vocabulary structure

[Peter Rockai]

On Tue, Jun 27, 2006 at 11:49:52PM +0100, Justin B Rye wrote:
> Peter Rockai wrote:
> > oh wait, i don't play that
> > one. Anyway, with your definition (from below), made-of:: is mixing
> > "implementation language" and "formats of data files in the package" which
> > look pretty unrelated to me. So not a good facet.
> 
> I was understanding it as meaning something like "what kinds of file
> the package is built from" - fonts, icons, Perl modules, C code...
> (Looks it up) Ah, the long description says "the languages or data
> formats used to make the package"; sounds about right.  It's true
> that the "made-from::code:c" and "made-from::code:c++" packages
> mostly end up on my system in a precompiled form; but that needn't
> stop us labelling them in terms of what they're constructed out of.

Well, that meaning is completely unobvious. If you have a facet for which you
can't make a name in 1-3 words (that would explain the semantics of the
facet), it probably has cohesity problem. It also makes it a lot harder to
search for things, since if you misunderstand the meaning of a facet (like i
until now did, never occured to me that "made of" would mean "contained in the
source package" -- and i am probably not alone, looking at all those packages
with generated documentation labelled by made-of::data:<format>). And if for
lang: "group" it means "implemented in" and for "data" it means "contained in
the result package" it is inconsistent and really confusing.

> > Well, my main problem is that made-of is poorly defined (from looking at its
> > tags, too).
> 
> Actually, rotten tagging would make it a lot easier to face the idea
> of just throwing tags out.  But what problems are there?  Is it just
> inconsistent decisions about whether trivial shell-scripts are worth
> adding a "made-of::lang:shell"?

No, see above.

> [...]
> > So we get development-language and
> > implementation-language. Does that sound plausible? Maybe with better facet
> > names?
> 
> Sounds okay to me.  I'm beginning to wonder whether this list is
> quorate, though.

What do you mean?

> > I am quite sure
> > that's not nearly enough to justify separating these two.
> 
> Er... okay, you've disproved something or other... but what has this
> got to do with the format:: facet?

That the understands-format:: and contains-format:: distinction is not
neccessary, since the exact meaning of format:: for a given package can be
determined completely by other tags. It probably still makes sense to separate
them, only that it's not completely neccessary.

> > Basically, the
> > problem with made-of:: is, that semantics seem to be "primary content of
> > package", which means that if it's data it basically cannot be a program
> > package which basically means there is no way it can "work with" anything.
> 
> A lot of foo-data/foo-doc/foo-common packages are tagged in parallel
> with foo(-main) - so apache2-doc's tags claim that it "works-with"
> HTML, and so on.

And those are wrongly tagged, since apache2-doc doesn't work with anything. I
have only checked for the made-of::data:* tags and looked for packages that
have a potential to "work with" something, not using tags because right now,
they are quite unreliable. Probably has to do with taggers being confused, too
:-).

> > On the other hand, if you had a contains:: facet, that would be somewhat
> > different.
> 
> So it would be okay for apache2-doc to be both "contains::html" and
> "works-with::html"?

Nope. First, contains:: was (initially) not supposed to be about formats but
about kinds of things found in the package. Like, executables, source code,
audio, (video? we have any?), text...

> > Question is, if such a facet would be useful at all. If it is, you
> > can have contains::<what>, contains-format::<format> facets. Similar to the
> > above language solution :-).
> 
> No, you've lost me.

Well, the contains-format:: would then specify what formats are to be found
inside the package. Like elf, plain text, pdf, html...

> > Well, the semantics of made-of::lang: are very obscure at best. They are
> > definitely different from what the name of the facet implies.
> 
> To me it it seems only very slightly forced; but if there are
> independent theoretical arguments against nested subfacets I'd have
> no problem with them being shifted back.

Well, so now i was explained how is "made-of" defined (at the start of the
mail), i can understand that it is OK, but most people don't and the taggings
don't correspond with the definition. Which makes the facet basically useless,
when you need to go to debtags-devel to find out what is it supposed to mean
:).

-- 
Peter Rockai | me()mornfall!net | prockai()redhat!com | +421907533216 
   http://blog.mornfall.net | http://web.mornfall.net

"In My Egotistical Opinion, most people's C programs should be
 indented six feet downward and covered with dirt."
     -- Blair P. Houghton on the subject of C program indentation