Andre Poenitz wrote:
>
> > I had a fast look and what I don't like is all this <xxx>some option</xxx>
> > stuff couldn't that be handled with something like <xxx value="some option">,
> > as this seems much nicer too me!
>
This is precisely why this output is handled by a small function that I
can redefine at will.
> I think Pierre-Olivier got it right in general. It is some kind of
> religious debate on what goes in <xxx>...</xxx> and what goes in "...".
>
> The best rules I found so far is:
>
> 1. If the removal of some item does remove something that is important for the
> understanding of the document, or if other items depend on its presence
> and/or value, it goes in <xxx>...</xxx>. If the item can be removed without
> consequences for the content of the document (like font size etc), it goes
> in xxx="...".
>
> 2. If the decision after rule 1 can not be made clearly in favour of
> xxx="...", use <xxx>...</xxx>.
>
> Of course this still leave room for interpretation, but most times it's
> pretty easy.
>
So I my interpretation is that <xxx> ...</xxx> is right in the
parameters section. Especially since some parameters like the preamble
would be strange in a <param name="preamble" value="\include ..." />
element. So I use the same format for all parameters.
But of course it is really easy for me to change this (and I don't
mind).
I will clean the code this week-end then start the parser. I would have
like to avoid writing many XMLRead methods and write a LyXXMLReader
class instead that only uses constructors of LyX objects but that does
not seem feasible (and maybe it is not a good idea).
Thanks for your opinions.
P.O. Gaillard
