On Tue, 2004-10-26 at 01:00, Josà AbÃlio Oliveira Matos wrote: > On Mon, Oct 25, 2004 at 05:15:39PM +0300, Martin Vermeer wrote: > > > > I still feel you're making this too complicated :-) > > Mais au contraire, mon cher ami. ;-) > > [This started as an simple answer and ended as a rant. You have been > warned... ;-)] > > ids are attributes that are allowed in every element. What I would to have > is a clever and clear way to deal with them. > > Take the ids for AGU as an example, that we have already discussed. One of > the proposal IIRC, was to allow all labels (ids) to appear in the graphical > interface, even the automatic.
No strong opinion on that one. > What I propose now is never to include the automatic labels, but to > translate from manual to automatic. This is something that lyx can do quite > well. > > Lets assume that your label called 'aha_gotcha' is located on paragraph > 151, according to AGU you should always use 'para151' to refer to that > lable, except if you add another paragraph before where it should be > 'para152'. The problem for us humans is that those references are not > stable, it is easy to refer to them as 'aha_gotcha'. As I understand it, the manual label overrides any automatic one. So in this case the previous paragraph would have label para150 and the next para152 (and these change dynamically as you add and remove paragraphs) but the label of this paragraph is and remains aha_gotcha. (For AGU, dunno about others) So translation from manual to automatic as you propose wouldn't be a good idea. Internally of course the counter should take the aha-gotcha paragraph into account, just not display/print its automatic label. > LyX always know where the label appears, so it is easy for it to do that > convertion automatically for the user. > > Now I would like to have a mini-language to put all this knowleadge in the > textclasses, the same we do with those $$i, $$o, $$b in the converter stuff. > > I don't want something that only works for AGU, I want something that > works for any XML based format. > > Have I been clear regarding my motivation? :-) Clear enough. If you can pull it off... -- Martin Vermeer <[EMAIL PROTECTED]>
signature.asc
Description: This is a digitally signed message part
