Re: Apropos commands and regexps

[storm@cua.dk (Kim F. Storm)]

Miles Bader <miles@lsi.nec.co.jp> writes:

> Eli Zaretskii <eliz@is.elta.co.il> writes:
> > > Wouldn't it be simpler (for a novice user -- and for advanced users
> > > too) to simply write one or more words (substrings) and then search
> > > for all combinations of those words (substrings) in the relevant list.
> > > 
> > > E.g. C-h a open file RET  would find any matching
> > > 
> > >         open.*file and file.*open
> >
> > Perhaps we should have a new command for that, and name it something like
> > apropos-keywords.
> 
> I agree, but I think it shouldn't use the wierd hack on regexp syntax,
> that's just confusing.

I agree that we might find something better than what I suggested; it's
a starting point which can be improved...

> 
> I'd say just separate the keywords by looking for commas or whitespace
> or either (each `keyword' could be a regexp though).  That would be
> both more convenient and also more familiar to people used to using
> typical keyword searches (e.g., in web search engines).

We could put \b around the words in the regexp if we don't want substring
matching.

The obvious problem restricting this to complete words is how to make
e.g. "list process" match "list-processes".

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk