How Soon We Forget
Huh, apparently O’Reilly are continuing their rich tradition of 100% buzzword-compliance by sponsoring an “Ajax Summit“.
I was struck by two passages in the write-up of the summit on the O’Reilly Network site. First:
The summit saw little or no interest in any sort of Ajax standard. [Adaptive Path’s Jesse James] Garrett noted that “the consensus of the group was that there are too many people trying to solve too many problems in too many environments for there to be one standard.” As such, he sees no universal Ajax toolkit forthcoming, possibly ever. “The variety and needs of Ajax developers are too diverse.”
Yeah! Ajax developers are wild, man! Standards? They don’t need no stinking standards!
… followed shortly by:
Back on the wider web, Ajax is being driven by a desire for cross-browser compatibility.
(snort)
By running away from the one way we’ve found so far that actually works to encourage cross-browser compatibility — open standards, openly arrived at — Ajax is fostering cross-browser compatibility?
Riiiiight. That must be why you have to write ’90s-style browser-sniffing JavaScript at the beginning of every Ajax app, so you can tell which XMLHTTPRequest invocation method to use — because browser-sniffing JavaScripts encourage cross-browser compatibility!
The future is bright — it’s just the people that are dim, I guess.
A web feed for this site is available,
Comments
quinn
May 14, 2005
12:48 pm
Your point is valid, and something I think people should be considering.
I can also see why Jesse James Garrett and other people I’ve talked to about this don’t see it as something that makes sense as a standard. Ajax really is a bunch of pieces that have their own standards and a methodology for using them. But it’s a powerful methodology.
I think the danger comes when the people in the Ajax community talk about going to the browser makers to extend some part of Ajax. They need to keep in mind that they might at that point be asking that browser maker to break a standard in a way that might encourage the bad kind of competition for the market of Ajaxian apps.
One of the things I really appreciated about the people I talked to writing this piece is how forthcoming they were about all the problems, and not just this one. Garrett, for instance, seems very realistic about what this is and what its downsides are, which impressed me. He brought up the UI problem inherent in Ajax apps without my asking. Not that many people I talk to about their babies manage this, it gives me hope that the market in general has actually matured past where we were in the 90s.
But like you, I’m not 100% sure. I really want Ajax to be all that. I personally want the platform independance that would give me.
Jason Lefkowitz
May 14, 2005
1:27 pm
Maybe the way to square the circle on this is thusly:
We don’t need a standard for “Ajax”.
We DO need a standard for XMLHTTPRequest.
And we DO need someone to articulate how the browser-specific extensions that “Ajax” as currently articulated is dependent on are going to be streamlined into the rest of the world. If I’m developing a browser right now, how do I implement “Ajax”? The IE way? The Safari/Mozilla way? Some _other_ way?
We could let the market sort that out, in which case Microsoft’s monopoly will leave us with another gimped implementation — or we could put this question front and center and get it into the standards pipeline now.
As I didn’t attend the summit, I can only go on what I’ve read; and my sense from that is that Garrett et. al. don’t really see this as an issue on the same level as Kewl Design Tricks. I’m sure Netscape thought it could survive on cool tricks forever too…
Alex Russell
May 14, 2005
9:41 pm
Hiya,
As one of the attendees of the “summit”, I think your criticism might be a bit off the mark. Given the number of people w/ competing interests who were there, I think what you were seeing was that no one wanted to say “oh, yeah, guess I’m gonna have to give up my baby…”. A lot of people are also thinking too small (IMO) and that limits their ability to think about a generic or “baseline” set of tools.
That said, standards can and should only happen when there are good implementations to standardize. We’re working on a good implementation with Dojo, but I would expect standardization to lag deployment significantly. That’s how good standards get written, after all.