MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Message-ID:  <3986.1295947424@cl.cam.ac.uk>
Date:         Tue, 25 Jan 2011 09:23:44 +0000
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@listserv.uni-heidelberg.de>
Sender: Mailing list for the LaTeX3 project <LATEX-L@listserv.uni-heidelberg.de>
From: Robin Fairbairns <Robin.Fairbairns@CL.CAM.AC.UK>
Subject: Re: \clist_length:N and \clist_nth:N
To: LATEX-L@listserv.uni-heidelberg.de
In-Reply-To:  Your message of Tue, 25 Jan 2011 08:47:00 +0000. 
              <4D3E8E04.8000802@morningstar2.co.uk>
Precedence: list
Status: R

Joseph Wright <joseph.wright@MORNINGSTAR2.CO.UK> wrote:

> On 24/01/2011 14:01, Lars Hellström wrote:
> > Isn't this "elt" an implementation detail for that type of list (various
> > \@elt tokens in 2e come to mind), and thus something that should be kept
> > internal rather than canonised in a public interface? The clean solution
> > for *both* types of list is rather to use "length".
> 
> Seems reasonable: one would normally talk about 'a long list of things
> to do', so lists to have 'length' :-)

in some ways of looking at a list, it has 'length'.

in plain english, i would say lists have "items" (or just "things")
rather than "elements", but "item" is already heavily overloaded in
the user's level latex-ese.  i don't have a problem with "element" (or
"elt" in command names).

> > Moreover, I get a vague impression that the term `elt' is part of the
> > pseudo-LISP heritage of LaTeX (emphasis on the "La"). If so, then that
> > is IMHO another reason to avoid it, as that heritage is full of square
> > pegs trying to fit in round holes.
> 
> Not being familiar with Lisp, I can only go on things like LaTeX2e's
> \@cdr, etc., which have much more sensible names in expl3.

the "lisp heritage" is no more than simple use of some lisp names for
some latex internal operations (e.g., car, cdr).  it's a long time (>40
years) since i learned lisp, but i don't remember a special name for
items in a lisp list.

> >> I didn't write "clist_nth" with a view of it being the permanent name,
> >> but
> >> now that I've written it I can't think of a (good) alternative. Any
> >> thoughts?
> > 
> > I think the verb you're looking for is "index", i.e., the command name
> > would be clist_index.

hmm.  index works, but i don't think it supports the "plain english" test.

> I'd imagine 'index' to be the other way around:
> 
>   \clist_index:Nn \l_some_clist { item } => Number

agreed.

> whereas what Will has implemented gives the 'entry', 'element', 'item'
> or some such name. ('element' seems to be discouraged based on the first
> part of your e-mail, so perhaps 'item' is better.)

i think element is as good as it gets, in this context.

> >  first: Return index of first occurrence of a particular item within a
> > clist,
> >         or -1 (given 0-based indices) if the item does not occur therein.
> >  last:  Return index of last occurrence of a particular item within a
> > clist,
> >         or -1 (given 0-based indices) if the item does not occur therein.
> >         (Note: Slightly trickier to implement.)
> 
> Both of these look relatively easy to do.
> 
> >  range: Return a subrange of the clist, i.e., if \a_clist is "a,b,c,d" then
> >         \clist_range:Nnn\a_clist{1}{2} returns "b,c". (I don't have an
> >         opinion as to what might be the best sense of "return" in this
> > case.)
> >  replace: Replace the material in a subrange of the clist by some other
> >         clist material.
> 
> More tricky. Let's sort the others first :-)

robin

getting better every day.  very slightly.