Lists Home |
Date Index |
> I used to get quite irritated by the fashion for adjectival
> company names
> (e.g. "Rational"). But one gets used to things - fashions change.
Perhaps. I certainly can't swear that the gerund form for technologies won't
become a trend. Personally I doubt it, but this falls into the realm of pure
speculation that XML-Devers tend to frown on.
Just to make sure that I'm driving my perspective home, let me make this
more general point. In a standards efforts that I am participating in (no
prizes for guessing which), we had lengthy discussions about which schema
language to use. Frankly, there was a lot of discontent about making XSD our
official choice, for a number of reasons, especially the quite unaesthetic
syntax and the lack of consistent tool support (so you never know which
features are actually usable). It's quite widely felt that the spec tries to
do too much. (If any XSD advocates would like to comment on these points,
I'm sure this would be welcomed.)
RELAX NG was much preferred by those familar with it. Its syntax is a lot
clearer. It's also more powerful, allowing the calculation of schema
intersections and unions. Finally, because it tries to do so much less, it
should be a lot easier to support consistently across tools. However, it is
widely perceived as having lost the "marketing battle" with XSD, so people
are relunctant to support it. Furthermore, it isn't entirely clear what the
RELAX NG folks' attitude towards derivation is. (I'd be very interested to
hear an official statement on this.)
This leads me to three potential conclusions:
1) All this is a storm in a teacup. XSD is fine, will be subsetted de facto
and it isn't so bad anyway. RELAX NG is a nice academic idea, but it didn't
gain enough support quickly enough from those who matters and is doomed to
the margins forever.
2) RELAX NG is going strong, there is a significant and active push to make
it a viable competitor, not just from a technology perspective but also in
terms of market adoption. But I (and many others) just don't know about it.
3) There is a strong potential role for RELAX NG in the schema world, but no
one is taking the role of "chief evangelist" and spreading the good word.
I'm sceptical about 2), since if this was the case, by definition it would
be more widely known. If 3) is the case, then ideas like simplifying the
name belong firmly in the context of an overall marketing discussion. The
fact that no one except Michael took the bait when I raised this point seems
to imply that this isn't of widespread interest, hinting that 1) might be
the case. Or maybe this point has been debated ad nauseum, and people have
exhausted themselves on it.
Any comments? I would emphasize once again that this is a real issue of
current relevance that is facing groups who are doing schema design and need
to decide what their canonical language is.