xml-dev - Re: New internet draft on layering protocols on top of HTTP

["Simon St.Laurent" <simonstl@simonstl.com>]

At 08:08 PM 5/8/00 -0400, Mark Baker wrote:
>> While it may affect protocols going through IETF process, it's not clear
>> (yet) that the current round of XML+HTTP protocols is going to be going
>> that route.
>
>Give it time.  The W3C is *not* a place to do protocols, while the IETF
>is *the* place to do them.  Seems like a no brainer to me.

The no-brainer for me is that many protocols may not rate a trip through
the IETF or any other organization, and that ad hoc is likely to prevail
over organized and even necessarily coherent.

(XML already has enough internal interoperability problems that I fear
we're used to them by now...)

>"Some URIs that purport to be HTTP are now complex listeners and
>responders, no 
>longer accepting 'ordinary' HTTP requests and modifying or ignoring the 
>HTML forms-based GET and POST approaches to transmitting information
>from 
>client to server."
>
>That's a Very Bad Thing, don't you agree?  If people want to reuse HTTP
>the protocol without thinking about the consequences to
>interoperability, they should be asked to use a port other than 80. 
>Ditto for the URI scheme, since expectations are that any HTTP client
>can GET any HTTP URI.  The alternative keeps me awake at night, worrying
>that the first vendor to release the understands-everything-on-port-80
>"web server", wins.

I didn't say it was a very bad thing, and I don't actually think it's a bad
thing.  I don't stay up at night worrying whether a particular URL is good
for a particular task - part of what we've learned from the Web is that
some level of 404 Not Found is okay.

>FWIW, I thought the "-xml" hack was a brilliant piece of engineering
>design.  Thanks for your work on it.

I'm glad _someone_ liked it.  I'd call it hackwork, not 'engineering
design', but it's a good way to adapt two-part MIME content-types for a
small set of problems that need more parts.

>- MIME has as much trouble integrating with "Web technology" as XML does
>(eg. CRLF vs LF).  It's heritage is email.
>- bigger and better things await namespace URIs

and both of those things worry me...  MIME and URIs may need to change to
accomodate such issues, as may HTTP - none of which will be easy.

Simon St.Laurent
XML Elements of Style / XML: A Primer, 2nd Ed.
Building XML Applications
Inside XML DTDs: Scientific and Technical
Cookies / Sharing Bandwidth
http://www.simonstl.com

***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************