Both the latest (0.2.1) and prior version of Album List Panel incorrectly sorts entries that contain accented characters.
-- To illustrate --
A Columns UI list shows the following word sequence correctly sorted:
Albert
Álvaro
Alvorada
Azerbaidjan
The same sequence would appear on a tree view as:
Albert
Alvorada
Azerbaidjan
Álvaro
(Note that Álvaro appears last, which is not correct).
The problem occurs with every accented character I tested: they're evaluated in such a way that they appear at the end, after all entries that have no accents.
Full Version: Album list panel (foo_uie_albumlist) - Sort problem
Hi,
I fixed it in version 0.2.2. Thanks for the report.
I fixed it in version 0.2.2. Thanks for the report.
Hi! 0.2.2 is indeed sorting right.
Another issue:
For classical composers, I subdivide the tree so that the first level shows only the initial letter of composers' names.
I note that 0.2.2 regards accented letters as if they were different from the non-accented ones, so that separate entries result for A, Á and  for instance.
Although that's not difficult to circumvent (I'm using $replace($substr(%composer%,1,1),Á,A,Â,A) to fix the problem), perhaps this behaviour could be addressed too.
Going further:
Some letters with diacritical marks should not imo be merged with their akins without diacritical marks; for instance Č (C with caron) and Š (S with caron).
For us western-european language users, such letters serve solely for the transliteration (according to ISO 9) of words from the Cyrillic to the Roman alphabet; since there are no Roman equivalent to them, they should be treated as different letters, not as accented letters.
----
p.s. For those interested: using the ISO recommendation for transliteration, Šostakovič stands for Шостакович, instead of the more popular Shostakovich (English), Schostakovitsch (German) or Chostakovich (French); transliteration makes it possible to revert to the original Cyrillic exactly.
Another issue:
For classical composers, I subdivide the tree so that the first level shows only the initial letter of composers' names.
I note that 0.2.2 regards accented letters as if they were different from the non-accented ones, so that separate entries result for A, Á and  for instance.
Although that's not difficult to circumvent (I'm using $replace($substr(%composer%,1,1),Á,A,Â,A) to fix the problem), perhaps this behaviour could be addressed too.
Going further:
Some letters with diacritical marks should not imo be merged with their akins without diacritical marks; for instance Č (C with caron) and Š (S with caron).
For us western-european language users, such letters serve solely for the transliteration (according to ISO 9) of words from the Cyrillic to the Roman alphabet; since there are no Roman equivalent to them, they should be treated as different letters, not as accented letters.
----
p.s. For those interested: using the ISO recommendation for transliteration, Šostakovič stands for Шостакович, instead of the more popular Shostakovich (English), Schostakovitsch (German) or Chostakovich (French); transliteration makes it possible to revert to the original Cyrillic exactly.
QUOTE(pstrg @ Jun 9 2007, 23:35) 
p.s. For those interested: using the ISO recommendation for transliteration, Šostakovič stands for Шостакович, instead of the more popular Shostakovich (English), Schostakovitsch (German) or Chostakovich (French); transliteration makes it possible to revert to the original Cyrillic exactly.
Offtopic and I'm not even sure how this stuff works internationally, but the Hungarian grammar rules say that a transliteration from Cyrillic should be written exactly as the pronounced word sounds (e.g. Sosztakovics)
Also, in Hungarian A and Á are completely different letters, so it would be strange and illogical to group them together by default. Especially when it comes to things like O, Ó, Ö and Ő
I had one native English speaker ask me why we need so many different Os...but these are not different Os but different letters. Also, S, Z, SZ and ZS are completely different and are considered single letters, even if they consist of two characters. In this case it would be logical for a Hungarian to group S separately from SZ, but I guess this would be strange and useless for an English speaker, for example.Sorry, I'm too tired for proper linguistics
But I guess you get my point...
QUOTE(kalmark @ Jun 10 2007, 09:33)
Offtopic and I'm not even sure how this stuff works internationally, but the Hungarian grammar rules say that a transliteration from Cyrillic should be written exactly as the pronounced word sounds (e.g. Sosztakovics)
Also, in Hungarian A and Á are completely different letters, so it would be strange and illogical to group them together by default. Especially when it comes to things like O, Ó, Ö and Ő
I had one native English speaker ask me why we need so many different Os...but these are not different Os but different letters. Also, S, Z, SZ and ZS are completely different and are considered single letters, even if they consist of two characters. In this case it would be logical for a Hungarian to group S separately from SZ, but I guess this would be strange and useless for an English speaker, for example.Sorry, I'm too tired for proper linguistics
But I guess you get my point... Interesting.
Not entirely off-topic, since it deals with fb2k components sorting patterns...
The Hungarian example you gave stresses the reason why ISO-9 was created - it is supranational.
It recommends a unique, biunivocal spelling instead of a phonetic transcription of a word from a different alphabet, which spells differently in Hungarian, German, French etc.
In most languages letters with diacriticals are not consider as different and dictionaries for instance ignore accents when sorting; take that with a grain of salt for I'm not familiar with East-European languages and less so with Hungarian (which not even Indo-European is).
In summa, with 0.2.2
sequential sorting is fixed
and
it is possible to decide whether first letters with and without diacriticals should appear alone or together by means of a simple expression.
QUOTE(pstrg @ Jun 9 2007, 22:35)
Another issue:
For classical composers, I subdivide the tree so that the first level shows only the initial letter of composers' names.
I note that 0.2.2 regards accented letters as if they were different from the non-accented ones, so that separate entries result for A, Á and  for instance.
Although that's not difficult to circumvent (I'm using $replace($substr(%composer%,1,1),Á,A,Â,A) to fix the problem), perhaps this behaviour could be addressed too.
Going further:
Some letters with diacritical marks should not imo be merged with their akins without diacritical marks; for instance Č (C with caron) and Š (S with caron).
For us western-european language users, such letters serve solely for the transliteration (according to ISO 9) of words from the Cyrillic to the Roman alphabet; since there are no Roman equivalent to them, they should be treated as different letters, not as accented letters.
For classical composers, I subdivide the tree so that the first level shows only the initial letter of composers' names.
I note that 0.2.2 regards accented letters as if they were different from the non-accented ones, so that separate entries result for A, Á and  for instance.
Although that's not difficult to circumvent (I'm using $replace($substr(%composer%,1,1),Á,A,Â,A) to fix the problem), perhaps this behaviour could be addressed too.
Going further:
Some letters with diacritical marks should not imo be merged with their akins without diacritical marks; for instance Č (C with caron) and Š (S with caron).
For us western-european language users, such letters serve solely for the transliteration (according to ISO 9) of words from the Cyrillic to the Roman alphabet; since there are no Roman equivalent to them, they should be treated as different letters, not as accented letters.
Hi,
I am not really going to change this for several reasons. It is basically difficult and I didn't write album list so am not familiar with its inner workings. I also can't see much use for it beyond the example of listing tracks by first letters of some field. There's also issues of what to merge fields into (e.g. {eé,ée}) and how to work this out and potential language differences etc etc....
QUOTE(musicmusic @ Jun 11 2007, 12:10)
Hi,
I am not really going to change this for several reasons. It is basically difficult and I didn't write album list so am not familiar with its inner workings. I also can't see much use for it beyond the example of listing tracks by first letters of some field. There's also issues of what to merge fields into (e.g. {eé,ée}) and how to work this out and potential language differences etc etc....
I think it's really better not to.
Special uses like that may be obtained by using buit-in functions.
Incidentally, compliments for the very quick correction that resulted in 0.2.2!
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.