Lists Home |
Date Index |
On 7/6/06, Bullard, Claude L (Len) <email@example.com> wrote:
The damming thing is, that is what the web does, not best or
> worst, but simply does at all.
Ok Len, the best technologies don't win, the winning technologies win.
Worse is better because that is all it
> is capable of.
I'm prompted to respond because the "worse is better" meme irritates
the heck out of me. Damnit the meme meme (or maybe just Richard
Dawkins) irritates the heck out of me. There's no doubt Darwinian
natural selection applies to critters, ideas, technologies. But the
environments for each vary widely. Understatement.
XML Schema and Relax NG maybe are examples of winner and better. But
the notion of winning implies a time of judgement - is the Web done
yet? At this point in time there are certainly places where XML Schema
looks a hobbled horse (e.g. Atom) and even places where one might
wonder whether Relax NG might be up to another lap (e.g.
microformats). Conversely there are places where each seems fittest
for the job. But the *general* issues and applicability of such things
only seems to happen as some kind of post factum lazy evaluation. Look
back in LX anger.
Perhaps this suggests there's something inherently flawed about
current notions of schema languages. Does syntax validation really
offer much for the vast majority of comms on the Web? Sure, I want my
bank interactions to fail on errors. (Truth be known I'm probably in
the Draconian/anal datahead camp. But I don't see a conflict between
that and pragmatism). But I only visit the bank once every few weeks,
I'm on the Web most hours of the day. There's pretty obvious value in
tight schemas in most *local* contexts, docs or data. But go big
distributed, how important is their role?
I dunno, probably most folks around here would have anticipated the
shift in popularity towards Turtle/N3 syntax from RDF/XML, had the
former been on the racing card. RDF schemas still works with that, but
offers no real notion of validity (unless you bring in DL consistency,
but that's another story). But what of JSON? Dark horse (to cliche out
We live and learn. Hobby horse de jour - Atom is a bugfix on the
first-fence pileup that is "simple" RSS. But neither would have been
necessary had a little more time been spent around metadata parts of
HTML. Poll HTML, why not? (c.f. hAtom microformat). Folksy tags also
had a chance as HTML keywords. What changed? Not the tech per se, just
the enviroment (including technologies, ideas and maybe critters).
Maybe on the Web natural selection affects the macro more than the
Whatever, my money's on the winner (TBD).