Lists Home |
Date Index |
- From: Sean McGrath <email@example.com>
- To: firstname.lastname@example.org
- Date: Sat, 30 Dec 2000 11:56:54 +0000
At 10:01 PM 12/29/2000 -0500, Simon St.Laurent wrote:
>As one of the poor schmucks who has the lucky job of explaining these
>issues to the world, I'm afraid it's my experience that this spec has sowed
>far more confusion than production-quality XML-centric software, and far
>more confusion than any of its peers. Your so-called success has been my
>misery - in books, in discussions, at conferences.
I feel your pain:-)
>For developers attempting to use namespaces, though, the muck has only
>grown deeper and deeper and deeper and deeper. Nearly two years later, we
>still don't have consensus on what namespaces are, or how best to use them.
The namespaces rec suffers from two things it seems to me:-
1) It flashes its eyebrows at a problem it does not solve
2) It is neutral w.r.t. the key issue of how to use the namespaces
it allows you to create.
I'm reminded of the short section of the SGML spec. (7.6.1?) which
flashed its eyebrows at the record end handling problem and confused
an entire generation of developers and implementors.
Unique identification is the prelude to shared semantics.
The two are so closely related that Joe Developer expects
that they would be addressed in the same spec. Furthermore,
Joe Developer sees a spec. on unique identification and consequently
expects the authors to have something to say on the substantive
issue of semantics to benefit from the uniquely identified components.
The namespace rec. addresses the former and then bewilders
a generation by going out of its way to say nothing about the latter.
The namespace rec. is short because it is only chapter one of
a greater spec. Perhaps publishing it so far in advance of chapter two was
As I have said before, perhaps it plays into
the hands of those who know that the real "lock in" lies
in the semantics which now, in the full glare of standards
compliance, they can make proprietary.