- From: David Poehlman <poehlman1@home.com>
- Date: Sat, 14 Jul 2001 09:54:43 -0400
- To: "wai-ig list" <w3c-wai-ig@w3.org>
----- Original Message ----- From: "Louis Atkinson" <louis@HIGHERMIND.ORG> To: <EASI@MAELSTROM.STJOHNS.EDU> Sent: Saturday, July 14, 2001 3:14 AM Subject: Re: Web Access; When the Rubber Meets the Road ******Delurking***** Hi, Everyone As a multimedia artist, with formal training as a painter, I must admit that I enjoy the visual communications abilities that the web enables. That, in fact, was something that caused it to blossom: the ability to have inline graphics was a step above what gopher or bbs ascii sigs allowed. That said, I am also a long time web design professional, have been watching the list for months, have been working on making accessible sites, and think that I also have some points to make about enabling the easing of providing suitable computerized interfaces to a wide variety of human interfaces. I agree that the noscript javascript tag is most useful, I mostly use javascript as a tool to enable my own development efforts. I have found that one of the most abused functions of javascript is the onmouseover effect, which has been in many ways rendered moot (in IE5 at least, for text links) by the combination of a title in the href and the a:hover element in CSS. The main problems with javascript are that it isn't universally supported by clients and can be turned off. The fact that it is a powerful scripting language that executes *on the client-side* leads to not only accessibility issues, but security ones as well. >The principle of the lowest common denominator still applies; the goal is to >raise the denominator. I have been trying to partially answer that question, and think I have found a partial solution. I'm sure many have heard of it before, but it is to abstract the data from the presentation by keeping the data in a database or XML file (which is in many ways a text file-as -relational database). Then to get the client's USER-AGENT http header (with server-side scripting being preferable to javascript) and serve the data with whichever interface is relevant back to the client. This causes some headaches during the development cycle (such as only being able to serve CSS based sites to certain clients, text based to lynx and search engines), but will get easier, and shouldn't preclude anyone from being left out. My main problem with it at this point is that I do have a list of the variety of USER-AGENT strings used by screen reading and other browsers. I have made a page that will gather that data (again, server-side, not javascript), and which puts it into a database, which is located here: http://www.highermind.com/design/beta/user_agent/index.html so if anyone with screen reading or other non-visual software would care to go to that page, it will help create a list of different clients that can then be filtered into the appropriate directions for easier serving. I will leave this list of clients online for anyone to use, and I hope that you can help me. As a note, though, to anyone using Netscape or IE on a Mac, I am using a javascript to gather the plug-ins names, and if anyone could suggest a better way to do it, it would be greatly appreciated. Also, I am using open source tools available by many hosting companies to do this. Best regards, Louis Atkinson n-dimensional artist Higher Mind Productions http://www.highermind.com/
Received on Saturday, 14 July 2001 09:54:52 UTC