On 7/24/08, Ted Zlatanov <tzz@lifelogs.com> wrote:
>
> On Wed, 23 Jul 2008 22:08:29 -0400 Stefan Monnier <
> monnier@iro.umontreal.ca> wrote:
>
> >> The advantage over categories, which you suggested, is that the
> >> completion code won't need categories supplied--it automatically gets
> >> them based on the word syntax, so it's less work for the programmer and
>
> SM> Maybe I misunderstand what you mean by word syntax.
> SM> Are you referring to something more like partial-completion-mode?
>
> That's close, but it looks for abbreviations rather than categories, and
> is nicest when you know the word you want in advance.  When you don't,
> you still have to walk through the list of candidates.
>
> The essential problem is that more than 40 or so candidates are simply
> unusable--it's not just slow, it's unmanageable.  You don't know what's
> at the end of the list until you go there, and by then you've forgotten
> what was at the beginning.  With categories you end up with meaningful
> prefixes instead, as in my example.  The category name is meaningful
> because the name of the thing you are completing is split in an
> informational way, not (as p-c-m does) by first letter, and then the
> first thing from the split string is used as the category.
>
> If this doesn't sound useful I'll drop it.
>
> Ted


To me it sounded very useful.