Re: [Bibulus-dev] Bibulus DTD: An concrete example

[Thomas Widmann]

Torsten Bronger <address@hidden> writes:

> Thomas Widmann <address@hidden> writes:
>
>> Torsten Bronger <address@hidden> writes:
>>
>>> [...]
>>>
>>> <city refid="de/Munich">
>>>
>>> with an appropriate city data base entry.
>>
>> Better, but <city country="de">München</city> would be better since
>> it would provide a sensible fall-back if it wasn't mentioned in the
>> city database.
>
> Yes, right.
>
>>> Both is unrealistic in my opinion, and unnecessary.
>>
>> We'll know more about realism soon, but it *is* necessary.
>
> I don't think so, but because you think it is and I must admit that
> I've thought about it only shortly, I'll try to help to find a
> solution.

Great!

> Is there any way to avoid creating a list of city names, at least
> for the most important, say, 50?  I don't see any.

No.  But I think it's important to note that the list probably will be
different from language to language.  Typically, a language will only
have specialised words for cities in their neighbouring countries and
some major world cities.  Or in other words: Although the Dutch and
French lists presumably will be full of Belgian place names, there's
no need to list them all in the English or Danish lists.

> It needn't be an XML database, a mere list should be enough.  But it
> must become part of the standard.

Sure, but of course it will be locally expandable, just as everything
else in Bibulus.

> Then take your example from above
>
> <city country="de">München</city>
>
> I think the only thing that is missing is language in addition to
> country.  We can assume a default language based on the country
> attribute, however this is a little bit fragile.  More XMLish is
> probably to use English as default and read the xml:lang attribute
> if present.  So, a Russian who want to write Moscow in Cyrillic has
> to use xml:lang in <city> or its ancestors,

Agreed.  At the very least, people should specify xml:lang on the top
<bibliography> level.

> *or* they have to write "Moscow".

Personally, I dislike falling back to English.  It's bad enough that
tags have to be in a particular language.

> The underlying city database will contain as primary keys the
> American English and the native name; other combinations like
>
> <city country="de" xml:lang="it">Aquisgrana</city>
>
> are allowed but rare enough to accept the possible failure.

Plenty of cities wouldn't have American English variants.  Let me
propose another way to deal with it:

Each city has a primary language (yes, there are some bilingual cities
like Brussels where this is going to cause problems, but most users'll
never know), e.g., København/dk/da, Aachen/de/de, Roma/it/it, Baile
Átha Cliath/ie/ga, Warszawa/pl/pl.

Each language then has a list of correspondences associated:

da.pm:
  Baile Átha Cliath/ie/ga <-> Dublin
  Roma/it/it <-> Rom
  Malmö/se/sv <-> Malmø

de.pm:
  Baile Átha Cliath/ie/ga <-> Dublin
  Roma/it/it <-> Rom
  Warszawa/pl/pl <-> Waschau
  
it.pm:
  Aachen/de/de <-> Aquisgrana
  Baile Átha Cliath/ie/ga <-> Dublin
  Warszawa/pl/pl <-> Varsovia

If we now cite your example (<city country="de" xml:lang="it">
Aquisgrana</city>) in a Danish bibliography, the following steps would
happen:

1) Because the language of the XML entry is Italian, we look up
   Aquisgrana/de in it.pm.  It is found, at we thus now that our
   internal form is Aachen/de/de.

2) Since the new context is Danish, we look for Aachen/de/de in
   da.pm.  It is not found, so we keep it in the original and output
   'Aachen'.

Another example (with fewer explanations):

XML: <city country="ie" xml:lang="da">Dublin</city>
Internal: -> Baile Átha Cliath/ie/ga
German context: -> Dublin

Does that make sense?

> My father, editor of a journal, told me that the abbreviations are
> the same even in journals of different publishers (which would be
> enough).  I hope the same is true for the full names of the
> journals.  I will look for this on the Internet.

Thanks!

/Thomas
-- 
Thomas Widmann          Bye-bye to BibTeX: join the Bibulus project now!
address@hidden                                <http://www.bibulus.org>
Glasgow, Scotland, EU     <http://savannah.nongnu.org/projects/bibulus/>