From matthew_heusser at mcgraw-hill.com Fri Jan 25 08:13:25 2002 From: matthew_heusser at mcgraw-hill.com (matthew_heusser@mcgraw-hill.com) Date: Wed Aug 4 00:01:18 2004 Subject: My Presentation ... Message-ID: <85256B4C.004E8F6F.00@corpny55wls01.mcgraw-hill.com> Folks: My presentation today has the potential to run long. We could mitigate this by my starting the presentation whilst folks eat, or else I could forcefully limit questions. Or I could just run long. Thoughts? Matt H. From matthew_heusser at mcgraw-hill.com Mon Jan 28 06:49:51 2002 From: matthew_heusser at mcgraw-hill.com (matthew_heusser@mcgraw-hill.com) Date: Wed Aug 4 00:01:18 2004 Subject: Spell Check ... Message-ID: <85256B4F.0046E5D6.00@corpnj148ls01.mcgraw-hill.com> This weekend, I was thinking that XP, Peopleware, SMNP, and wireless networking are cool, but they aren't really ... perl. So, my proposal is that interspersed with various CS- appealing topics, we drill down deep into perl and cover some things that many folks are 'scared' of. Regular Expressions. Hash Tables of Arrays of Hash tables (and thier uses). Packages. Which got me to think: Making a spell checking package would be pretty darn easy(*). In fact, I could do it for around $500, (assuming we find a sponsor ... isn't there some design firm that wants this functionality somewhere in GR? :-) and get GR.PM to Code Review it an QA it. I would donate 80% of the proceeds to GR.PM, and we could use the money to pay our speakers, or subsidise lunch, or have door prizes, etc. (I'm guessing this would make attendence and presentations shoot up.) This could make an interesting open-source project ... Then again, If someone asked nicely, I'd probably do this for free. (Or use it as an independent study in Grad School or something ...) Comments? regards, Matt H. (*) - Literally, you'd pass in a string, I'd call a function to remove punctuation (which also turns "can't" into "can not", etc.), then I'd call split() ... then read the dictionary into a hash table, do a foreach loop, and return an array ... The biggest challenges are getting an ASCII english dictionary and performance ... I think if we embed the dictionary into the perl script and pre-compile it (then put it onto a caching server) we're okay ... hmm .. From albert.tobey at priority-health.com Mon Jan 28 09:49:14 2002 From: albert.tobey at priority-health.com (Al Tobey) Date: Wed Aug 4 00:01:18 2004 Subject: Spell Check ... In-Reply-To: <85256B4F.0046E5D6.00@corpnj148ls01.mcgraw-hill.com> References: <85256B4F.0046E5D6.00@corpnj148ls01.mcgraw-hill.com> Message-ID: <1012232959.975.0.camel@zorak> Matt, you really can't keep doing this! It takes me hours to clean the flame-thrower! ;) Umm. There are already a bunch of spell checkers out there that are both perl and C/C++ (cross-platform). Linux Journal actually published a neat article that will spell check your typos and make sure they fit with the desired keyboard layout ;) In the article, the author talked about how typos differ from QWERTY keybards to DVORAK keyboards. His program would detect QWERTY typos and translate them to the proper keyboard-typo-type. $500 for a spell checker in perl (or any other language) is rediculous. For one in C, it is knavery as there are 3 or 4 well-known API's in the wild that are free (as in beer or GPL). Beyond that, AFAIK, in the Microsoft world you have COM objects so you can tie into the same spell checker used by Word and Outlook. You can also connect to COM objects from ActiveState perl, last I saw. I'm not a Microsoft programmer, so I'm not too sure about that. I know you feel that software should have a stiff price, but this is downright rediculous. None of the options I mention below require open sourcing your project, and furthermore, a couple of them don't even require you to give a link to the source (although, with perl it would be a good idea if you didn't bundle it). A quick search on CPAN results in these: Lingua::* There are at least two spell checkers for English in this group on a quick scan ... Text::Pspell, an interface to the C library pspell Text::Ispell, an interface to the unix spell checker, ispell There is also the Apache module mod_speling (yes, that's what it is called), among many other options. But, all in all, if there is some sucker^H^H^H^H^H^HCIO out there that will pay $500 for something he can have (or might already have) for free, then I guess that's up to him. Maybe I can sell him some stock in my website which will make money by not asking for any and not selling anything ;) -Al Heckler Extraordinaire and Unix Administrator On Mon, 2002-01-28 at 07:49, matthew_heusser@mcgraw-hill.com wrote: > > This weekend, I was thinking that XP, Peopleware, > SMNP, and wireless networking are cool, but they > aren't really ... perl. > > So, my proposal is that interspersed with various CS- > appealing topics, we drill down deep into perl and cover > some things that many folks are 'scared' of. > > Regular Expressions. Hash Tables of Arrays of Hash > tables (and thier uses). Packages. > > Which got me to think: Making a spell checking package > would be pretty darn easy(*). In fact, I could do it for around > $500, (assuming we find a sponsor ... isn't there some design > firm that wants this functionality somewhere in GR? :-) > and get GR.PM to Code Review it an QA it. I would donate > 80% of the proceeds to GR.PM, and we could use the money > to pay our speakers, or subsidise lunch, or have door prizes, etc. > (I'm guessing this would make attendence and presentations > shoot up.) This could make an interesting open-source project ... > > Then again, If someone asked nicely, I'd probably do this for free. > (Or use it as an independent study in Grad School or something ...) > > Comments? > > regards, > > Matt H. > (*) - Literally, you'd pass in a string, I'd call a function > to remove punctuation (which also turns "can't" into "can not", etc.), > then I'd call split() ... then read the dictionary into a hash > table, do a foreach loop, and return an array ... > > The biggest challenges are getting an ASCII english dictionary > and performance ... I think if we embed the dictionary into the > perl script and pre-compile it (then put it onto a caching server) > we're okay ... hmm .. > > > ******************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the Priority Health Information Services Department at (616) 942-0954. ******************************************************************** From matthew_heusser at mcgraw-hill.com Mon Jan 28 11:11:43 2002 From: matthew_heusser at mcgraw-hill.com (matthew_heusser@mcgraw-hill.com) Date: Wed Aug 4 00:01:18 2004 Subject: Spell Check ... Message-ID: <85256B4F.005EE147.00@corpny55wls01.mcgraw-hill.com> Gee Al, you really should like, come to meetings and stuff. At the meeting, someone said it was worth 300 man-hours of effort. I figured I'd see if they'd pay $500.00. :-) For that matter, I then wanted to give 80% of the proceeds to GR.PM, mostly because I knew it wasn't worth that much. Still, free books and lunches are like, cool and stuff. But, even if I downloaded a module from CPAN and installed it, I'd charge a one hour fee. Hence the remaining 20%. (Most enterprises still need to pay somebody to FTP and download and call the script.) Besides that, note that I admitted I'd do it for free. The point was not to get $$$, the point was to write a package in Perl and then show the GR.PM's how to write packages in Perl. Last time I checked, that was our mission. So it's redundant. So what? Make sense? Matt H. From brandon at squareonedesign.com Wed Jan 30 08:44:08 2002 From: brandon at squareonedesign.com (Brandon Gohsman) Date: Wed Aug 4 00:01:18 2004 Subject: Spell Check ... In-Reply-To: <85256B4F.005EE147.00@corpny55wls01.mcgraw-hill.com> Message-ID: Matt, I sent your idea to Lin and Mike, the partners here at Square One, to see if they are interested. If they aren't, I am going to have to berate them. Here is a problem, however, that I constantly have to deal with: We don't host anything ourselves. Because of this, we don't have administrative access to any of the web servers that our sites end up on. This can make installing a Perl module more of a political issue than a semantic one. Any suggestions on how to get around this sort of ilk? Thanks, Brandon Gohsman Square One Design 560 5th Street NW, Suite 301 Grand Rapids, MI 49504 T 616.774.9048 F 616.774.8003 E brandon@squareonedesign.com W http://www.squareonedesign.com/ -----Original Message----- From: owner-grand-rapids-pm-list@pm.org [mailto:owner-grand-rapids-pm-list@pm.org]On Behalf Of matthew_heusser@mcgraw-hill.com Sent: Monday, January 28, 2002 12:12 PM To: albert.tobey@priority-health.com; grand-rapids-pm-list@happyfunball.pm.org Subject: Re: Spell Check ... Gee Al, you really should like, come to meetings and stuff. At the meeting, someone said it was worth 300 man-hours of effort. I figured I'd see if they'd pay $500.00. :-) For that matter, I then wanted to give 80% of the proceeds to GR.PM, mostly because I knew it wasn't worth that much. Still, free books and lunches are like, cool and stuff. But, even if I downloaded a module from CPAN and installed it, I'd charge a one hour fee. Hence the remaining 20%. (Most enterprises still need to pay somebody to FTP and download and call the script.) Besides that, note that I admitted I'd do it for free. The point was not to get $$$, the point was to write a package in Perl and then show the GR.PM's how to write packages in Perl. Last time I checked, that was our mission. So it's redundant. So what? Make sense? Matt H. From matthew_heusser at mcgraw-hill.com Wed Jan 30 09:28:26 2002 From: matthew_heusser at mcgraw-hill.com (matthew_heusser@mcgraw-hill.com) Date: Wed Aug 4 00:01:18 2004 Subject: Spell Check ... Message-ID: <85256B51.00556883.00@corpnj148ls02.mcgraw-hill.com> To: Bradon: CC: GR.PM For CS 658, I just put the Perl packages in the same directory as the script, and that seemed to work. I suspect that if they are in the same directory as the script _or_ in your PATH (or some "lib") environmental variable, perl will find the package. If you host on a UNIX box and want to use a "built-in-UNIX-API" calls, then I suspect you (or I) can write a small "C" function to get a string and return "words not found", and call this from perl relatively easily. With competant system admins (ok, Al, I have to admit, competant admins make a difference), I've had no problem asking a question like "Hey, I need TCL or G++ or whatever-else free module, and I need it in my Path", so you could probably just get that Perl/Unix-API module FTPed to the server. Hopefully, you've got competant admins, right? Final thought: So far, all these spell-checking ideas require an explicit trip to the web-server. You'll have click a "submit" or "spell check" button, and the web page could return, just as before, with a list of failed words at the top. (before your text box. Or you could try to replace words in the box, but I don't like that idea.) If you want to avoid the submit, it would be cool to write a Java App that checks through a socket connection. (a "web service" ... wow ...) We could hook it up through JavaScript and not require a Web-Page "Refresh." There might even be a free or Cheap component out that does this. But, thing is, we're a perl users group, and I don't do Java. But to a java person, I'll bet it's a neat idea ... regards, Matt H. From bill_day at mcgraw-hill.com Thu Jan 31 09:54:30 2002 From: bill_day at mcgraw-hill.com (bill_day@mcgraw-hill.com) Date: Wed Aug 4 00:01:18 2004 Subject: Networking question Message-ID: <85256B52.00578F51.00@corpny55wls01.mcgraw-hill.com> This is not a Perl question, but I know there are opinionated net savvy people on this list. I'm taking CS 654 (Computer Networking) at GVSU this semester. The class requires a survey paper. The potential topics that the professor has suggested are really lame. I thought it might be fun to do something like "A comparison of Linux vs. BSD IP implementations, a source code perspective". So I went and took my first look at the Linux kernel source code (net directory). This looks like it might be over my head (I'm more of a UNIX wannabe than a guru). On the other hand this is the one time in my life when it may make sense to dive into this code. My questions for the list are: Is this a reasonable topic? Is my scope too large? Does anybody know of good resources other than just the source code? If I'm out to lunch, does anybody have ideas for a cool topic for a paper for a networking class? Whatever I end up doing, I'd be willing to present at a future meeting. Your input is appreciated. From beaker457 at hotmail.com Thu Jan 31 11:57:29 2002 From: beaker457 at hotmail.com (Ben Carlson) Date: Wed Aug 4 00:01:18 2004 Subject: Networking question Message-ID: Bill, Networking certainly is a good topic and has been developing on and off for many years so I'm sure there is a great deal of depth to cover with comparing two operating systems (BSD and Linux). To communicate via the net don't both BSD and Linux both have to output a certain protocol, maybe redefining your scope to one OS and concentrating on the reasons why they chose certain avenues to get to the final result could be a possible topic of discussion that might give you a more defined goal. Personally I would just be happy to understand how one of them implements the IP stack and protocols. I would certainly focus on Linux, for the most part a great deal of it's implementation of things are simpler. This not only makes for an Operating system you can understand, but one you can tweak and make better. Good luck, Ben Carlson >From: bill_day@mcgraw-hill.com >To: grand-rapids-pm-list@happyfunball.pm.org >Subject: Networking question >Date: Thu, 31 Jan 2002 10:54:30 -0500 > > > > >This is not a Perl question, but I know there are opinionated net savvy >people >on this list. > >I'm taking CS 654 (Computer Networking) at GVSU this semester. The class >requires a survey paper. The potential topics that the professor has >suggested >are really lame. I thought it might be fun to do something like "A >comparison of >Linux vs. BSD IP implementations, a source code perspective". So I went and >took >my first look at the Linux kernel source code (net directory). This looks >like >it might be over my head (I'm more of a UNIX wannabe than a guru). On the >other >hand this is the one time in my life when it may make sense to dive into >this >code. > >My questions for the list are: Is this a reasonable topic? Is my scope too >large? Does anybody know of good resources other than just the source code? >If >I'm out to lunch, does anybody have ideas for a cool topic for a paper for >a >networking class? > >Whatever I end up doing, I'd be willing to present at a future meeting. > >Your input is appreciated. > > _________________________________________________________________ Send and receive Hotmail on your mobile device: http://mobile.msn.com From albert.tobey at priority-health.com Thu Jan 31 21:48:05 2002 From: albert.tobey at priority-health.com (Al Tobey) Date: Wed Aug 4 00:01:18 2004 Subject: Networking question In-Reply-To: <85256B52.00578F51.00@corpny55wls01.mcgraw-hill.com> References: <85256B52.00578F51.00@corpny55wls01.mcgraw-hill.com> Message-ID: <1012535291.1191.3.camel@zorak> Hi Bill, Being the leading Linux bigot on this list, I feel qualified to chime in here. The linux vs. bsd debate has been beat around quite a bit already. Until kernel 2.4, Linux's network stack was actually based on the BSD stack. OS-X (darwin) would make for a nice fresh comparison, but it is actually based on the BSD network stack with a handful of changes. What you'll find is that it's difficult to gauge the efficiency of a network stack all by itself. Many things affect the throughput of the stack such as NIC driver quality, VM, I/O latency, buffering, and point of focus (security vs. speed vs. scalability). These may be interesting things to gauge all on their own. You'll also notice that there isn't much to be gained from analysis of the two source codes unless you start breaking them down into extensive mathematical analyses of various algorithms within. For the most part, even this data would be useless as the two stacks differ quite a bit on how they approach various problems. If you really think you can get a lot out of reading the source, then this would be a fairly original examination as most comparisons to date are done on the system level with standard benchmarks and empirical measurements. What might be more interesting and refreshing is a look at how all of the IPv6 compliant operating systems are doing things. This is still mostly uncovered territory (thus more interesting) and varies a lot between systems. This would probably be a bit smaller scope, but would involve more research and could make way for more background information (read: fluff) in case you're running short on pages. -Al On Thu, 2002-01-31 at 10:54, bill_day@mcgraw-hill.com wrote: > > > > This is not a Perl question, but I know there are opinionated net savvy people > on this list. > > I'm taking CS 654 (Computer Networking) at GVSU this semester. The class > requires a survey paper. The potential topics that the professor has suggested > are really lame. I thought it might be fun to do something like "A comparison of > Linux vs. BSD IP implementations, a source code perspective". So I went and took > my first look at the Linux kernel source code (net directory). This looks like > it might be over my head (I'm more of a UNIX wannabe than a guru). On the other > hand this is the one time in my life when it may make sense to dive into this > code. > > My questions for the list are: Is this a reasonable topic? Is my scope too > large? Does anybody know of good resources other than just the source code? If > I'm out to lunch, does anybody have ideas for a cool topic for a paper for a > networking class? > > Whatever I end up doing, I'd be willing to present at a future meeting. > > Your input is appreciated. > ******************************************************************** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the Priority Health Information Services Department at (616) 942-0954. ********************************************************************