RE: Issues from Microsoft ISSUE #1 Move 1.2.3 to level 3?

I'll try to respond to these...

	The concern is this:  There is no way that I know of for a Web
author to add real-time captions himself.  He must use a professional
service, which can be quite expensive.   

This is all true.  Realtime captioning is neither cheap nor easy for web
authors to create.  

	 This is the only one of our guidelines that requires hiring an
outside professional.  Amateurs can caption, and even audio describe,
pre-recorded multi-media.  They may not do it as well as professionals,
but it is possible to do it.   

They definitely don't do it as well as professionals, particularly AD.

	 The live captioning requirement essentially puts level-2
compliance for live multi-media out of the reach of all but the largest
of media companies.   Is this really what we want to do?

This is a bit of a stretch.  Within the US, there is a service that is
run by caption colorado that does realtime captioning via the web
(http://www.captionedtext.com/Embedded.aspx).  This service is free for
federal employees (http://www.fedrcc.us/fedrcc/).  Other groups,
including NCAM with captionkeeper (http://www.captionkeeper.org) and
WebAIM have created tools to present caption data acquired from a
stenographer.  

	Even if we say that requiring a service is ok (and I'm not at
all sure that it is), the technology for including that service is not
described well in the techniques.  How do you integrate the captions
into your site?  Can you generate the SMIL that's in the techniques in
real time?   

If this is a smil presentation, the smil would be static and it would
reference a URI where the captions were streaming from.

	 How?  Gregg mentioned putting the text coming from the
professional service in another window.  How does that work?  What
technologies are used?   

There's one example at the captionedtext.com link above.  There is no
standard way to deliver caption data, offline or realtime, so each
player does it differently.  Windows Media can deliver realtime captions
too. 

	 If you had the text before the live broadcast, is it possible
to synchronize it with the broadcast?  If so, how?   

There are services that will automatically synchronize a transcript with
the audio, but I'm not sure that this can happen live so much as
"near-live". These technologies aren't perfect yet, but they are used in
production settings currently.

	 Is it possible to convert captions from television content to
text that's usable on a web site?  How?  

Line 21 captions are not in a convenient XML format, ready for XSLT, but
there are tools that will do this.  NCAM has been doing this for at
least 5 years.  The technologies aren't common or widely used, but that
may be a chicken/egg thing.
 
Hope this helps,
AWK