emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: A proposal for a friendlier Emacs


From: Alexander Adolf
Subject: Re: A proposal for a friendlier Emacs
Date: Tue, 29 Sep 2020 19:08:14 +0200

Richard Stallman <rms@gnu.org> writes:

> [...]
>   > Why does it matter whether we call it "category" or "tags".  The
>   > important part is what the user thinks about a package.  And a single
>   > package can satisfy several different needs.
>
>   > So I think the idea of categorization that assumes a single category
>   > per package is basically unworkable, because it's too restrictive, and
>   > won't stand the test of time.
>
> I agree.
>
> In any case, the implementation of this doesn't need to make an assumption
> about how many categorizations a package can have.
> [...]

It seems I was talking about possible technical solutions instead of
semantics, which appears to have gotten us into confusion.

My intent was to attach one new concept to any given package in the
repository. This concept would define the "init configuration topic" of
the package, so as to allow Custom code to auto-generate an init file
for that topic, and collect the config for all packages of that topic in
this init file.

In this thread, Eli has suggested a second concept to be attached to a
package, which is a (virtually) unbounded list of words and terms, which
would be matched against search terms users employ to find a package for
their needs.

I would hence view these as two distinct concepts, serving distinct
purposes. The "init configuration topic" concept is chiefly technical
information to partition init code into smaller units. The "unbounded
word list" concept is chiefly UX information as it is used to improve
search results' relevance.

Note that I have deliberately shied away from choosing any labels for
either of the concepts, under which they would be used in a vocabulary
or user interface (such as for instance "tag" or "category").

While I see Eli's suggestion as a worthwhile addition (as we appear to
be debating the metadata data model of the package repository), my
intent was to use the concept I described to enable a new approach to
Custom handling init files.

So I don't see the two concepts as mutually exclusive, competing, or
otherwise in conflict. The could very nicely coexist, or be adopted
independently of each other. Just let's not mix them up; that's all.


Many thanks and looking forward to your thoughts,

  --alexander

Attachment: signature.asc
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]