Thread Subject: Re: Second Life
This archival content is maintained by WebAIM and NCDAE on behalf of TEITAC and the U.S. Access Board . Additional details on the updates to section 508 and section 255 can be found at the Access Board web site.
From: Sean Hayes
Date: Fri, Aug 03 2007 3:20 AM
- Return to this mailing list's archives
- View all messages in this thread
- Next message in thread: Jim Tobias: "Re: Second Life"
- Previous message in thread: Gregg Vanderheiden: "Re: Second Life"
- Messages sorted by: Author | Thread | Date
Second life apparently uses XMLRPC and UDP as its network protocols . XMLRPC does run over HTTP but this is mainly for setup (logging in etc), the actual model rendering protocol appears to be UDP based.
I share Peter's concern over using HTTP or indeed any other specific technology as the defining attribute of a class of software for 508, as it creates too many gray areas.
Outlook is in one mode by this definition a web application, as it is able to communicate with the Exchange server using HTTP tunnelling, yet in a second mode it is not. Other network connected applications use similar tunnelling. Parts of many sites use RTP or UDP, FTP or other network protocols.
This whole Web/Software/Content debate is extremely problematic.
Firstly we have to try and distinguish between applications and data, but this is difficult because software is essentially "data in motion". Whether data is "in motion" can be a judgement call, for example Spreadsheets, 3D models, Presentations and so on can all be 'put in motion' in ways which make them appear and behave like software. Lisp is a classic example of a system where the difference between data and program is a point of view.
Besides that debate, we also have some defining attributes of software:
On one dimension we have
1 - Client only
2 - Client Server
We can distinguish sub categories of client server like:
2.1 - Connected, where the application communicates with server applications for data, but the application itself is wholly running on the client (it may have been downloaded before executing)
2.2 - Paged, where the application itself is loaded on demand from a network in small pieces.
(Another type of connected is distributed, where the application itself is partly running on a server and partly on a client. The distinction between distributed and connected is fairly subtle though)
Another dimension we have discussed is:
A - Platform hosted
B - Non hosted.
Platform hosted means there is a runtime layer isolating the application from an underlying platform, non-hosted means no such isolation is happening (although the boundaries of this get a bit fuzzy when you consider an operating system)
A Web browser is typically Connected, Non hosted (but examples of other types exist), the data rendered within it is historically paged, platform hosted, but with "Web 2.0" is increasingly becoming connected, platform hosted. A web browser is capable of putting web data "in motion", by suitably interpreting form elements, script and CSS. But simple text based html is not really an application.
Second life is connected, non-hosted I believe (although it could equally be other); the data within second life is not generally considered software, in the sense that the client is really presenting a view of something being calculated on the server. One doesn't typically think of the 3D locations in second life as applications although you could certainly argue that they are 'data in motion', and I believe it is possible to upload and run 'automata' scripts within the second life server as part of a 3d model.
An email program is again typically connected, non-hosted (but could be 2.1A, 2.2A, or even 2.2 B); the content within it is typically regarded as just data, but examples exist where the email is 'put in motion' by the email client. Word processing, Paint and Draw type programs also typically regard data as passive, but often have features which are able to put their data 'in motion'.
I very much doubt that we will be able to come up with clean dividing lines for all this stuff within the 508 recommendation. What we need to strive for is to create rules that no matter how data is being presented, it can be utilized directly or adapted by AT for PWD. I think focusing on the user need, rather than the product characteristic is the way to organise the regulation; with individual clauses containing any necessary conditionals limiting their scope.
Accessibility Business Unit
mob +44 7977 455002
office +44 117 9719730