Hi Ken,

2009/3/23 G. Ken Holman <[email protected]>:
> Thank you for supplying so many examples, as that made testing a result
> easier.

Thanks for taking the time to come up with such a detailed answer and examples.

>>The problem is that I think this requires the group containing type
>>and card_data to be in order;
>
> It does not. You can interleave with an interleaved group.

I guess my confusion stemmed from making an implicit bad assumption. I
ran your schema through trang to get a .rng version and everything
became immediately clear.

> p.s. I'm not sure why you are making the input so free-form. Back in
> the SGML days I believe we (the community of SGML users and document
> designers) ended up coming to the conclusion that if the order of
> elements is not important, impose one on the users so that they don't
> somehow get the misinterpretation that order is important. If they
> are allowed to choose they order, they might think this gives them
> some control over the meaning. So if there is a user-controlled
> semantic by their choice of order, then by all means give them that
> choice. On the other hand, if there is no order to the information,
> then give the users a single order to enter it so they understand
> they can't assume order has any importance. Just a thought for you
> to consider.

I agree, and I'm going to copy-paste this paragraph to support my opinion. :-)

Dusty

Reply via email to