Content-Type: text/plain; charset=iso-8859-1
Mime-Version: 1.0 (Apple Message framework v1081)
References: <4C967C9B.5010706@gmx.de>
            <71ED7A5B-DE11-45D9-9785-95A1DC795080@gmail.com>
            <4C9724B4.3010301@residenset.net>
Content-Transfer-Encoding: 8bit
Message-ID:  <AA223041-D9B8-4EE4-BF8A-5782FAC2078A@gmail.com>
Date:         Mon, 20 Sep 2010 19:58:35 +0930
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: Will Robertson <wspr81@GMAIL.COM>
Subject: Re: boolean expressions in ExplSyntaxNames
To: LATEX-L@listserv.uni-heidelberg.de
In-Reply-To:  <4C9724B4.3010301@residenset.net>
Precedence: list
Status: R

On 20/09/2010, at 6:39 PM, Lars Hellström wrote:

> My opinion, if anyone cares, is that using && and || at all in the syntax of \bool_if_p:n is stupid eye-candy. The stated situation is just one of the ways in which it leads to problems;

Sorry for the bug. I don't anticipate further problems of this sort, since \ExplSyntax(On/Off) guards against it.

> another is that it suggests || and && may sensibly be mixed within the condition, which isn't the case either.

How do you mean they can't be mixed? The documentation gives the example

\intexpr_compare_p:n {1=1} && (
  \intexpr_compare_p:n {2=3} ||
  \intexpr_compare_p:n {4=4} ||
  \intexpr_compare_p:n {1=\error} % is skipped
) && !(\intexpr_compare_p:n {2=4})

> The natural way to express AND or OR of a bunch of boolean conditions in a macro language such a LaTeX is the LISPish one: have macros/functions whose meaning is "AND of my arguments" and "OR of my arguments". IIRC, that would be
> 
>  \bool_and_p:nn{<predicate 1>}{<predicate 2>}
>  \bool_or_p:nn {<predicate 1>}{<predicate 2>}
>  \bool_and_p:nnn{<predicate 1>}{<predicate 2>}{<predicate 3>}
>  \bool_or_p:nnn {<predicate 1>}{<predicate 2>}{<predicate 3>}
>  ...
> 
> for as long as one likes to continue that list.

It would be straightforward to add these to the syntax; there's already \bool_not_p:n and \bool_xor_p:nn. No reason not to have both; I just prefer infix notation to

\bool_and_p:nnn
{ \intexpr_compare_p:n {1=1} }
{
  \bool_or_p:nnn
  { \intexpr_compare_p:n {2=3} }
  { \intexpr_compare_p:n {4=4} }
  { \intexpr_compare_p:n {1=\error} }
}
{ \bool_not_p:n { \intexpr_compare_p:n {2=4} } }


On 20/09/2010, at 7:40 PM, Chris Rowley wrote:

> My intuition is to agree with Lars, for the designs reasons he gives and because one mission of this language is that it is reasonably human understandable and consistent.
> 
> My prejudice is not to use things like Ś &  because _I_ have never used them elsewhere

They're natural to me, but our mileages no doubt vary, as they say.

-- Will