Personally, I manage this frustrating state of affairs by spending
time/energy on efforts like EXSLT, EXPath, EXQuery (and I hope EXProc
in the future). These efforts inevitably are less rigorous but occur
on shorter timescales, this is especially important in our field ....
10 years is far too long for any standard to be generated, if only for
the fact that advances in hardware will change a lot of underlying
assumptions during these timescales making it doubly difficult to
navigate a standard to some useful outcome.
Also I have found that these 'EX' type ancillary standards provide
valuable feedback to 'official' standard bodies.
I don't see any barrier (other then the familiar constraints of
time/effort) where we could spec out extensions to the browser and
request/build implementations in Firefox, Webkit and Chrome.
Perhaps I am being naive, but I think this approach would be an order
of magnitude shorter then trying to change existing constraints with
existing standard orgs.
Yes, speaking at least from our shared experience with EXSLT, that got pretty good uptake from browser vendors, among others, and this is despite the fact that we never were really able to get our act together and polish all those spec modules. But I think the key was in keeping things simple and modular (that word "tractable" again), and focusing on the most urgent needs first (e.g. exsl:node-set). And that was my entire argument against the huge, monolithic specs that were being developed at the time.
As to Mike's point about just having the browser vendors provide the hooks for implementation in whatever language, regardless of security problems, wouldn't that require users to install a plug-in? Isn't that well demonstrated to be impractical, except perhaps in Intranet?
