Lists Home |
Date Index |
> Right. The important thing is to recognize that we *agree* how to
> process things, how to interpret them and in some applications, what
> the various tags "mean". Nothing intrinsic in XML provides anything
> here. GI's, even with namespaces, only work by application convention.
> So, to bring this all back in a circle. I dislike namespaces (or more
> precisely, the standardization of namespaces) because they complicate
> things (look at all the W3C specs) and *don't give you anything you
> couldn't already do without the standardization*. You and I could have
> agreed to use gi prefixes for exchanging data, and got along equally
> as well.
Sure. Not having been a participant in the XML namespaces debates of
1997-1998 (pre 1997?), my reaction is perhaps different.
Suppose we forget prefixes and attributes, then clearly, xmlns can be
treated as just a plain 'ole attribute.
<foo gavinns="org.example" />
and clearly some sort of AF processor might rename this to
regarding unique names, and preventing clashes, DTDs get around this problem
with entity names purely by using somewhat unique prefixes. the binding of
prefix to URI is not necessary, at all, for the stated goal of creating
unique names, particularly locally unique names.
at the end of the day (actually it is friday morning, so i reserve the right
to change my mind tonight :-)) the xml namespaces _convention_ has been
widely adopted. I completely agree that there is nothing unique, and we
could have agreed on a different convention, but we (the we of the larger
xml community) has essentially agreed on this convention, as judged purely
by the number of W3C and non-W3C specifications, software programs (e.g.
SAX2, IE5 etc) and instance documents that use namespaces.
I think there is great value in conventions such as XML namespaces, and the
value of cooperation outweighs any deficiencies, and perhaps even any
complexity, introduced as a way of getting different people to sit down at
the same table.
A misconception that many technology folks have (and I am specifically not
addressing this to you in person, or to anyone in particular) is that the
technology exists to be pure, or elegant, and that code exists to be
efficient. The technology exists for what it _does_, for the benefits that
it delivers to those who use it. The benefits of cooperation outweigh
elegance of syntax. This is particularly true of XML. S-expressions which
have existed at least since the 1960s really are more elegant, compact etc.
than XML, but so what?