#686908 Wrong indexterm sorting order for Norwegian Bokmål (nb)

#686908#5
Date:
2012-09-07 04:17:25 UTC
From:
To:
I ran into this problem while using docbook-xsl and fop (and also
dblatex) to generate a PDF for the translated version of the Free
Culture book by Lawrence Lessig.  See
<URL: https://github.com/petterreinholdtsen/free-culture-lessig/> for the
complete source of that project.

The indexterm sorting order is wrong for the nb language choice when
using docbook-xsl and dblatex.  See the included docbook file for an
example demonstrating the sorting, and the correct sorting order.  This
sorting order make both docbook-xsl and dblatex unfit for typesetting a
Norwegian book. :(

I've already reported a similar problem with dblatex, so this bug report
is about the docbook-xsl manifestation. :)

<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook V4.5//EN"
"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd">
<!--

The nb indexterm sorting order should be "Alle inne", "Bare tull",
"Ærlig talt", "Øvrigheta", "Åndalsnes" under separete letter "A", "B",
"Æ", "Ø" and "Å".  But docbook-xsl sort it

Symboler:
Ærlig talt

A
Alle inne
Åndalsnes

B
Bare tull

O
Øvrigheten

while dblatex sort it like this (note the underscore as the "letter"
for symbols):

_
Ærlig talt
Øvrigheten
Åndalsnes

A
Alle inne

B
Bare tull

Both sorting orders are wrong.

#686908#8
Date:
2012-09-07 16:01:53 UTC
From:
To:
Am Freitag, den 07.09.2012, 06:17 +0200 schrieb Petter Reinholdtsen:

This might be due
to /usr/share/xml/docbook/stylesheet/docbook-xsl/common/nb.xml not
customizing normalize.sort.{input,output}
from /usr/share/xml/docbook/stylesheet/docbook-xsl/common/en.xml, so it
maps "Æ" and "Å" to "A" and "Ø" to "O". I saw, that you customize
nb.xml. Can you please add the above variables, customize the mapping
and check, if that works?

Regards, Daniel

#686908#13
Date:
2012-09-07 18:55:41 UTC
From:
To:
[Daniel Leidert]

I tried, and it did not work.  I tried the attached patch, and got the
(more correct) 'Register' string shown verifying that the correct file
was edited, but 'åndsverklov' is still sorted as 'A' and not 'Å'. :)

#686908#22
Date:
2013-05-26 13:30:34 UTC
From:
To:
Hi,

I was playing a bit with the issue you reported. I found this:
http://www.sagehill.net/docbookxsl/IndexIntl.html

Looks like it's not that easy to solve on the maintainers  level and
probably not a bug in docbook-xsl. Please raise your question on a
docbook* mailing list to get a solution for it. In the current situation
I'm in favor of treating your report as not-a-bug, closing it soon.

Regards, Daniel

#686908#25
Date:
2013-05-26 13:30:34 UTC
From:
To:
Hi,

I was playing a bit with the issue you reported. I found this:
http://www.sagehill.net/docbookxsl/IndexIntl.html

Looks like it's not that easy to solve on the maintainers  level and
probably not a bug in docbook-xsl. Please raise your question on a
docbook* mailing list to get a solution for it. In the current situation
I'm in favor of treating your report as not-a-bug, closing it soon.

Regards, Daniel

#686908#30
Date:
2013-05-26 14:19:00 UTC
From:
To:
[Daniel Leidert]

You lost me there.  Can you explain a bit more?  Where do you believe
it is a bug, if it isn't in docbook-xsl?  If it is a bug somewhere
else, where should this issue be reassigned (closing seem like the
wrong approach, if you believe it is filed against the wrong package).

#686908#35
Date:
2013-05-26 15:11:09 UTC
From:
To:
Am Sonntag, den 26.05.2013, 16:19 +0200 schrieb Petter Reinholdtsen:

I meant to say, that is isn't a bug at all. From my understanding of the
page cited: The index sorting you get is a _non-configurable_ feature of
docbook-xsl, called the "basic" index sorting mechanism (check out the
index.method parameter). The stylesheets know two more sorting
mechanisms as described on this page. However you have to write the
necessary customization templates yourself.

So from my point of view, there is nothing I can do for you as
maintainer of docbook-xsl. Best I can do is to send you to the
docbook-apps mailing list, where people can probably help you.

That's why I'm in favor of closing your report.

Regards, Daniel

#686908#38
Date:
2013-05-26 15:11:09 UTC
From:
To:
Am Sonntag, den 26.05.2013, 16:19 +0200 schrieb Petter Reinholdtsen:

I meant to say, that is isn't a bug at all. From my understanding of the
page cited: The index sorting you get is a _non-configurable_ feature of
docbook-xsl, called the "basic" index sorting mechanism (check out the
index.method parameter). The stylesheets know two more sorting
mechanisms as described on this page. However you have to write the
necessary customization templates yourself.

So from my point of view, there is nothing I can do for you as
maintainer of docbook-xsl. Best I can do is to send you to the
docbook-apps mailing list, where people can probably help you.

That's why I'm in favor of closing your report.

Regards, Daniel

#686908#43
Date:
2013-05-26 16:38:19 UTC
From:
To:
[Daniel Leidert]

Very nice to get the web reference.  Had a closer look and it seem to
me like the kosek input method, as defined on the web page, support
all the features needed, but the
/usr/share/xml/docbook/stylesheet/docbook-xsl/common/nb.xml file
contain the wrong ordering in the <l:letters> block.  I did no quite
understand why this ordering method could not be enabled by default
for the norwegian locales.  Can you explain why?  I would expect it to
be possible to implement the docbook rules to behave correctly for any
language in the world.

Right.  I hope you are wrong, based on my observation above, but if I
am wrong I understand that docbook is not flexible enough to provide
correct index sorting for Norwegian Bokmål (nb) and Nynorsk (nn) out
of the box, and that there is nothing you can do about it.

#686908#48
Date:
2015-07-20 06:50:45 UTC
From:
To:
[Petter Reinholdtsen]

Hi.  The index sorting order problem still exist in version
1.78.1+dfsg-1, and I am not sure how to try to fix it.  Did I test the
wrong thing?