Sorting compound names

[jda]

>>Well, here important to note.
>>
>>Index By words -- cannot be sorted.
>>
>>I did think that if last names always have
>>
>>     prefix Name
>>
>>Then if we ignore prefix then Name is the only word,
>>And we get simple Index String, which we can sort.
>
>Names is a messy business. Also prefix can be multiple words (like 
>in 'van der Waals') but that would not cause trouble as both can be 
>excluded.
>
>Since Jon does not need to sort the whole database but only the 
>subset selected by the user, he might be better of to build a 
>sortable list in his environment, excluding the prefixes as 
>specified, and sort them outside Valentina.
>
>At least this is what I am planning to do for the publication 
>database I am preparing to work on myself.
>

But as always, things are not as simple as they might appear.

In sorting a list of records for a bibliography, one may want to sort 
first by author (with or without considering name prefixes) and then 
by date -- in fact, this is quite common. What I output is not just 
the names, but a whole citation (e.g. Brenstein, R (2004) Posts to 
the Valentina list. 45:1-3). Each element I have listed here (author, 
date, title, volume, pages, and more is kept in a different indexed 
field and assembled to yield the citation above).

Being able to ignore words on a search would be cool, but what I need 
to do is to ignore them on a *sort*. For example, to yield this sort 
order (German sort, by author,date):

von Allen, G (2003) Whatever. 2:34-35.

Brenstein, R (2004) Posts to the Valentina list. 45:1-3

Note that von Allen sorts before Brenstein because von is ignored.

So, Ruslan, if you are saying it would be possible to specify a list 
of words to ignore in a sort, that would do the trick.


Jon