- From: Al Gilman <asgilman@iamdigex.net>
- Date: Sun, 17 Dec 2000 11:17:44 -0500
- To: Davey Leslie <davey@inx-jp.org>, Marti <marti@agassa.com>, "Charles F. Munat" <chas@munat.com>, <jim@jimthatcher.com>, "'Kynn Bartlett'" <kynn@idyllmtn.com>, <w3c-wai-ig@w3.org>
At 11:32 PM 2000-12-17 +0900, Davey Leslie wrote: >Thus spake Marti on 00.12.17 8:57 PM at marti@agassa.com: > >> and it seems to me >> that some of this is like insisting all the stairs be removed when ramps are >> installed. Sure, that would guarantee the same access for all, but it would >> punish those who walk by making them take a longer route. > >Hi Marti, > >I wonder if that's the correct analogy. It seems to me more like widening a >doorway so that all may pass more easily. I don't believe that use of CSS >makes anyone "take a longer route." Or does it? > Many of the people building user agents for mobile devices of cell phone size and smaller represent that it does, although others contest this point. The fact that its syntax is inconsistent so it adds parsing code to the application is an example of how this is true that is easy for many people to see quickly. But basically, the answer is yes, it inherently does. The GUI is generally considered to be more cognitively accessible than command line interfaces because it allows "direct manipulation" of what is affected by user actions. There is a direct parallel to this in "direct qualification," the expression of properties in_situ at the point where they apply. This is logically more direct, and it takes both mental discipline and computational effort to do the same thing in a more indirect, rule-based way. The indirection can have real benefits, but it definitely has real costs, too. I think I recall Tim Berners-Lee talking about a principle of "lowest-level language" but I am failing to find it in the "architectural design notes" section of the W3C site. It is a little like Occam's Razor. We do have the line attributed to Einstein: Everything should be as simple as possible -- but no more." The "principle of lowest level language" treats abstraction as complexity, and says "What people will use in the end is the medium that gets the message across with the least amount of abstraction overhead." Of course, in my work I am constantly up against the fact that, like left-brained and right-brained people, people differ with regard to the extent that they find abstraction adds or removes complexity. Consider the other Einstein quote about "Nothing is so practical as a good theory." Science and technology is rife with examples where patterns or rules simplify the world around us and make it easier to understand and manipulate. The rule-based approach comes naturally to those of us who actually got the message in high school plane geometry and mastered proofs. But it is more indirect and more demanding on the infrastructure. We need some of that indirection to achieve the abstraction required to retain flexibility in what the User Agent does with the received data. Graceful transformation requires an infrastructure of encoded abstract knowledge. Getting the abstraction in the encoding requires some indirection, the best we know how to build languages and formats today. But we must not assume it comes free. At least in terms of where PF is dealing in the halls of the XML working group, every ounce of indirection has to be justified. There is a lot of discussion on this point that is buried in Member private space from the PF consideration of the XSL FO document (technology). One of the things we have been failing to do is to connect that discussion with the publicly visible record of the WAI. We don't have a Note on "The access perils and safety net of XSL FO" to go with the more upbeat Notes on CSS, SVG, and SMIL. Al >Regards, >Davey Leslie >-- >"Keep your monkey up!" >Davey Leslie >davey@inx-jp.org >
Received on Sunday, 17 December 2000 11:16:52 UTC