[
Lists Home |
Date Index |
Thread Index
]
> The argument in the thread is whether supporting entities but not using
> them must cause show-stopping space or time degradation compared to
> not supporting entities, as was claimed. (And, more than that, the
> claim that supporting internal entities requires multiple buffers.)
My interpretation is slightly different. The thread started with the
discussion of how SOAP disallows DTD's. If DTD's and their internal
entities were allowed, then they must be supported. I was saying that if
you didn't have to have entities, then you could make a small, fast, and
efficient implementation. You're saying that if you don't *use* entities,
then no penalty. (Did I get that right?) Those are similar, but not the
same. If the spec doesn't disallow entities, then your code must handle
them. If the spec dosen't allow them, then your code can be smaller and
faster.
I guess it's like "not(not(a)) != a".
If I haven't made my point, I think I've at least argued it into the
ground.
/r$
|
