New tagcolledit
Erich Schubert
erich@vitavonni.de
Wed, 21 Apr 2004 00:51:08 +0200
Hi,
> By looking at the newly downloaded tag collection, I see that
> not-yet-tagged tags are being generated without facet/namespace: could
> you please update your scripts to generate them as
> special::not-yet-tagged* ?
Yes, this was a mistake by me using the old scripts.
> What about guile? I'm fine with treating it as a separate language,
> too.
"Guile is a Scheme implementation designed for real world programming,"
If guile thinks of itself as scheme?
> Navigation seems to be a bit awkward, probably because I got used at
> seeing the facets/namespaces as "points of views" instead of tags, and
> I'm finding strange to see them intermixed with tags.
Yes, but this is what i got without modifying my code. ;-)
Also i find that it slows you down to select a namespace, then select
some facets, then switch to another namespace etc.
I find it nice to have the interface suggest ways to refine the search
also on a namespace level.
I think it will help to write descriptions for the facets and
namespaces; probably the Namespace should be part of the facet
description, like this:
Tag: use
Description: By field of use
Tag: use::editing
Description: by field of use: data editing
Grouping them is of course something i have always wanted, i think i had
a "flag" added for such cases, but i hadn't implemented that.
("use" would be flagged "keep-grouped")
> ...and so on. It should also give efficient and compact navigation.
Just throwing the single-words at the users is bad IMHO.
AFAICT they prefer a short sentence; especially to some technical term.
> > Please check out the list of used tags, too.
>
> Sorry, I can't understand: what do you mean here?
Sorry, badly written. I meant the complete tree of tags like you are
shown when editing a package.
> I'll start sending you tagcoll-diffs generated with
> debtags update; debtags mkpatch <file>
> that is, diffs against the latest debtags-available database.
against which version? i used the one you sent via mail.
> We could use the same logic to define "utility": a program which is
> necessary in some task, but which purpose is not to perform that task.
So a screwdriver is not an utility? "cat" is no utility?
On the contrary: a "utility" is designed to perform just a small simple
task; but you will rarely use it alone but in order to modify or extend
the behaviour of other applications, on its own it will be mostly
useless.
I think of utilities like "glue applications", or "building blocks".
> In light of this, I added "utility" to "role::". I'll add the
> description to the vocabulary.
utility is a role, yes.
Greetings,
Erich