- From: Marcos Caceres <marcosc@opera.com>
- Date: Sun, 20 Dec 2009 21:42:35 +0100
- To: Alex Russell <alex@dojotoolkit.org>
- Cc: Maciej Stachowiak <mjs@apple.com>, "Klotz, Leigh" <Leigh.Klotz@xerox.com>, Jonas Sicking <jonas@sicking.cc>, Boris Zbarsky <bzbarsky@mit.edu>, WebApps WG <public-webapps@w3.org>, Forms WG <public-forms@w3.org>
On Sat, Dec 19, 2009 at 4:01 AM, Alex Russell <alex@dojotoolkit.org> wrote: > > On Dec 18, 2009, at 3:09 PM, Marcos Caceres wrote: > >> On Fri, Dec 18, 2009 at 4:40 AM, Maciej Stachowiak <mjs@apple.com> wrote: >>> >>> On Dec 17, 2009, at 3:15 PM, Klotz, Leigh wrote: >>> >>>> OK, so is the conclusion that XHR is implementable only in HTML5 and >>>> should be re-titled "XMLHttpRequest in HTML5" or something similar? >>> >>> I think your premise is false, and I don't such a retitling would be >>> helpful. The XHR spec does not require a full implementation of HTML5. It >>> only references some concepts from HTML5. The XHR spec could be >>> implemented >>> in an SVG or HTML4 or XHTML 1.0 implementation that did not support most >>> aspects of HTML5 at all, as long as it could satisfy the requirements >>> implied by those definitions. Your proposed title change would imply that >>> the XHR spec could only be implemented by an HTML5 UA, but that is not >>> accurate. >>> >> >> So, basically, what you are saying is that you can't pick up this spec >> and, say, implement it in [insert favorite programming language] >> easily without a whole bunch of baggage from HTML5? Seems like pretty >> poor engineering, but that might not be the fault of the specification >> (i.e., given that XHR is a reverse engineering of something that is >> closely tied to browsers). Its a shame though that we can't liberate >> these things from browser behavior so they are more generally >> applicable. I've seen XHR implemented in other classes of product, but >> it'd be a shame if such products can't ever conform to the spec. > > That's sort of a perverse way to look at it. It's not like XHR is a *good* > API. It's a passible Win32 COM interface, but you'd want a lot more control > over many aspects of the HTTP discussion if you were doing this in an > environment that's not a browser. What other environment has a same-origin > policy? Unless the other language you're talking about is C++, I don't think > anyone should be toting XHR around with them like it's some sort of a > liberated gem. It's a bad JS API and would be as bad (or worse) in many > other dynamic languages. Yeah, you are right. I guess we get so used to having these crappy retrospective APIs around that one forgets that things could be done in better ways - thankfully decent frameworks have been built around them to make these things usable. -- Marcos Caceres http://datadriven.com.au
Received on Sunday, 20 December 2009 20:43:30 UTC