[mb-users] Sortingorder Dutch tussenvoegsel

[Lauri Watts]

On Mon, Apr 21, 2008 at 8:01 AM, Sander van Zoest <sander at vanzoest.com> wrote:
> On Mon, Apr 21, 2008 at 12:52:40AM +0200, Fred Marchee wrote:
>  > > Perhaps it is also better to put a text in the wiki (that's where it
>  > > all started) that: "Last name, tussenvoegsel, First name" leads to the
>  > > same result as "Last name, First name tussenvoegsel" and that we have
>  > > to do with a sortingmachine working this out the right way. It helps
>  > > against stupid guys like me ;-)
>
>  I think what is going on here, is that the meaning for SortName seems to
>  have changed or it was never understood equally. I guess that is the beauty
>  of the style guide in that it allows everyone to come to the same conclusion.
>
>  It is clear that the reason behind "Groot, de, Boudewijn" was not fully
>  understood by many, as clearly indicated by <http://wiki.musicbrainz.org/SortNameStyleDiscussion>. The point was to have 3 separate columns that would be sorted in that order, but it seems to be pretty clear that we are moving away from that. If the comma separated parts are never treated as different entities than it is clear that you will never be able to sort using th SortName appropriately and you are better off using it as ArtistListDisplayName or something like that.

There's also your local settings to consider.

'Collation' (ie, sorting) order can be set independently of language
and character set on some operating systems, but defaults on all to
the language you are currently running in.  Add to this the fact that
not all software knows what to do with the sortname field.

If you are running software that doesn't actually use the sortname
field, the order of sorting then falls back to the collation order
your software uses, and that will change depending what language you
run it in.

Further, a lot of software carries it's _own_ rules for how to sort
dutch (and other) surnames, those rules may not be the same when the
software is running in Dutch or when it is running in English.

The order on the websites given as examples almost certainly depend on
the collation environment variable on the webserver they are running
on, and a lot of people are not even aware you can change that (both
as a default and on the fly for each page.)   So as examples, you then
have to consider 'are these serious greybeards who have debated this
for 23 years and come to the conclusion this is how you sort
tussenvoegels' vs 'are these a bunch of people who are being held
hostage by the fact they don't know how to change the LC_COLLATE
variable on a hosted linux webserver, and have decided to live with
the results'.

Not just webservers either.  Swedish actually has two collation
orders: One in common use, and one legacy one used mostly in
phonebooks, because the 'common use' one, if applied to phone listings
would have resulted in a whole lot of people being moved around, and
not being able to find themselves.  And the 'phone' order does cute
things like considering 'Karlsson' and 'Carlsson' as both starting
with K, and Sorting W into V.  There is no W in the Swedish alphabet,
the only place you find it, and therefore need to deal with it, is in
names; My name sorts near the start of the V listing in the phonebook.
 But the 'cheapy' throwaway phone directories sponsored by local
companies, tend to not have the production values of the real phone
book, and at least the one I have lying around here is not using the
phone sort order.  Someone else held hostage by their software
apparently put that together.

All of which doesn't mean we don't need to sort out which way around
this all goes, and having an actual standard that makes sense to dutch
speakers would be great.  We just have to remain aware that a lot of
software is going to utterly ignore it and sort everything as it
pleases anyway.

-- 
Lauri Watts