[
Lists Home |
Date Index |
Thread Index
]
At 03:06 PM 7/25/2002 -0600, Uche Ogbuji wrote:
> > For a brief description of the kinds of problems I've heard in the field
> > (at conferences and an ACM tutorial), see:
> > http://lists.xml.org/archives/xml-dev/200207/msg00965.html
>
> >From my POV, Tim's point comes out unscathed.
>
>To the extent that I understand your three candidates for interop
>problems (and I think I only vaguely understand them), I don't see how any of
>them are code-level implementation problems. I also haven't really seen such
>problems in practice, though I must admit that the various Python XML
>developers are a relatively close bunch and misunderstandings get hashed out
>rapidly.
It's not an implementation problem as in "Namespaces are hard to process in
various XML tools." I've written a fair number of those tools myself, and
the scope rules are a lot trickier at that level than the URIness.
It's an implementation problem as in "George thought the URI meant XYZ so
did YTQ, Gina thought the URI meant ABC so did QTM, and Joseph just ignored
the damn thing and relied on checking prefixes since God knows it was
easier and it worked."
Keeping the developers building generic XML tools on top of these things in
line is a much smaller problem than keeping programs who build their own
programs on top of those tools from believing that their assumptions are
widely shared. URIness is a key aspect of that confusion.
"If you're explaining, you're losing." [1] We're doing a lot of explaining.
[1] - JC Watts (Republican Congressman from Oklahoma) on why he's retiring
from politics, as quoted by Larry Lessig in his OSCON keynote yesterday.
Simon St.Laurent
"Every day in every way I'm getting better and better." - Emile Coue
|
