On Tue, Apr 15, 2003 at 05:34:01AM -0400, Bryan Ewbank wrote:
> Iain Buchanan wrote:
> >>The existing \todo is a perfect example of the behavior, but the desire
> >>is to have a user-defined title for the generated page that contains the
> >>reverse index. What we know as \todo is just an alias for \reverse todo.
> >I agree this might be a good way to solve it without getting too much
> >'feature creep'.
> I think there is just a small difference here - I see this as a simple
> refactoring of the language, not feature creep.
> In this case, for example, there's a very handy implementation that is
> bound to a specific data type ("todo"), but it could perhaps be
> reimplemented as a more generic abstract base class or template from
> which the user can construct their own implementation for other data
> types ("wibni", "feature", "my code", ...).
> Imagine if there were no \page command, but instead commands for
> specific page names. In the same way, there is currently no command for
> a generic reverse index (that is, B seen in A adds: (1) B in A (2) A in
> B (3) links between). That appears to be inconsistent with the other
> elements of the language (\page), so rewriting the feature to use a
> different notation is reasonable.
> We'd of course need to keep \todo as an alias for "\reverse todo" or
I think this is indeed a good plan. I've already made \todo and friends
more generic in the past to facilitate something like this. I was thinking
about \xrefitem as a command name (which sounds more intuitive to me
than \reverse, but this is open for discussion).
The syntax would be as follows:
\xrefitem "listfilename" "Item Heading" "List Title" text about the item
The \todo, \test, \bug and \depreciated command are then built in aliases
for this command (which then happen to be i18n aware). Other user-defined
commands can then be added using aliases.