- From: Henrik Frystyk Nielsen <frystyk@w3.org>
- Date: Mon, 19 May 1997 14:56:30 -0400
- To: Ross Patterson <Ross_Patterson@ns.reston.vmd.sterling.com>, http-wg@cuckoo.hpl.hp.com, confctrl@isi.edu
At 10:30 AM 5/17/97 EDT, Ross Patterson wrote: >Henning Schulzrinne <schulzrinne@cs.columbia.edu> writes: > >> I seem to remember from the PEP >>draft that the URLs serve to (a) identify (b) describe the extensions. I >>don't see how downloadable extensions would work, given the diversity of >>servers, platforms, languages, etc. > >I agree, although Java and the "servlet" movement might change all that. >But frankly, I'm more concerned about the "nonstandard" part of my >"nonstandard downloadable" comment. I see PEP as primarily advancing >the Balkanization of HTTP that began when the major browsers started to >deploy unilateral and (at the time) unpublished extensions to both HTTP >and HTML. Experimentation is good for interoperability, but variant >"standards" are not. This is actually what I would call an argument _for_ PEP and not against ;-) What is happening now is that more and more applications become dynamically extendible but there is no mechanism for expressing this on the wire. If the only mechanism to describe dynamically extensible applications are using static specifications then we get the tension that we see now in HTTP. Note, that it is not a "either or" situation between PEP and new versions of HTTP, RTSP, etc. PEP extensions which become popular can get integrated into the base protocol as they evolve and become ubiquitous. >Nothing here requires that a server go obtain the extension and run it, >but market pressure will inevitably lead there, especially the first >time either Netscape or Microsoft publish a downloadable extension. >Just look at the Netscape plug-in market - it's what HTTP will wind up >looking like after that. The problem of trust and which extensions you can download is real but is not particular to PEP - it's the same problem every time you download something onto your computer. Thanks, Henrik -- Henrik Frystyk Nielsen, <frystyk@w3.org> World Wide Web Consortium http://www.w3.org/People/Frystyk
Received on Monday, 19 May 1997 11:59:16 UTC