andy Lester => andy at petdance.com => www.theworkinggeek.com => AIM:petdance





From andy at petdance.com  Sun Jun 28 12:21:34 2009
From: andy at petdance.com (Andy Lester)
Date: Sun, 28 Jun 2009 14:21:34 -0500
Subject: [Chicago-talk] What I've been doing lately: Not much Perl
Message-ID: <BF25C781-B313-460D-AB97-582B8A5E57B9@petdance.com>


I've been pretty busy lately with my just-released book "Land The Tech  
Job You Love", and getting ready for the webcast I'm doing with Chad  
Fowler, Ruby guru, on Wednesday titled "Radical Career Success in a  
Down Economy".

To sign up for the webcast, go to http://www.oreillynet.com/pub/e/1360

If you're interested in the book, check out http://www.pragprog.com/titles/algh/land-the-tech-job-you-love

I'm hoping that once I get the initial publicity about the book out of  
the way, I'll have more time to deal with my backlog of module fixes.   
Most of my Perl these days has just been maintaining Perlbuzz.

xoxo,
Andy


--
Andy Lester => andy at petdance.com => www.theworkinggeek.com => AIM:petdance





From andy at petdance.com  Sun Jun 28 12:22:25 2009
From: andy at petdance.com (Andy Lester)
Date: Sun, 28 Jun 2009 14:22:25 -0500
Subject: [Chicago-talk] Promote Perl 6 by saying "Perl 5"
Message-ID: <5A5B1B0A-FF2A-4E11-9766-246EF8F7CCFD@petdance.com>

 From http://perlbuzz.com/2009/06/promote-perl-6-by-saying-perl-5.html

Perl 6 is hurtling toward completion. The specification is nearly  
complete, and Rakudo now passes 68% of the specification tests.  
Applications like November are being written in Perl 6. Perl 6 is no  
vaporware, and the day when Perl 6 is ready for widespread use is  
coming quickly.

Perl 6 has been in development since 2000, and in that time many  
people may have forgotten about the plans we've had for Perl 6. There  
may be those who have never even heard about the plans for Perl 6.  
Those of us who live in the Perl world are aware of the great changes  
afoot, but there are plenty of people who are not.

I think that the time is right to help make those people aware of Perl  
6, and to remind them constantly of what's coming. My proposed  
technique is simple and it takes advantage of the key elements of time  
and repetition to help remind everyone about Perl 6.

     We need to stop referring to Perl 5 as "Perl" and start calling  
it "Perl 5."

Specifying the "5" in "Perl 5" calls attention to the fact that there  
is more than one Perl. It makes the listener or reader who is unaware  
of Perl 6 wonder why the 5 is specified. For the reader who knows  
about Perl 6, hearing "Perl 5" reminds her that Perl 6 also exists.

I don't think it will be too tough. All I ask is that, at the very  
least, when writing about Perl 5 in your blogs or mailing lists that  
you specify the version you're talking about. It doesn't even need to  
be every instance. I'm guessing we'll find that repeatedly saying  
"Perl 5" in a long message will get tedious both for writer and reader.

I think the way to look at it is that "Perl 5" is the formal name for  
the language, and later references can refer to it as "Perl," almost  
like a nickname. Just that first reminder of "Perl 5" will be enough  
to help lodge in the reader's brain.

With enough time & repetition, it will get to be habit in our minds.  
With enough time & repetition, the computing world will be reminded of  
Perl 6 coming soon.

xoxo,
Andy

--
Andy Lester => andy at petdance.com => www.theworkinggeek.com => AIM:petdance





From shlomif at iglu.org.il  Sun Jun 28 13:14:38 2009
From: shlomif at iglu.org.il (Shlomi Fish)
Date: Sun, 28 Jun 2009 23:14:38 +0300
Subject: [Chicago-talk] What Perl stuff is everyone doing?
In-Reply-To: <6D262F32-338D-4C14-B684-8D243EF403BD@petdance.com>
References: <6D262F32-338D-4C14-B684-8D243EF403BD@petdance.com>
Message-ID: <200906282314.38742.shlomif@iglu.org.il>

Hi all,

On Sunday 28 June 2009 22:20:13 Andy Lester wrote:
> What kinds of Perl things is everyone doing these days?
>

well, I've been working on quite a few Perl projects, actually. The first one 
worth mentioning is CPANHQ. See my blog post about it:

http://community.livejournal.com/shlomif_tech/28209.html

Essentially CPANHQ aims to be an open-source, unified, meta-data-enhanced, and 
community-driven interface to CPAN. At the moment, we have little to show for, 
but we've been making progress. CPANHQ was my chance to finally get my feet 
wet with Catalyst and DBIX-Class, and they both seem pretty nice so far.

I also did some work on File-Find-Object and File-Find-Object-Rule :

http://www.shlomifish.org/open-source/projects/File-Find-Object/

In case, you don't know, F-F-O is an object-oriented alternative to File::Find 
that can be instantiated, has an iterative/incremental interface and can 
return result objects. File-Find-Object-Rule is a port of File-Find-Rule to 
use it instead of File::Find, and lately, I've made its ->start() and -
>match() methods truly iterative instead of using ->in() which just collected 
all the results into one humongous array and kept shifting stuff from it. This 
was a severe mis-feature of File-Find-Rule, IMO, and was implied by the 
limitations of File::Find.

Another thing I've done was did some cosmetic work on the Build.PL files of 
some my CPAN distributions. Namely I added some links to resources and 
keywords (that can also be thought of as tags or labels) as specified in the 
META.yml spec:

http://module-build.sourceforge.net/META-spec.html

resources are displayed in the the distribution's pages on search.cpan.org and 
other sites. Author keywords (when they exist, and very few have been defined) 
are currently under-utilised. However, but I recently added some rudimentary 
support for them in CPANHQ (see above), and I hope that they will soon become 
more popular.

A different aspect of my cleanup was clarifying the licensing terms of the 
distributions. My original distributions are under the MIT/X11 licence, but I 
wrote "All rights reserved" there as well, which I was told was incompatible 
with it (or with "The same terms as Perl" for that matter). I also added a 
COPYING file where appropriate.

I've finished reading Mastering Perl. I "forked" its github repository and 
made some corrections, which brian d foy has pulled. I did some similar forks 
too that way for Perl projects.

I did some other work on the CPAN modules that I maintain see:

http://search.cpan.org/~shlomif/?R=D

As a result of experimenting with these various Perl-based technologies, I've 
been building many CPAN distributions as RPM packages on my Mandriva Linux 
system. I found this bash function useful:

<<<
function up() { urpmi "perl($1)" ; }
>>>

So I can say "up MyModule::Sub" and it will install the package containing the 
Perl .pm "MyModule::Sub".

Sometimes building the packages was straightforward, but often I needed to 
apply some patches to the build .spec or the tests.

On the non-Perl related (but still open-source) front, I've been working on 
http://fc-solve.berlios.de/ . I released a stable version only to discover its 
build process was severly broken (especially on Windows) so I had to release a 
x.y.1 release.

Regards,

	Shlomi Fish

-- 
-----------------------------------------------------------------
Shlomi Fish       http://www.shlomifish.org/
Interview with Ben Collins-Sussman - http://xrl.us/bjn8s

God gave us two eyes and ten fingers so we will type five times as much as we
read.

From michael at potter.name  Sun Jun 28 15:20:43 2009
From: michael at potter.name (Michael Potter)
Date: Sun, 28 Jun 2009 17:20:43 -0500
Subject: [Chicago-talk] What Perl stuff is everyone doing?
In-Reply-To: <6D262F32-338D-4C14-B684-8D243EF403BD@petdance.com>
References: <6D262F32-338D-4C14-B684-8D243EF403BD@petdance.com>
Message-ID: <2379dacc0906281520o76e2c1c0x480c8ee18bc42f52@mail.gmail.com>

On Sun, Jun 28, 2009 at 2:20 PM, Andy Lester <andy at petdance.com> wrote:

>
> What kinds of Perl things is everyone doing these days?
>

Mongers,

I am just wrapping up a data conversion project using Perl 5 and Groovy.  I
used Groovy to do some address validation using a Java based package called
MelissaData.

Besides the pretty complex conversions I wrote dozens of reports to identify
inconsistent data so it could be cleaned up before the conversion.

The client was pretty happy with how fast I could create new reports
and make changes to the conversion algorithm.  I did not miss an opportunity
to mention that I was using Perl.

About a year ago I was working on a project to generate Java code using Perl
5.  The Perl reads XML to generate Axis2 based webservices.  The Perl code
generates .java, .wsdl, and .html.  The XML is documentation of mainframe
transactions.  The idea was to turn mainframe transactions into webservices.

Anyway, I am looking for another gig;  I am especially interested in the
Financial SW.  I have been studying CUDA in hopes of using it for financial
SW.  For those of you who may not be familiar: CUDA is a C library used to
turn an nVidia graphics coprocessor into a math coprocessor.

-- 
Michael Potter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090628/2b22c041/attachment.html>

From richard at rushlogistics.com  Mon Jun 29 09:34:25 2009
From: richard at rushlogistics.com (Richard Reina)
Date: Mon, 29 Jun 2009 12:34:25 -0400 (EDT)
Subject: [Chicago-talk] =?utf-8?q?Database_or_Code=3F?=
In-Reply-To: <4a46a046.48c3f10a.13a7.3c27@mx.google.com>
Message-ID: <20090629163425.62D8B2142@victory.xo.com>

This query turned out to be just what I needed thank you very much.


---- Chicago.pm chatter <chicago-talk at pm.org> wrote:
>
> Whoops, missed by carrier.
> 
> Select airline, depart, arrive, count(*) from flights group by airline, depart, arrive;
> 
> -----Original Message-----
> From: imran javaid <imranjj at gmail.com>
> Sent: Saturday, June 27, 2009 3:55 PM
> To: Chicago.pm chatter <chicago-talk at pm.org>
> Subject: Re: [Chicago-talk] Database or Code?
> 
> There are many ways to slice and dice this. One is to build a hash of hashes:
> $data->{$airline}->{$origin}->{$destination}++
> 
> 
> 
> On 6/27/09, Richard Reina <richard at rushlogistics.com> wrote:
> > I have to create a summary report out of a database that contains data about
> > flight departures.
> >
> > Example:
> >
> > Chicago, IL Phoenix, AZ United
> >
> > Each report has to summarize an airline's quantity of unique monthly flights
> >  so that in the end I get:
> >
> > Continental Airlines:
> > Chicago, IL to New York 21
> > New York, NY to Dallas 17
> > Houston, TX to Dallas 45
> > Chicago, IL to Phoenix 5
> >
> > I can't find an SQL query that allows me to parse the data in such a way so
> > I was wondering if I could get some general opinions on what would be the
> > easiest way to do this in perl?
> > _______________________________________________
> > Chicago-talk mailing list
> > Chicago-talk at pm.org
> > http://mail.pm.org/mailman/listinfo/chicago-talk
> >
> 
> -- 
> Sent from my mobile device
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
> 
> _______________________________________________
> Chicago-talk mailing list
> Chicago-talk at pm.org
> http://mail.pm.org/mailman/listinfo/chicago-talk
> 

From Andy_Bach at wiwb.uscourts.gov  Tue Jun 30 10:22:01 2009
From: Andy_Bach at wiwb.uscourts.gov (Andy_Bach at wiwb.uscourts.gov)
Date: Tue, 30 Jun 2009 12:22:01 -0500
Subject: [Chicago-talk] Dr. Dobb's Update - 06/30/09 - Book review of
 "Land the Tech Job You Love"
In-Reply-To: <12592586.156902161246378402759.JavaMail.?@rbg01.pdkp2>
References: <12592586.156902161246378402759.JavaMail.?@rbg01.pdkp2>
Message-ID: <OFD460E81E.8E7F3AFC-ON862575E5.005F299E-862575E5.005F4734@uscmail.uscourts.gov>

Latest Dr. Dobbs update 
http://links.techwebnewsletters.com/servlet/MailView?ms=MzM0NzQ5MDkS1&r=MjIxNTU0Mjk1NAS2&j=NTIyODE4ODQS1&mt=1&rt=0

has a review of Andy Lester's new book!
http://dobbscodetalk.com/index.php?option=com_myblog&show=Land-the-Tech-Job-You-Love-Book-Review.html&Itemid=29

Land the Tech Job You Love Book Review
It goes without saying that times are tough and finding the perfect job is 
even tougher. Chicago-based technologist and author Andy Lester offers 
no-nonsense job search and interview advice in his recent Pragmatic 
Bookshelf title, Land the Tech Job You Love. Read on to find out if the 
tips he shares are worth the book's cover price. [...] 



a
----------------------
Andy Bach
Systems Mangler
Internet: andy_bach at wiwb.uscourts.gov
Voice: (608) 261-5738; 
Cell: (608) 658-1890

The trouble with doing something right the first time
is that nobody appreciates how difficult it was.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.pm.org/pipermail/chicago-talk/attachments/20090630/e5c1c8ac/attachment.html>