Re: XMLP Issue 450 resolution

Hello xml-dist-app,

> With regards to Content-Transfer-Encoding, we believe that the criteria
> for its use are already well-established by the MIME RFCs, and do not 
> need amplification by our specification.

We agree that this feature is already well specified, and merely
thought that it should be clarified that this feature should be used
here.

> With regards to Content-Length, we point out that MIME already has 
> well-specified mechanisms for delimiting message parts, and although 
> length delimitation is attractive, it is not within the scope of the 
> Working Group to invent a new MIME packaging mechanism (use of 
> Content-Length in SIP, as in HTTP, is specific to that protocol).

We agree that length delimitation is attractive - indeed, essential,
if the object is to embed arbitrary binary data.

The alternative of a random separator string is computationally
unattractive if the software verifies that it is indeed unique, and
relying on statistical likelihood without verification was described
by some as "a fragile hack".

-- 
 Chris Lilley                    mailto:[email protected]
 Chair, W3C SVG Working Group
 Member, W3C Technical Architecture Group