Page 1 of 1

Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 3:55 am
by jdeperi
The composer W?adys?aw ?ele?ski (http://imslp.org/wiki/Category:%C5%BBel ... ys%C5%82aw) is devilishly hard to find. (And just look at that URL!) One has to search for him using the letter Z with the appropriate diacritic. Of course, he's not even browsable because his last name does not start with a Z without diacritic.

What is the best-practices way of handling diacritics? It's an issue not only with composers but with works as well (cf. viewtopic.php?f=2&t=2124&p=11137&hilit= ... ics#p11137, for example). ?ele?ski's works, for one, are practically impossible to find unless one is both aware of the precise orthography and able to reproduce it for entry into one's web browser.

Jonathan

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 4:00 am
by jdeperi
Is there a way to configure Mediawiki to perform searches in a diacritics-insensitive manner? Google does this. It's the only way I can conceive would solve the problem entirely.

Jonathan

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 4:40 am
by horndude77
Try these two searches out:
http://www.google.com/search?q=site%3Ah ... e%C5%84ski
http://www.google.com/search?q=site%3Ah ... g+Zelenski

I was hoping that google would be a good workaround, but it's not (unless there are some weird options to make it work).

Searching for Dvorak seems to work ok though:
http://www.google.com/search?q=site%3Ah ... org+Dvorak
http://www.google.com/search?q=site%3Ah ... %99%C3%A1k

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 6:06 am
by jdeperi
The Google search works for Zelenski as well as Dvorak. If plain Zelenski is used, some results (including the one at the very top) have the plain-Z spelling, but some have the spelling with diacritics.

At any rate, I still consider it a serious problem that searching for plain Zelenski on the Project Petrucci wiki itself doesn't show any of the composer's works. I even went so far as to edit the composer's category page, http://imslp.org/wiki/Category:%C5%BBel ... ys%C5%82aw, to include the following:

Code: Select all

|Alternate Names=Zelenski, Wladyslaw
And note what was already there:

Code: Select all

{{DEFAULTSORT:Zelenski, Wladyslaw}}
I don't know enough about the underlying implementation (of Mediawiki itself and of the Project Petrucci-specific layer on top) to understand exactly what these directives do.

Jonathan

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 8:30 pm
by Lyle Neff
This is a problem that I've brought up before.

The only thing I know of to do for now is to create redirect page that has a diacritics-free version of the composer's name and redirects to the actual page.

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 9:25 pm
by Davydov
Using the {{DEFAULTSORT}} tag serves two purposes:
1) To make sure the composer appears in the right place in alphabetical lists
2) To allow for searches on names without diacrtics, which will pick up on the form shown in the tag.

For example, searching on "Zelenski" (using the "Comp[oser]" button*** by the search box) will lead you to "Category:?ele?ski, W?adys?aw". This is because the last line of the page includes the text {{DEFAULTSORT:Zelenski, Wladyslaw}}, which is normally invisible.

I added quite a few of these {{DEFAULTSORT}} tags a few months ago, but since then other composers have been added with diacritics in their names where tags may be required.

*** Note that if you use the "imslp.org" button, or just press enter when searching for a composer, this won't work properly.

Re: Difficulties with diacritics in composer names

Posted: Sun Jun 28, 2009 10:58 pm
by jdeperi
Aha! I had no idea there was a distinction between a composer search and a general search (obtainable by hitting enter instead of clicking on one of the search buttons). Ever since I started using the IMSLP (a couple of years ago!), I could never understand why searching for e.g. "Bach" didn't show the composer's page at the top of the results. This had always tripped me up...

In light of my experience, I propose a couple of ideas:
[*] How about changing the label of the composer search button from "Comp." to "Composer"? The button has a layout row to itself, so why not make it clearer what it's for?
[*] Why not make it so that searches entered by pressing Enter show both 1) composer search results, then 2) musical work search results. That way if a user does a general search, but of a composer's name, then the composer's page will show up at the top of the results with individual works right below.

What do you think?

Jonathan

Re: Difficulties with diacritics in composer names

Posted: Mon Jun 29, 2009 5:23 am
by Davydov
jdeperi wrote:Aha! I had no idea there was a distinction between a composer search and a general search (obtainable by hitting enter instead of clicking on one of the search buttons). Ever since I started using the IMSLP (a couple of years ago!), I could never understand why searching for e.g. "Bach" didn't show the composer's page at the top of the results. This had always tripped me up...
Yes, it was a long time before I found out about it as well. Before that I'd always assumed that "Comp." stood for "Composition"!
jdeperi wrote: In light of my experience, I propose a couple of ideas:
[*] How about changing the label of the composer search button from "Comp." to "Composer"? The button has a layout row to itself, so why not make it clearer what it's for?
On my display the "Comp." button is to the right of the "IMSLP" one, and there wouldn't be enough space there to expand the text. Maybe screen resolution is a factor?
jdeperi wrote: [*] Why not make it so that searches entered by pressing Enter show both 1) composer search results, then 2) musical work search results. That way if a user does a general search, but of a composer's name, then the composer's page will show up at the top of the results with individual works right below.
I understand that Feldmahler is working on an uograde to the site, so maybe that's something he will take into consideration (if it's technically possible).