[
Lists Home |
Date Index |
Thread Index
]
> It is now crystal-clear that allowing qnames to escape from element &
> attribute names into content was a terrible mistake that we're now
> stuck with forever. I struggled against this idea but lost. -Tim
I struggled against pretty well the whole namespaces REC and lost.
I actually remember a stage when I thought "it doesn't matter that this
spec is bad, because it's so bad that no-one will ever use it". But the
lesson from history is that no spec is ever so bad as to ensure its
failure, any more than any spec is ever good enough to guarantee its
success.
And it's bad for exactly the reason mentioned in the blog - it's not
properly layered. You can't define the fundamental naming and addressing
mechanism as an application layer on top of XML and expect it to work.
Michael Kay
|
