Lists Home |
Date Index |
Elliotte Rusty Harold wrote:
> Perhaps. It seems you're arguing (and please correct me if I'm wrong)
> for a binxml that has more information than a plain vanilla XML
> document. Your vision of binxml includes type information. That is, it
> knows what's a string and what's an int and what's a date and what's
> binary data and so forth. Is that accurate?
Half-right, half-wrong :)
In the case I quote, yes indeed in order to produce the binary infoset
corresponding to the description made in the article one would need to have
access to type information. Note that I was only providing an interpretation of
the text in the article, sorry if it was insufficiently detailed.
My vision of binary infosets includes using as much metadata as possible to
obtain more efficient compression. In the presence of a PSVI, the product I work
on will make use of the typing metadata to better encode its content. In the
presence of a vanilla XML Infoset, it'll encode it as such. In other words, the
tool is open to what you want to do with it.
> If so, it seems that your BinXML is really not BinXML at all. Rather
> it's a BinPSVI or something like that.
You will certainly have followed this permathread long enough to notice that
I've stated a dozen times that "Binary XML" and all variations on it are
language abuses and that "Binary Infosets" is imho much more appropriate. The
PSVI is one kind of Infoset.
Robin Berjon <email@example.com>
Research Engineer, Expway http://expway.fr/
7FC0 6F5F D864 EFB8 08CE 8E74 58E6 D5DB 4889 2488