Re: Form-based Device Input and Upload in HTML

James Salsman (jsalsman@corp.webtv.net)
Fri, 14 Nov 1997 21:02:47 -0800


Message-ID: <346D2CF7.733D@corp.webtv.net>
Date: Fri, 14 Nov 1997 21:02:47 -0800
From: James Salsman <jsalsman@corp.webtv.net>
To: Peter Flynn <pflynn@imbolc.ucc.ie>
CC: www-html@w3.org
Subject: Re: Form-based Device Input and Upload in HTML

Peter,

Thanks for your message:

>... File uploads were proposed ... for uploading files of
> static data ... of pre-known size, not for interactive 
> streaming....

I'm not proposing streaming at all; I think using HTTP 
for real-time transactions is a Bad Idea.  That is why 
the draft specifically suggests using RTP instead for 
real-time needs.  Activating streaming from a microphone 
with an arbitrary page of HTML is a serious security 
headache, probably enough to be illegal in California. 

The proposal is for buffering before the upload, just 
as <input type=text> buffers your keystrokes before 
uploading when you submit the form.  The MAXLENGTH 
attribute and the DEVICE-MAXLENGTH Pragma HTTP header 
are proposed to negotiate buffer length, in section 2 
of the draft. 

> I think the idea is fine, but I'd like to see input from 
> more people in the audio/video field.

I've recieved very positive comments from a gentleman at 
Lucent about the use of 'audio/L16;rate=11025' instead of 
'audio/basic' -- I'll forward all the private comments I 
receive and an updated draft (if any) to the W3C lists 
www-talk, www-html, and www-multimedia after the vote on 
HTML 4.0.  (As you could probably tell by the embargoed 
status, the proposal doesn't take effect until after the 
HTML 4.0 draft is approved.)

>... can I send again my long message about the WebTV 
> additional elements that I put into HTML Pro? 

Absolutly; please do.  The problems with our on-line 
documentation are understood and being worked on, and 
any help would be great.

Sincere regards, 
-- 
:James Salsman, WebTV Networks  
::<jsalsman@corp.webtv.net> 650-614-8465