Re: [jats-list] JATS 1.1 request: kwd-group in more locations

Subject: Re: [jats-list] JATS 1.1 request: kwd-group in more locations
From: Wendell Piez <wapiez@xxxxxxxxxxxxxxx>
Date: Tue, 23 Oct 2012 09:03:12 -0400
Hi,

On Mon, Oct 22, 2012 at 5:15 PM, Maloney, Christopher (NIH/NLM/NCBI)
[C] <maloneyc@xxxxxxxxxxxxxxxx> wrote:
>> > I think we're already partway in that direction. The fig and table-wrap
>> > models both already allow attrib and permissions. So if we were to bundle
>> > them now, it would be a non-backward compatible change.
>>
>> Yes, I saw that today ... boo.
>
> You could add <sec-meta>, and then deprecate the use of <permissions> inside those
> elements for new document instances.  But I don't know if there's an effective way
> to communicate that something valid according to the DTD is deprecated.  It does
> bring to mind the DtdAnalyzer, where you could put it into a schematron section
> of a structured comment, but that's not in widespread use (yet), so I don't know if
> that would be "effective".
>
> This discussion also makes me wonder why <permissions> is included in <sec-meta>,
> but <attrib> is not.

I don't think there's a problem in general with a model's growing
"organically"; quite the contrary. A model that has evolved in
response to actual practice and need is usually better than a model
that has been imposed by someone's notion of clear and consistent.
(Even mine. :-)

So I don't want to give the impression that I think there's something
wrong with the current design, even if my earlier idea of a simple and
clean yet more general extension won't work for reasons I hadn't
thought of. (Happens all the time.)

That being said, while it's true that things (in some other imaginable
world) could be simpler and more consistent, a strategy of extending
backwards-compatibly, and then deprecating features, is not simpler --
most especially when you end up pushing normative strategies for
dealing with the resulting complexity (which everyone then has to
assimilate and then, maybe, vary).

In other words, I think this is probably a case where simplicity and
usability trumps "correctness".

Cheers,
Wendell

-- 
Wendell Piez | http://www.wendellpiez.com
XML | XSLT | electronic publishing
Eat Your Vegetables
_____oo_________o_o___ooooo____ooooooo_^

Current Thread