[
Lists Home |
Date Index |
Thread Index
]
> Huh? Grammars aren't good at expressing position independence? I'm
> having troubling thinking of any grammars for any language (not just
> XML) that don't handle constraints similar to the one that the poster
> describes.
It's common to have such constraints in a language, but they aren't
generally part of the grammar.
For example, a grammar can say that you can have zero or more attributes,
but it can't express the constraint that their names must be distinct.
Similarly, a constraint that you can have any sequence of A, B, and C
elements provided that there's at least one A and not more than three Cs is
not a grammatical constraint.
> XML Schema was a classic case of design by committee with individual
> features grafted onto it by the different participants.
Ad hominem arguments always weaken your case. In this case, one could argue
the opposite criticism: that the group was too wedded to the conceptual
integrity of the theoretical framework they had chosen (namely, to design a
grammar-based constraint language), and not flexible enough to bend the
rules to meet practical user requirements.
Michael Kay
http://www.saxonica.com/
|
