Re: Long file names in Dired

[Marcin Borkowski <mbork@mbork.pl>]

On 2015-04-24, at 12:19, Rasmus <rasmus@gmx.us> wrote:

> Marcin Borkowski <mbork@mbork.pl> writes:
>
>>>> I. HATE. BIBTEX. WITH. A. PASSION.
>>>> [...]
>>>
>>> I find the format OK.  I can't imagine a vastly superior format for
>>> metadata...
>>
>> 1. Multiple authors are separated with the five-character string
>> " and ".
>
> I type quickly enough for this not to bother me.  Perhaps ";" would be
> better, but for the sake of comparability I can live with " and ".  OTOH
> it may make human parsing more easy when reading the bib file.

It's not about typing.  It's about the stupidity of the format itself
(see below).

>> Can you imagine a more stupid idea?  What if you want to quote
>> an anonymous report whose author field should say "National Aeronautics
>> and Space Administration"?  Yes, you /can/ quote the " and ", but this
>> shouldn't be even needed.
>
> Author = {{National Aeronautics and Space Administration}}

I know.  But see above and below.  Why not use amsref's approach of
"repeatable fields"?

>> 2. BibTeX separates the author name into (at most) four components,
>> called "first", "last", "von" and "jr".  No support for: Chinese names
>> (not fitting into this format)e and names in some other languages (I'm
>> forgetting now which ones).
>
> Another flaw is the auto-guessing of names e.g. for three names one must
> use "last, first middle" vs "first last" or "last, first" for two names.
>
>> 3. What about alphabetic sorting in various languages?
>
> This is orthogonal to the format.  I believe Biber reads unicode.

Fair enough.

>> 4. Since BibTeX is aimed (after all) at a typesetting system, what about
>> diferent hyphenation patterns?  Imagine an article written in Polish,
>> citing a paper by a German (so the name should be hyphenated differently
>> than the rest of the document), the paper being in English, and
>> published by an Italian university.  Where's support for that (i.e.,
>> language field for the author, title etc.)?
>
> Biblatex will read language field.  I don't know how it uses it.  In any
> case, that's orthogonal to the bibtex format as a "database".

No, it's not.  As I hinted above, what's really needed is a language
"field" for /any other field/, separately.  One "language" field per
/entry/ is not enough.

>> (Not to mention several authors from various langauge backgrounds.)
>
> If there's a name you can't type, I guess it's the fault of unicode.

See above.  Due to the idiocy of the " and " approach, even if there
were a method of declaring different languages for different fields, it
wouldn't help here.  This is not about scripts, but about
e.g. hyphenation patterns (which are language-dependent), or decomposing
names into parts (which BibTeX /the program/ can't do correctly anyway,
and the format doesn't help a lot, either).

>> 5. What about all these funny characters not found in English, and in
>> particular sorting using them?  (See the docs for the xindy indexing
>> tool for examples of nontrivial problems regarding locale-aware
>> sorting.)
>
> This is a flaw with the bibtex program.  Biber is better.

Of course, you're right here.

> BTW: Emacs 24 can't do local-aware sorting (but string-collate-lessp is in
> 25).  This would suggest to me that it's a nontrivial problem.

Yes, it /is/ nontrivial.  (See xindy's docs for real-life examples.)

>> 6. What about custom citation styles?  Have you seen the syntax of the
>> bst files?  If yes, you know the pain, and if not, you'd better not look
>> at them, for your own good...
>
> Yes.  I used to use custombib or whatever it was called to generate bst
> files.  Pain.
>
> Again, this is orthogonal to the bibtex file-format.

Again, you're right here.

>> 7. BibTeX is not really "case-sensitive" or "case-insensitive", it's
>> "case-destroying".  IOW, /you have to quote capital letters in titles/
>> so that BibTeX doesn't convert them to lowercase.
>
> I think you can use: {{}}.

Yes, but /why/ should the user be enforced to do /that/?

>> BTW, BibLaTeX addresses (AFAIK) problems 3, 5, 6 and 7.
>
> Would 3 not be biber?  For 4 perhaps you can use the langid field (pp. 24
> in the "texdoc biblatex" in TL14).  I don't know if it's as extensive as
> you need.

By "BibLaTeX" I mean the whole thing - the LaTeX package /and/ Biber.

>> The amsrefs package addresses problems 1, 2, 4, 6 and 7.
>
> I don't know it.  The AMS math packages are good quality.

Yes, at least from the user's perspective - reading their code can
sometimes reveal some funny skeletons in the closet;-).

>> Again: nice, and thanks for the tips.  But still, BibTeX is a wrong tool
>> for that.
>
> I'm using the *format*, not the binary "bibtex".  Biber is good, and via
> reftex Emacs has good support for the bibtex format.

So I heard.  It's a pity that reftex doesn't support amsrefs.  (I'm
considering helping to remedy this situation.)

>> Note that no BibTeX based solution helps /if I want to be able to move
>> files between directories/.
>
> In the above system you'd move *folders* rather than files.  An entry is a
> container (folder) of:

So it's even worse.  What I want is the ability to e.g. simply move the
pdf from the directory "to-read" to the directory
"articles-on-fixed-points".

> —Rasmus

Best,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University