Fwd: [mb-style] [Clean up CSG] Correct punctuation (was: Typography)

[Bogdan Butnaru]

Yeah, I don't know why I keep missing these list addresses :) Here was my post.

---------- Forwarded message ----------
From: Bogdan Butnaru <bogdanb at gmail.com>
Date: Feb 12, 2008 1:51 PM
Subject: Re: [mb-style] [Clean up CSG] Correct punctuation (was: Typography)
To: dave at gasaway.org


On Feb 12, 2008 6:08 AM, David K. Gasaway <dave at gasaway.org> wrote:
> On 11 Feb 2008 at 17:44, Bogdan Butnaru wrote:
> > I understand and (partly) agree all three arguments I left standing
> > above. But there's one thing that's bothering me: they can just as well
> > apply to all our other rules.
> I can't outright disagree with any of what you said.  Instead, I've
> been pleading that we consider the relative value versus the cost.  In
> other words, a little bit of incorrectness might be worth the savings
> in hassle for the users.
> > [...] The
> > proposal was just to define correctness in the (few and rather uncommon)
> > cases where the current rules are "just use ASCII", and replace the rule
> > with "Please do it the nice way if you can and care.
> My problem with this approach the end result: A database where the
> rules are not applied uniformly and the data is inconsistent.  Worse,
> this inconsistency is *explicitly condoned* by the guidelines.

Both good points, and please let me point out that they can be combined :)

I assume we'll all agree that the point of the guidelines is to strike
a balance between (in)correctness, user hassle and (in)consistency.

(Why? We'll always have some inconsistencies, because of errors that
are not spotted, different interpretations and ambiguities in
guidelines. I can give examples if you want.) Even the fact that
they're guidelines instead of rules suggests they're not absolute (and
the http://wiki.musicbrainz.org/StylePrinciple page denies it
explicitly).

But let's return to that balance. The "correct punctuation" guideline,
if applied, would give us some more correctness at the expense of user
hassle. The "ASCII is OK too" guideline would transform the user
hassle into inconsistency. I submit that the trade-offs are
advantageous, because the hassle or inconsistency would be:

1) small in absolute terms, because the punctuation we're discussing
is uncommon.
2) even smaller in "perception" terms, because people who notice these
kinds of things would have already fixed the releases they have, and
the rest... won't notice.
3) traded-off nonlinearly for correctness (which is why MB works in
general): a single user can fix a tag (on unit of hassle) and everyone
uses it (many units of correctness).

In conclusion, there's a trade-off, but I think it's a worthy one.

I haven't forgotten the "explicitly condoned" part. I don't think
that's a big issue (in _this_ case), because (a) the domain is small
(few cases where the problem arises), and more importantly (b) we
already do this implicitly in many more cases. (For instance, unless
it's a total mess, almost no-one no-votes a release with incorrect
caps, we just fix them and maybe leave a note. And for non-English
titles, many voters don't even notice them; those who do go back to
clause one... I can give some more examples.)

As I said before: "if a tree falls in the forest and nobody's there to
hear it, does it make a sound?". It is interesting as a philosophical
question, but for adopting guidelines I prefer the more practical "Who
cares, listen to rest of the forest!"


-- Bogdan Butnaru — bogdanb at gmail.com
"I think I am a fallen star, I should wish on myself." – O.



-- 
Bogdan Butnaru — bogdanb at gmail.com
"I think I am a fallen star, I should wish on myself." – O.