[Chicago-talk] New system architecture

Joel Limardo joel.limardo at forwardphase.com
Mon Jun 1 10:37:52 PDT 2009


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



More information about the Chicago-talk mailing list