PDA

View Full Version : Implementing Case Insensitivity



Dondi
2005-05-25, 13:09
Currently running May 20th version of the nightlies and (not sure if this bug existed prior or not) but the # of Artists is influenced by capitalization in this version for sure.

a-ha, A-Ha, A-Ha are listed as 3 separate artists as opposed to the same artist (Same for artists such as Ani Difranco vs Ani DiFranco, so it isnt the hyphen in the a-ha example).

Is there, or could there be, an option to IGNORE CAPITALIZATION/CASE??

This would be of great help for those with large libraries :)

I also have the server set to rescan every day at 11am..... I added ~2,500 new tunes and the last 2 rescans have not picked up the new tunes -- I am still in the process of retagging my entire libray (~70,000 tunes) and I am up to the letter 'U' at the moment; Monday I added the entire 'T' folder which has yet to be included on the daily rescans

-- Dondi

John Hernandez
2005-05-25, 13:45
Dondi wrote:
> Currently running May 20th version of the nightlies and (not sure if
> this bug existed prior or not) but the # of Artists is influenced by
> capitalization in this version for sure.
>
> a-ha, A-Ha, A-Ha are listed as 3 separate artists as opposed to the
> same artist (Same for artists such as Ani Difranco vs Ani DiFranco, so
> it isnt the hyphen in the a-ha example).
>
> Is there, or could there be, an option to IGNORE CAPITALIZATION/CASE??
>

Such an option might fix counts (ie library statistics reporting),
assuming no counterexamples exist, but when it comes to actual Browse
and Search results, a proper implementation requires some interactive
mechanism to tell Slimserver which of the detected variants is the
correct one to display. I can think of some non-interactive
compromises, like defaulting to all caps, but these are all suboptimal.
The bottom line is that you can't just "ignore" the case, but rather
you want to justify or standardize case for output.

While you wait, consider running a tag editing tool that will help fix
your tags. I've seen some that claim to do case fixing, like
http://www.wma-mp3.com/mp3_tag_studio.html but never tried them. Even
if your library is large and haphazardly tagged, pecking away at it may
prove worthwhile.

Dondi
2005-05-25, 14:08
Such an option might fix counts (ie library statistics reporting),
assuming no counterexamples exist, but when it comes to actual Browse
and Search results, a proper implementation requires some interactive
mechanism to tell Slimserver which of the detected variants is the
correct one to display. I can think of some non-interactive
compromises, like defaulting to all caps, but these are all suboptimal.
The bottom line is that you can't just "ignore" the case, but rather
you want to justify or standardize case for output.

While you wait, consider running a tag editing tool that will help fix
your tags. I've seen some that claim to do case fixing, like
http://www.wma-mp3.com/mp3_tag_studio.html but never tried them. Even
if your library is large and haphazardly tagged, pecking away at it may
prove worthwhile.

I use Tag & Rename and there is an option for Capitalization of first word. However, since tagging my huge library has taken some time, Tag & Rename has since abandoned the ALL MUSIC service and instituted Amazon.com as the source. Once imported, there are variants on case etc. Examples that reflect the most frequency of case inconsistency are words like 'of', 'and', 'the'. On one album Dion and the Blmonts may be represented as 'Dion and the Belmonts' while the next album may be represented as 'Dion And the Belmonts' while another appears as 'Dion And The Belmonts'. In the interim I have been going back through the library and trying to standardize the case on the artists that have multiple-listings based on case sensitivity; I am up to 'A' so far :)

-- D

mherger
2005-05-25, 14:09
> a-ha, A-Ha, A-Ha are listed as 3 separate artists as opposed to the
> same artist (Same for artists such as Ani Difranco vs Ani DiFranco, so
> it isnt the hyphen in the a-ha example).

This has been discussed lengthy befor:
http://forums.slimdevices.com/showthread.php?t=14278

There are two parties:

a. what you see is what you have in you tags. Fix it using an appropriate
tool.
b. let's have slimserver do some magic and guess what you might want to see

b. will be hard to implement. Case sensitivity is only a start. Next we'll
have "Beatles" vs. "The Beatles", "Tony Levin" vs. "Tony Levin Band", "22
Pistepirkko" vs. "22-Pistepirkko", "Björk" vs. "Bjork" etc. Don't expect
magic to happen and do your homework :-). If you've got a big collection,
then it's even more important you do it correctly.

--

Michael

-----------------------------------------------------------
Help translate SlimServer by using the
StringEditor Plugin (http://www.herger.net/slim/)

Dondi
2005-05-25, 14:51
> a-ha, A-Ha, A-Ha are listed as 3 separate artists as opposed to the
> same artist (Same for artists such as Ani Difranco vs Ani DiFranco, so
> it isnt the hyphen in the a-ha example).

This has been discussed lengthy befor:
http://forums.slimdevices.com/showthread.php?t=14278

There are two parties:

a. what you see is what you have in you tags. Fix it using an appropriate
tool.
b. let's have slimserver do some magic and guess what you might want to see

b. will be hard to implement. Case sensitivity is only a start. Next we'll
have "Beatles" vs. "The Beatles", "Tony Levin" vs. "Tony Levin Band", "22
Pistepirkko" vs. "22-Pistepirkko", "Björk" vs. "Bjork" etc. Don't expect
magic to happen and do your homework :-). If you've got a big collection,
then it's even more important you do it correctly.

--

Michael

-----------------------------------------------------------
Help translate SlimServer by using the
StringEditor Plugin (http://www.herger.net/slim/)

I have been a long-time SLIM devices customer and because of my enjoyment of the devices that I have owned over the years, I started retagging my library over a year ago meticulously -- I have averted the article issue by making sure any band that uses the obligatory "THE" is wiped-out, i.e., "The Beatles", "The B52s" are, in my library as "Beatles" and "B52s". My post however had nothing to do with the articles, just case insensitivity.

After perusing the thread cited above, I must admit I concur with Phil Kerman's stance and confusion as to why anyone would want the listings of "Tower of Power" and "Tower Of Power" to be separate and unique artists. There were good points cited in the thread, yet I still see no argument as to why someone wouldnt want those 2 bands listed as the same. Someone asked, how would the server know which one is correct??? Can it take the first instance?? I dont care if the resultant artist listing is 'Adam and the Ants' as opposed to 'Adam And The Ants'.... BUT I do care alot that they are listed separately in the hierarchy AND that I have to sift back through my entire library to fix it so that the listing for a particular artist appears in a single location.

The articles situation is already an option... and an issue I have successfully averted the first time through my library (which I havent completed yet). I am weary from tagging, someone save me.

-- D

dean
2005-05-25, 14:56
The bug is open: http://bugs.slimdevices.com/show_bug.cgi?id=1548

and Vidur's looking into it.

Stay tuned.