Le 02/04/2016 00:12, Karl Berry a écrit :
> I think that it would be valuable if we had an environment index and a
> control character index.
>
> Sorry, but I disagree. I do not want multiple indexes. I think they
> make things harder for readers, because they create multiple places to
> look for the same information. (Indeed, I'd like to merge the concept
> and command indexes. I inherited them.)
>
> What I can agree with is that index entries could contain text
> like " (environment)".
Ok, what I was unhappy with is that the findex of environments was
inconsistent (the word `environment' was not always used). Also it is
quite evil to duplicate information, and repeating in many places the
word `environment' was evil to that extent.
So what I did is create an EnvIndex macro to produce the index
entries. So in result we have a single index, as you are willing, but it
is made consistent and futureproof, ie easy to maintain. Also, each
EnvIndex call indeed creates two entries, one is
@r{environment}, @code{\env\}
and the other one is
@code{\env\} @r{environment}
So we can find all the environments in the same place in the index by
looking at word `environment'
>
> Since they are not so many,
>
> Precisely because there are not so many, I don't think there is any
> great problem with having the list in the text. A whole separate index
> would be overkill to me. -k
This means that they should be all indexed with some macro like this:
@macro CtlCharIndex {cc}
@findex @r{control character}, code{\cc\}
@findex code{\cc\} @r{control character}
@end macro
This way all the control characters could be found in the same place by
looking for `control character' in the index. The same argument as for
using EnvIndex macro apply here : consistency & futureproofness.
V.
