Re: Comments on draft-ietf-conneg-media-features-02.txt

[koen@xxxxxxxxxx (Koen Holtman)]

Larry Masinter:
>
>I think the issues of 'text only' content negotiation are sufficiently
>complex and affect a sufficiently different review article that 
>'textonly' features should be handled in a separate media feature document.

Hmm, I think that would be overkill.  I think I'll just register my
textonly tag when the registry is open.  I am interested in the use
case of labling the version best suited for lynx/speech/braille users,
in the limited scenario of a (web) server that has a few static
versions available.  I am not interested in writing a document that
provides tags for the server-side fine-tuning of rendering on
terminals, braille devices etc.  I think the internet community needs
a textonly tag on a short term basis, while there is no big need for
server-side fine-tuning of textonly content.

This still leaves us with my observation that the draft would be
better as a style guide if a boolean tag were included.  Any ideas on
less controversial boolean tags?

>I've always assumed 'rational' meant 'expressed as a ratio' and suggest
>that the registry draft be updated to make that determination. Even if you'd
>prefer decimals, 256/100 == 2.56 easily enough.

Let me put it this way: I think the registration draft either needs to
define a standard (ASCII) representation for the rational type, or else
it needs to declare explicitely that the choice of representation is
left to the individual negotiation protocols using such tags.  I would
prefer if a standard representation were defined, as this is better
for overall interoperability.

Question to the chair: I see that the -registration- draft is now in
IESG last call.  Do you want me to announce/forward the above comment
in a more formal way?

>Larry

Koen.