Lars =?iso-8859-1?Q?Hellstr=81=F6m?= writes:

 > My point was that by providing some kind of "compability template" for
 > galley-related stuff one could make the transition from the LaTeX2e to
 > LaTeX2e* much easier.

i'm not sure that this is the way to go. but then i'm not at all sure about
the galley prototype at all (in comparison to most of the rest (even though
they are too prototypes only))

i rather think somebody (and not necessarily me, since i might just run in
circles) needs to take a heart and do some further thinking about capturing
those problems with the paragraph galleys.

In the mean time one might in fact have a set of simpler templates that model
some of the galley interfaces, eg a "justification" template that accepts the
same arguments like those in xhj but does only this in a more direct manner.

 > Today it is hard to start using galley simply because
 > you have to redo _everything_ in your layout that galley is involved
 > in---thus you face an enormous amount of work even though each part of it
 > may have become much simpler. If one could redo e.g. the frontmatter using
 > the galley interfaces and then select the compability template (which would
 > essentially have galley try its best to imitate the old interfaces) for the
 > reminder of the document, then I suspect that more people would consider
 > trying it out.

that might be an option if we find a way to sort of confine galley lowlevel
stuff to certain parts of the document but I don't see that really working
since the model is so different from lowlevel TeX. Thus you will not find it
easy or possible to switch between these models.

 > It's basically the old "how do we get enough people to switch to XXX"
 > problem that I think needs to be addressed.

true, but galley as such seems to me a difficult candidate to chose a path,
mainly (as I said) because I'm less anyway.

my current path of thinking (which is a good bit future) is to finish of all
major parts first, some of which will me usable standalone perfectly and if
those eventually form a more or less consistent base then perhaps even spend
some money on converting major packages to properly work with that kernel.

that would then form a kernel + "enough" packages that a large amount of
document could be processed without the need for development "beforehand" and
would also serve as a good model for building further extensions.

just a thought though (yet)

speaking of stand alone packages: the frontmatter stuff that i'm intending to
finish for Gutenberg 2001 in May (as a prototype :-) is of that nature, it
would be usable with just 2 (or perhaps 3) basic modules, ie template.sty (or
a successor incorporating your suggestions), xtools.sty (which is a compressed
version of l3expl code suff like property lists queues etc in @ notation) end
perhaps a variant of xhj as a standalone version.

frank