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

[Torsten Bronger]

Halloechen!

Thomas Widmann <address@hidden> writes:

> Torsten Bronger <address@hidden> writes:
>
>> Thomas Widmann <address@hidden> writes:
>>
>>> [...]
>>>
>>> 3) The <place> doesn't have any structure.  Would this be better:
>>>    <place>
>>>      <city>Boston</city>
>>>      <state>Massachusetts</state>
>>>      <country>USA</country>
>>>    </place>
>>>    ?
>>
>> I don't think so.  The question is whether somebody would need
>> that, e.g. "I want to see all entries with publishers from
>> Boston" or "I want to have the city printed in small caps".  But
>> in my opinion this is not necessary.
>
> Agreed.  But there could be a case for leaving out certain parts
> (e.g., I might include "USA" in the example above if publishing in
> Europe, but not if publishing over there).

Granted, but then I'd suggest to have only <address> (with #PCDATA
inside) and <country> (with optional ISO code attribute) in <place>.

> Also, some placenames have to be translated, e.g., Aachen should
> come out as Aix-la-Chapelle in French, and Glasgow as Glaschù in
> Gaelic.  This might be easier to do correctly if the structure is
> there.

This I don't like.  The city name is some sort of part of the
publisher's name and should not be translated (except for the case
if the publisher does so itself).

Moreover, this approach is fragile because it relies only on mere
string comparison (there are no codes for cities).  This means that
if two cities share the same name but only one has a foreign
translation, this fails.  The other way round, there are multiple
spellings for some cities:

Frankfurt                         New York
Frankfurt am Main                 New York City
Frankfurt a.M.                    New York City/New York
Frankfurt a. M.                   New York City (NY)
Frankfurt a.&#160;M.              ...
Frankfurt&#160;a.&thinspace;M.
...

Tschoe,
Torsten.

-- 
Torsten Bronger, aquisgrana, europa vetus