[
Lists Home |
Date Index |
Thread Index
]
> Well, we [Objective Systems]have implemented "SAX-like"
> which I guess is what you experts are now calling "SAD".
> Is that good enough for "proof of concept"?
This is certainly a first step on the way to providing a
real SAX implementation. Let's hope that you guys take the
next step and do full and correct SAX 2 (even if you feel
that it is technically less efficient or less "good" in some
way.). A partial implementation, or even a "SAD" interface
has value on its own, however, it is of limited utility to
those of use who want our binary streams to work properly and
well with XML focused tools.
It is possible to build a SAX interface to binary
encodings in such a way that no one and no code can "smell"
the difference between that implementation and one which was
written to parse XML textual data. Such an "odorless"
implementation is what we really need to work well in this
environment. Sure, the implementation might provide
additional features, etc. (i.e. SAD stuff) However, the SAX
stuff should be implemented with religious attention to
correctness and conformance.
bob wyman
|
