(Sent from my phone. Apologies for brevity.)

On 18/08/2009, at 0:38, "J.Fine" <[log in to unmask]> wrote:

> Manuel wrote:
>
>>> Thank you.  It's not the only nice thing we can do.  Here are some  
>>> more:
>>>
>>>  * Allow digits in control sequence names
>>>  * Allow period in control sequence names
>>
>> I think the naming scheme (at the programming level at leat) has been
>> discussed
>> previously and is now considered stable. Let's not re-discuss again  
>> and
>> again
>> points were decisions are already made. (And btw, I personnaly  
>> think the
>> namign
>> scheme in expl3 is good.)
>
> My suggestion is compatible with the LaTeX3 naming scheme.  Did you  
> think it was not?

Period (I think) and digits (definitely) must all be catcode other to  
work inside numexpr and so on. I wish it were otherwise, but I don't  
think it's nice to have to use scantokens to sanitise input for all  
numeric work.

>
>>>  * Allow '~' to produce a space /in all circumstances/
>>>
>> What would be the purpose?
>
> So that
>   \this ~ \that
> has as space between \this and \that, for example.  Or to get  
> several spaces in a row.

~ is currently catcode space in expl3 since space itself is catcode  
ignore. So I don't think this suggestion is feasible either.

Will