User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:24.0)
            Gecko/20100101 Thunderbird/24.6.0
MIME-Version: 1.0
References: <m3k34m4vjd.fsf@somewhere.org>
Content-Type: text/plain; charset=UTF-8
Message-ID:  <5429D0DF.7060608@morningstar2.co.uk>
Date:         Mon, 29 Sep 2014 22:36:31 +0100
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: Joseph Wright <joseph.wright@MORNINGSTAR2.CO.UK>
Subject: Re: [l3keys] Suggestion: Add a key property for specifying key that /has/ to be used
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
In-Reply-To:  <m3k34m4vjd.fsf@somewhere.org>
Precedence: list
Envelope-To: <rainer.schoepf@GMX.NET>
Content-Transfer-Encoding: 8bit
Status: R

On 29/09/2014 10:37, Denis Bitouzé wrote:
> Hi,
> 
> among others, `.value_forbidden:` and `.value_required:` key properties
> from `l3keys` package are useful: they let us specify that a given key,
> when used, either cannot or must receive a value.
> 
> What could be useful as well is a key property, say `.required:`, that would
> specify that the corresponding key /has/ to be used. This could be helpful for
> instance in the case of a document command `\MyModuleSetup` for setting up
> a module `mymodule`:
> 
>     \DeclareDocumentCommand \MyModuleSetup { m }
>       { \keys_set:nn { mymodule } { #1 } }
> 
> where some module property /has/ to be specified.
> 
> I know this can be achieved with something like test of existence of
> some tokenlist and message emitted in case of nonexistence but a high
> level key property for this could be nice.
> 
> May I ask what is your opinion about that?
> 
> Thanks in anticipation.

Sean has pointed to a similarity to dealing with mandatory arguments at
a design level. This request seems a little different as l3keys isn't
meant to be targeting that area. However, there is clearly overlap.

At present, key properties mainly apply to individual keys in a 'stand
alone' sense. This request is different as it's actually about an entire
set of keys. To me, that seems a bit odd: it's pretty rare that buried
inside a set of optional settings is one or more mandatory ones. There
is also the question of what 'required' might mean, as

    key = \empty

or similar might well be equally invalid in the real use case but would
presumably pass such a test.

What I think we need before we can really comment is some solid proposed
use cases, showing why the property is needed as opposed to other
approaches as you outline.
--
Joseph Wright