Lists Home |
Date Index |
> In theory Xpath1 processors can drop comments, but in
> practice they don't. But Comments are inherantly second class
> citizens anyway. 0's in character data are not (or were not
> until recently) No Xpath processor is going to drop leading
> zeros from source data. If I decide, in my query to use that
> data as a number then obviously 01 = 1 but if I want to check
> then I know it will work.
> If I switch to Xpath 2 (and it's a big if) then I no longer
> get that assurance, only your "hope" that I'll find
> processors that don't mess up my data.
OK, let's be clear. There are some big companies who think that it's a
good idea to store XML in relational databases. This means shredding
your document, turning it into tuples, and then putting it back together
again on retrieval. If you do this, you are going to lose fidelity. If
you're using XML to represent simple tabular data, you probably don't
care. If you're storing documents such as legal contracts, then you
almost certainly do.
There are other companies (like mine) who think that shredding your
document is a lousy idea and that users will prefer to buy XML databases
that don't do it.
The standards game is about identifying where we can achieve commonality
and where we agree to differ. If the spec gives options that reduce
interoperability, you can be pretty sure that's because the alternative
was to have no spec. In the end, the market will decide what the
requirements are, and they may well go beyond mere conformance with the