Will Robertson wrote:
> On Wed, Sep 17, 2008 at 4:49 PM, Joseph Wright
> <[log in to unmask]> wrote:
>> I've been giving a little thought to how I might implement something
>> like siunitx in expl3. An initial assessment points to two support
>> packages: array and amstext. Both strike me as pretty much essential
>> for LaTeX3, so I wondered if it is likely that either functionality
>> might make it into the new kernel?
> Good question :)
> I suspect the underlying interface to tabular material will remain
> rather similar. (Although I'd like to write a better *user* interface,
> one day.)
Something like ctable, or in some other way?
> Either way, I can't see anything happening in the near future since
> porting array would be a big job
It's not that long! Seriously, the key issue is of course whether the
entire approach is correct.
> and maths is all rather up in the air
> at the moment. (breqn is taking form but it's very early days yet;
> it's current incarnation on CTAN is probably the last that *won't* use
> the expl3 bundle, though.)
> But there's no problem, of course, using their "LaTeX2e" macros in your package.
Well no, but this comes back to a point I've raised before. It's hard
to know what might need to be written in LaTeX3 as there is no way to
tell what will and won't be in the kernel + required packages (or even
what might be in!). So I can't tell if I might be better knocking up
something like amstext myself, or whether things like the font detection
I do in siunitx will still have to be done in the package, or whether
the kernel might handle some of this.