Updated 2003-08-05 12:21:27

Some notes, reviews, etc on Tcl'2003.

See Tenth Annual Tcl/Tk Conference. CF: My notes are a bit long, so I'll put them here, rather than merge into the Tenth Annual Tcl/Tk Conference page.

Many years ago (Tcl in Monterrey, with the Tk workshop), Jacob Levy saw me with my guitar and said I should drag it out. So, I opened the conference with a song and dance. Actually with a couple songs, since I dance worse than I sing.

This is a 'standard' 12 bar folk blues riff, I do it in E using variations on the 'Stealin' E-E6-E7-E6 pattern.
 Tcl Coder
        Copyright 2003 Clif Flynt

 I used to write my programs in C and FORTRAN-2,
 Just give me access to a teletype, and there was nothing I wouldn't do.
 I'd print out next month's calendar, with last month's Playboy diva,
 Even play the drum solo from In-In-Gadda-da-Vida.

 But I don't do that any more.  There's better programs to be done.
 These days I write my code in Tcl, just type it in and let it run.

 Yeah, I'm a Tcl Coder,
 I wrote 4 apps so far today.
 Give me a keyboard and an editor,
 And try to keep out of my way.

 Now, C has got it's pointers, its mallocs and its frees,
 A single buffer overrun and your program's on its knees.
 But now I code in Tcl, I don't worry 'bout those things.
 Don't deal with Mallocs, Frees or Ints or Floats, when everything's a string.

 Yeah, I'm a Tcl Coder,
 I wrote 5 apps so far today.
 Give me a keyboard and an editor,
 And try to keep out of my way.

 Some folks like a structure, and some folks like a class.
 You get a little feature creep, and your design is on its ass.
 Get all your spec's up front, and that won't happen,
                so they say.
 But I live in the real world - give me associative arrays.

 Yeah, I'm a Tcl Coder,
 I wrote 6 apps so far today.
 Give me a keyboard and an editor,
 And try to keep out of my way.


I moved from there to a song I wrote based on Mark Twain's "Life along the Mississippi". The song describes a Riverboat pilot's career from dream-filled kid to technolgically obsoleted specialist (post railroad). In a 20 year span being a RiverBoat Pilot moved from a job that didn't exist, to a licensed monopoly, to a job that didn't exist.

I don't want being a Tcl expert to follow that pattern.

And then I introduced Larry McVoy for the KeyNote address.

Larry berated us sternly for leaving Tk in the 1980's for look & feel. He pointed out that Tk is great when an engineer needs to hack a quick GUI, but is death if you are making a commercial product. He wouldn't have a company if it weren't for Tk and the early GUIs he made with it, but now he'd hit a ceiling and couldn't compete in the 'Real World' with the normal Tk look & feel. He was ready to leave Tk for Java to get the look & feel he needs to compete in marketplace.

By the end of the conference, he's sold on the Tcl Community, what we (and Tk) can do, and is helping organize an effort to revamp the Tk appearance.

We had a couple of 'normal' type Tcl application talks, and then a paper by Fred Hansen, an SEI researcher who admits that Tcl is suitable for large projects, but again felt that Tk contributed to bad GUIs, and felt that there were no examples of good GUIs. He got a lot of pushback from the audience. Folks like Steve Landers, Brian Griffin and Brian Oakley who build good looking commercial GUIs in Tcl/Tk pointed out that good GUIs can be done in Tk. On the other hand, the GUI's most of us build are simple compared to the 300+ screens in the application they were studying.

We went from there to a panel with Mark Ackerman (an HCI prof from UMich), Larry McVoy, Fred Hansen, and Brian Oakley debating Tk/Tcl strength/weakness. Again, there was a lot of push from folks who have 'fixed' Tk to look better, and some good points on all sides. My take is that it's so easy to get out a Tk GUI that folks can create bad ones. It takes longer to make *anything* in Java, so folks are forced to spend longer thinking about the problem, and may do a better design on first go (though they could go through several Tk revisions in that same time and *could* end up with a better GUI in less time. Too few go through the revisions.)

One of the conclusions was that you need a Human Factors expert somewhere in the mix if you are going to create top quality commercial GUIs. The engineer who understands the implementation details is probably not the one who can abstract things to a user and GUI level.

lv Any chance that Larry could help recruit a human factors expert for the core tk work?

MR Larry may have some resources to kick in, and we've got some "moving forward" stuff in the works. But it's not really an HF person needed for the core work.. anyway, more later.

Bryan Oakley Yeah, human factors has more to do with how to use widgets rather than defining what specific widgets might look like. We don't need an architect to tell us we need more bricks :-)

lv I understand your point - my thinking was that with a HF expert in the mix we make certain that the bricks we build can be used for the types of architecture that people want to build. No use building bricks that are only good for decorating 18th century chapels if people are wanting to build 21st century homes...

After that, the banquet. Ken and I missed this as we did the final prep for a networked Tcl Quiz game that everyone could play. Jeff Hobbs presented the ActiveState awards to Cameron Laird and Richard Suchenwirth.

The networked quiz game seemed to go over well. Each table was a team, with a laptop they clustered around, and Ken Jones MC'd the questions and recounting the score. The questions were courtesy of Ken Jones, game engine courtesy of me, prizes from ActiveState, Prentice Hall (4'th edition Welch), Morgain Kauffman (2'nd edition Flynt), and T-Shirts from my first book.

The Bof on how to do manager speak to sell Tcl turned into a bit of a self-help encounter session. Maurita does "how to speak to management" seminars for engineers professionally and had a few points on how to present data to managers (these are big-picture folks - tell them how much money it will save) and other technical types (These are detail oriented folks - explain how much more productive it is, show them examples.)

Ken Jones, Cameron and I led a Publish/Publicity in Tcl/Tk BoF next. We wandered all over the map in the hour and a half or so we talked. We suggested looking at vertical markets (trade rags in non programming fields) to publish Tcl success stories, KJ, CL and I explained how we got into writing the book, and what it's like to write one, John Harris extolled the virtues of Literate Programming. I suggested folks throw a Tcl/Tk BoF at any conference they attend, just to let the world know we're still alive, and describe the latest additions to the language. There was more I've fogged on.

Friday was more 'normal' type Tcl papers and presentations.

At the Core Team Meeting we started getting into what can be done to revamp Tk.

The wine BoF went well. We held it in the suite the hotel gave me when they saw how much equipment we were moving and storing each day. When you find someone elses clothes in your living room the next morning, you know it was a good party. I finally kicked folks out a bit after 01:00 so I could get to sleep.

A local bookseller brought copies of Welch's, my and Libes' books for sale. Ken and I autographed copies of our respective books.

There was a lot of general socializing, and a lot of "how do we fix Tk" talk. A half dozen of us held a "Games in Tcl" BoF in the corner discussing various GUI tricks, looking at the games and what can be done to update look&feel and what is already cool. Kevin Kenney led a lot of this. No surprise, you can do a lot with the Tk canvas.

Saturday was a 'normal' set of sessions, and the final Town Hall set up a bunch of action items for the next few months to start reworking Tk. Larry and Jeff had their heads together a lot over Thur, Fri and Satdy. I don't know everything they've discussed, but I'm certain we'll be seeing some changes. Brian Griffin put together a set of option commands and dropped them onto the widget demo to show how easy it would be to make Tk look modern, and Larry was impressed and amazed. I believe we'll see those options made available soon. -- (Any chance we can get a rundown of these options? I'm quite curious to see what was done.)

Saturday evening, the last dogs went out for dinner. This group included Jeff Hobbs, Andreas Kupries, Brian Oakley, Larry McVoy, Steve Cassidy, Peter Spjut, Sheila Herndon, and me. Jeff and Larry kept their heads together, with Andreas and Brian Oakley kicking in comments, and me listening to two sets of conversations and following neither well enough to contribute (but throwing in the wrong comments anyhow.)

Aside from that. Nothing much happened at Tcl'2003.

Mac Cody Sounds like a good conference. Sorry I missed it. What is going to be the mechanism for posting the proceedings from the conference? It would be great if output from the BoFs could be posted as well. It think that there are a lot of things that can be learned from them, especially the BoF on selling Tcl to our managers.

DKF: Not all the BoFs had output. The Games BoF was just fun.

CF: I'll encourage anyone with comments/reviews of BoFs, papers etc to post them somewhere. The Conference CD-ROM and Paper Proceedings will be available for purchase at some price similar to at the conference as soon as I figure out how to move the money and 'stuff'. What I can do for A/V materials depends on just what I can put together. There is, of course, no monetary budget for reworking the video materials, so that will move slowly.

[ Category Community | Tcl Conferences ]