Saturday, July 26, 2003
Tim Bray asks: What can we do?
The W3C TAG group is faced with a long-running debate on the nature of resources, which as he states, don't really exist an any level that matters to code. On the other hand, removing all mention of resources (the R in URI) would cripple the usefulness of the developing webarch document.
So do this: Still talk about resources, but clearly separate that discussion from concrete, code-affecting concerns. There are multiple ways to look at the 'resources problem', none of which are strictly right or wrong, so give us an unbiased report on all the main ones. The main challenge in this approach is getting folks to agree that their favorite world view isn't the only game in town. It'll be tough, but is there any other way forward?
Monday, July 21, 2003
What's been bugging me about REST
REST, the influential PhD dissertation by Roy Fielding, is a huge success. It has shaped the viewpoints of countless Web developers, especially in the standards arena. It's even helped SOAP 1.2 become a better citizen on the Web. But there's always been something about it nagging at the back of my mind. Now I know what it is.
I've worked with Web developers for many years, and they tend to be a pretty concrete bunch. They write code that deals with HTTP requests for a file or Servlet responses to POST messages--hypertext dispensers. Ordinary developers don't talk about whether a http URI can actually stand in for physical objects like cars in machine-readable assertions. Before today, I never was able to internally square the development reality with the elegance of REST.
In a posting to www-tag today, I offered that both the 'hypertext dispenser' way of looking at the web AND the abstract resource/representation viewpoints are valid interpretations. The webarch document should be divided into two levels: that of the concrete, running Web code (which is unquestionably a success), and that of interpretations. Interpretations are important, because they shape how the people writing the code for the Web approach various problems. In order to be most useful, the webarch document shouldn't be in the business of pushing one particular interpretation. Instead, it should guide Web developers toward understanding the interpretation that differs from the "one they grew up with", which will lead to a better, more informed developer base. And eventually, a better Web. -m
Sunday, July 13, 2003
This was my first time giving a tutorial, and my longest talk yet (3 hours, not counting breaks). The audience was packed, and attentive. They had lots of good questions. Next time I go in front of a crowd to teach XForms, I'll be better prepared.
Got applause when I mentioned that the book will be free (as in freedom) under the GFDL. Guess that issue resonates with that particular audience. :-) The hand-out for the tutorial also included early proofs of 3 chapters from XForms Essentials, which went over very well. I had a good chat with Todd Mezzulo, over ways to effectively market the book.
Some of the most interesting parts of the conference happened outside of the sessions. I had meals with (in order, roughly) Antone Quint, Eric van der Vlist, Simon St. Laurent, Guido van Rossum and Paul Prescod, Edd Dumbill, Mitch Kapor and most of the OSAF team, and on the last day, Wil Wheaton who autographed Dancing Barefoot.
Next year, I'll have something of my own to sign. -m
Friday, July 11, 2003
Episode 0596003692: Killer Guinea Fowl from Outer Space
XForms Essentials is now available on Amazon for pre-order. A minor mystery is how they calculate "sales rank" for something that can't be bought yet. For now, it's at 1,841,722. :-)
Buy from Amazon.com
Bonus points for noticing the subtle error in the listed title there... [Update: It's fixed, thanks for playing...] -m
Saturday, July 05, 2003
Preparing for the tutorial session at OSCON, I again came across Sebastian Schnitzenbaumer's project: A full implementation of XHTML, XForms, CSS, SVG, XFrames and other neat stuff.
Oh yeah, it's written in Flash+ActionScript, so it runs on any remotely recent browser platform.
This project has come a long way since I last looked at it. I'm impressed. -m
Tuesday, July 01, 2003
For external use only. I doubt
the enforcability of click-through licenses anyway. Copyright 2003 Micah Dubinko. All rights reserved.