Version 27 of Ninth Annual Tcl/Tk Conference

Updated 2002-09-21 10:50:55

September 16-20, 2002 is the time, and Vancouver British Columbia Canada [L1 ] is the city (home of ActiveState) for the Ninth Annual Tcl/Tk Conference. Check http://www.tcl.tk/community/tcl2002/ for all the details.


RS: As one of the Tclers that cannot attend, I would be very happy if the lucky ones in Vancouver could let us know how it goes (maybe right here on this page?)

LV I found an open terminal, and of course the second thing I did (after sending mail to the family) was stop in to update the page!

I arrived yesterday pm. The tutorials are said to have gone well - I have talked to Ken Jones and Clif Flynt who were tired but pleased with the way their presentations went.

Spent the afternoon talking with stevel, Steve Ball, dkf, aku, Clif, Mark Roseman, tclguy, Mac Cody, kennykb, and a number of others whom I have met over the past 3-4 conferences. It certainly is good to see them all again. Donal, Kevin, David Enry and myself grabbed a quick bite at a wonderful greek restraunt, then walked over to Activestate for the reception. Food, libations (for those who ate less than I at dinner), and loud music was the rule of the evening. I got to say hi to a number more Tcler's - like Matt Herdon of ActiveState - the tools manager - and Richard Hipp, as well as others.

After being on the go for about 20 hrs, I went back to the room, decellerated with a bit of tele time, and fell asleep.

This morning, I'm up, chatting with the early birds (Mark informs me that Vancouver is a premier place for Starbucks and other coffee shops...) and waiting for breakfast. We'll let you all know how things are going through the week.

Marty Backe - Thanks for the update. Keep'em comin. Dave Griffin - Hello to all. Sorry I can't be there (again) this year. Sigh!

Ro - Thanks - I wish I was there! Vancouver is far from Ottawa though ;( I can't wait to read Steve Landers' and Jean-Claude Wippler's papers! Anybody know if the talks are going to be available in mp3? I hope so!

stevel - a copy of my paper might be at [L2 ] and JCW's slides at [L3 ] <wink>

YES! Awesome Steve! I was going to go to sleep, but screw it now, I've got this to read! Your paper looks great => THANKS -- ro

LV I don't notice anyone video or audio taping the sessions, so I must assume that the sessions will not be available to watch. I have not yet seen any word on availability of the papers online, nor of the wonderful Proceedings CD that is a part of the conference.

This morning, we had a keynote by Mark, talking about his 10 years with Tcl/Tk and how it played an integral part in the acquisition of Teamwave, a talk by our own jcw about the wiki, and 2 talks (so far) on uses of Tcl in applications - one, by the author of http://www.starfishsystems.ca/ , a remote system admin package, and one by Hugh Dunne about testing high performance fiber optic switches. One more talk, and then time to go to lunch!

Okay, the last talk of the morning was by William Duquette, about his use of tcl in testing deep space network uplinks. I got a kick out of an exchange during the Q&A, where Will assured us that even though the energy used for the uplink is quite powerful, there are lots of safety interlocks to prevent the equipment from frying people or aircraft. That's nice to know! I would hate it if Tcl became known as the software with which to fry humans... LV

A dozen or so of us went out to a nearby food court for lunch - sushi, Vietnamese, other Oriental, something from a French Crepes stand, and I saw at least one person eating Arbys...

Back from Lunch, Jeff Hobbs presented a State of Tcl address [Anyone have a URL for data from that paper]? Mostly the same good info that Jeff presents - historical look back, recent release info, hope for the future. Jeff will, I hope, present results of his Oustervote poll. One piece of good news - MacOS download numbers appear to be slightly up recently. There's hope that a combo of the Tk Aqua port and Tclkit/Starkit will result in a new generation of user/developers on the Mac! But we need lots of people helping in all the areas.

In the afternoon, LV was pressed into service (without any arm twisting) to chair the Tcl/Tk Extensions and Deployment sessions - or what he called Tcl Down Under due to the dominance of Aussie Tclers presenting in this session.

Harvey Davies gave us a look at nap - if you are doing linear algebra, or data visualization, take a look at it; someone referred to it as APL (well, really J) in Tcl.

Next came Steve Ball - always a pleasure to sit under his tutelage. Steve talked about the fantastic performance from libxml2 and the new binding to libxslt. More info on XML is coming today in the form of a paper and a Birds of a Feather (BOF) session tonite.

Last, but certainly not least, was stevel - talking about Starkit and Starpack. There is quite a bit of enthusiasm here - Brent Welch, D. Richard Hipp, as well as some of our new acquaintences, are all thinking of interesting ways they can make use of all of these things. One of the tips that was mentioned was that our keynote, Mark Roseman, has been having some success combining TclPro compiled Tcl code with Starkits - to end up with the ability to slightly obfuscate small pieces of applications (perhaps like license managers). Obviously it isn't designed to be high-security - just something to make things a little more challenging for those whose delight is to beat the system.

Next, we chatted a while and then moved on to the conference dinner, held here at the Crowne. A wonderful salad, samon with roasted veggies and a butter sauce, followed by cheesecake (and libations for those who wanted) was enjoyed by all with whom I spoke.

After dinner, the ActiveState Active Awards presentations were held. [L4 ]. jcw wins one of the awards! [Justice prevails!] [No, wait, that makes it sound as though it would have been unjust for someone else to win. That's not it; it's just that--well, see for yourself how much jcw has contributed.] More on this in the press release!

Following dinner, at least 2 BOFs were held. One was on New features requested for wiki and the other was on Tcl 8.4 features. Unfortunately, a combination of too little sleep and a migrane forced LV to retire earlier than he would prefer. Hopefully someone else will follow up on the remaining festivities last nite. (In fact, apparently LV fell asleep during the features of Tcl/Tk 8.4 BoF - a picture of him snoring in a corner was seen on Thurs am... Sigh - even that sleep didn't relieve the headache much... and now adds to his public embarassment.

LV Thurs started out with a breakfast of crepes/Belgium waffles, granola/cereal, pastries, juice, coffee, tea, and good conversations. Then, we got to see a talk about an interactive presentation program called Javanti - see http://www.javanti.org/ . If you like to see curiousities - take a look. One of the interesting features of the powerpoint like program is that it uses Jacl to provide the ability to associate Tcl scripts to elements on the screen, which means that elements can be interactive, and actually can communicate with one another!

Next was Thomas Burch's work on the electronic version of the 33 volume german dictionary Deutcsche Worterbuch. This series, initiated by the Brothers Grimm in the late 1800s, is a massive work (300 million plus characters!) and is used for scholarly research in the origin of words and their usage. Consists of nearly a gig of data to format with appropriate highlighting, etc. to match the printed product.

Lastly, Simon Hefti talked about combining tDom, mod_websh and apache2 to do dynamic docbook rendering.

We took a break, then had the works in progress session. During this, a presentation on the batteries in the Batteries Included ActiveTcl product were shown by Jeffry Hobbs, Michael A. Cleverly showed nstcl (a layer above various current Tcl database extensions, allowing one to do a bit more generic database programming, Peter MacDonald discussed the fact that he has been rewriting TkHTML to use objects and has been deconstructing BrowseX into components so that he could create a BLite (light browser) for use in a wireless meter reading service that is coming on line soon. He also discussed how he has revamped TML (his template markup languge) and it now defaults to rendering to a text widget, but can generate HTML (except no tables), Latex and plain text as well. Next came Scott Thibault, who showed us an interesting approach to developing Palm applications that his company has developed - called X-Acute (pronounced execute), it is envisioned as a Visual Basic alternative. One writes the application in an XML form, but can embed Tcl to perform some interactions. Applications can make use of PalmOS's persisten storage. On the desktop one compiles these programs into a bytecode, then installs that onto the palm, where an 80k runtime interprets the application. The runtimes are also available for the desktop I believe.

Next we broke for lunch - a private meal at the hotel's restraunt, featuring a fusion/sushi motif. More great conversation ensued. Two of the ongoing topics during breaks, meals, etc. were ways to create some sort of centralized service for developers so that they could register their source distribution, announcing it perhaps through freshmeat, then people interested in a binary version might come to a Tcl Web site, select the platforms and add-ons they wanted, and they would then receive a starkit containing the binary for the platform(s) selected. Another topic was some sort of digital signature process for code on the wiki - however, the sociological issues seem larger than the technological ones.

After lunch, we returned to papers, where Scott Thibault presented his company's work on supporting Tcl bytecodes in a custom processor, then aku talked about the Cisco work on modularizing the core (quick summary - miguel produced a nice modification that may be useful for the core, but community should NOT expect a fully modular tcl from this work - the effort was to reduce the binary size and stack size but was not to turn tcl into a select only the features you want distribution. The results are in the tcl.sf.net cvs in its own branch and is tcl 8.3.4 specific at this time. The specific size savings were not known, and the impact on performance were not tested.

Last, but not least, on Thursday was Mac Cody's talk on Palm Tcl and Toucan. Work continues on this product, which one user described as the first PalmOS development product which actually lived up to its description.

After a break, an application show and tell was held. Komodo was shown, as was information by Xerion Pharmaceutical AG (xerion-pharma.com) regarding a bioinformatiocs server and some system admin tools were discussed.

Paragon - a 3D (opengl) file format verifier, written by a gaming machine software company, was shown.

Bryan Griffen of Model Tech showed some of the widgets his company had developed in their work on ModelSIM, a chip simulator. There were a nice tree widget, dockable toolbar, resizable/movable panes, etc. Some of these widgets, written in incr tcl, are hoped to be released back to the community in the future.

Ed Hume, of http://www.hume.com/ showed a package manager application. It used metadata provided with a package to show a user what packages were available, assist in installing new packages or new releases, remove packages, etc.

Ed also showed a nifty data entry system which included a persistent in memory sql database to which applications could subscribe to particular data so that when that data changed, the application would receive notification. It had lots of tools for displaying tables of data, charting the data, creating data entry screens, etc. See Ed's web site for more details.

We then had a few hours break - so LV slipped away to his room and tried to sleep for an hour, so that he could stay awake for the evening's BoFs. It seemed to help a bit ... there wasn't any obvious snoring going on from him this time.

First informal group was for scientific computing and visualization. The author of nap was looking for co-developers to help extend and improve his code. He was looking for fast rendering of x-y plots of tens of thousands of points, FFT and other image manpulations, etc. People suggested Vtk, Snack (which might have some fft code), e4graph, and others which unfortunately LV didn't get in his notes. Hopefully some of the others there will chime in here.

Next was an XML BoF chaired by Steve Ball. He showed a few new tools under development - XML Tool (on the tclxml cvs repository) which is a wizard like application to step through checking the well formedness, validity of an xml document and then to apply XSLT transformations. It's GUI is configured with XML. After than, a Simple Content Management System with makefile like processing was shown, followed by InfoEdit (an XML editor). Also mentioned is upcoming online tutorials on TclXML by Zveno. Finally, another XML editor, by Roy Numri (spelling?) was shown. Its user interface also was configured, dynamically, by XML statements.

Next was a BoF on needs/concerns regarding the Tcl package command. I will defer to dgp providing a summary of the events.

Lastly was a BoF on AOLServer, which also will await notes from one or more of the attendees, as this writer slipped out and back to bed for a few more hrs of sleep...

On Friday, things were busy, as the time was short, our attention span grew shorter, and yet there were so many things we wanted to do.

A paper by miguel, kennykb, and tclguy led off the morning. It covered the bytecode optimizer - but the presentation, instead of being a dry paper about bits and bytes, covered highlights of some of the changes as well as a number of helpful tips on writing Tcl scripts in a way to get more performance out of the machine. Expect to see this info in some best practicess pages in the upcoming days. Next was jcw's paper on critcl. The presentation went well (there were no pitchforks and angry mobs...) but I got the impression that there were no dancing peasants in the street either... Finally, Andrej Vckovski preseneting a paper on Netcetra's comparison of Jython, Rhino and Jacl to determine what scripting language is the leader for Java. Their decision to use the Bean Scripting Framework to abstract away the details was the result - the distinguishment between the 3 contenders was just too close. After a short break (as people were saying their goodbyes, etc.) We came back for the last matters of the day. Brian Griffin presented a paper on Model Tech's work to make their Tk applications look more like windows (which was an excellent example of how if one spends the effort tuning (and writing custom code) one can get better results, was a great job. The final paper was one which unfortunately did not get presented due to information export restrictions. In its place, Kevin and Jeff presented the awards for the conference programming challenge. Details on this will follow on its own page in the days ahead.

Closing remarks and town hall were the finale of the week. There were no serious complains about the conference. In looking ahead, 3 possibilities were suggested about another Tcl gathering:

  1. Yet another Tcl conference as this year
  2. A Tcl track in a Small Languages Conference by Usenix or at an Open Source Conference by O'Reilly
  3. A Tcl workshop at one of the major Usenix conferences.

Several of us attended expressed the notion than a workshop at a major conference would provide us additional ammunition to be able to attend - likelihood of relevance of other material to work might be used as additionl justification, cross pollination within the larger population possible, ability to do Tcl papers in other tracks, etc.

It remains to be seen what will work out - but people should consider contacting tclguy to get on a mailing list to discuss the future direction.

The afternoon was spent doing more sushi with conference attendees, then discussions at activestate by a smaller group followed by Korean food with a smaller group followed by final goodbyes on the part o LV. Now I have 5 minutes to check out and head back to the US. More notes will likely arrive, plus watch http://www.tcl.tk/ for papers and hopefully slides from presentors in the days ahead.