The End of "Genres"

[varnis]

I would like to pre-empt the request and sign up as part of the project team, it seems like a worthy and interesting endeavour and perhaps Ill discover some new music I never considered before

[Operalala]

This form of "Tagging" can work only if it is standardized, and modularized (like the current 3 fields in the workpage template: Genre, Instrumentation and Language).
1. Work type/Genre - This must be both logical and not infinitely large.
2. The Language for vocal works - no specialist vocabulary for each different language - no Lieder, Gesanaenge, Melodies, Chansons, Romanzy, Pesni, ad infinitum
3. A standardized set of instrumentation tags/categories

Standardizing these terms is where all the debate has been going on, whether it's catgories or tags.

[Davydov]

I would like to pre-empt the request and sign up as part of the project team, it seems like a worthy and interesting endeavour and perhaps Ill discover some new music I never considered before

Thanks, Varnis, and I'll be in touch when the time comes!

[steltz]

no specialist vocabulary for each different language - no Lieder

etc.

I would have to disagree with this -- for reasons we have discussed endlessly elsewhere. People will search with common terms.

[Davydov]

This form of "Tagging" can work only if it is standardized, and modularized

As I said before, the work will be done by a small group of volunteers following a series of logical guidelines There's clearly a demand for categories like "Lieder", which will be tagged as such only where this was the composer's original designation (so all songs in German will not automatically be assumed to be Lieder, for example). But people will also be able to go straight to the categories called "Songs" or "Compositions for voice and piano" and access all the songs, lieder, chansons, etc. together from one page; or if they specifically want "Lieder", they can go straight to a page for that as well. So I believe people will find it logical, standardized, and comprehensive (without being infinitely large).

[vinteuil]

Wow...lots happens while I'm away. Number of thoughts
1. This looks great. I would suggest some sort of way to prevent the capitalization and mispelling errors by way of drop-down menus, or sets of links or something.
2. @Steltz: a) Glad you noticed the star . b) The OP project has, IMHO, not been running smoothly, as there are currently 3 members who are uploading anything. So, while we are working hard, it could go better.
3. @Operalala: It is actually not terribly difficult to get at least 2 or 3 users together to do this. You'd be surprised (and could do with volunteering for one or two teams
4. @P.Davydov: I love the idea of your permanent team of people doing this. However, the major problem is that many, if not most of the people on the team will be people like the current CRs and admins, who are already doing projects like this, and are rather heavily loaded. So I'd like to see some interest aside from yourself, varnis, steltz, probably KGill, and myself.

[Operalala]

Pearlnerd, I'm more concerned that this be set up in a manner that will be self-sustaining and future-oriented, and not just suit somebody's desire to have the word "Lieder" to look up.
I would love to see drop-down menus, if that can be implemented on the work pages. This will require a finite and well-thought-out list of Genres/Work Types and Instrumentations (as well as Languages).
The problem we've had, both with the categories and with this tagging proposal is that that the Work Types have not been well-thought-out.

Steltz, I must ask you, have you ever searched for any vocal music other than German? How did you manage that, seeing that it was probably not labeled a Lied in any way.

[Davydov]

Hi Snailey! Who gave you permission to go away?

Wow...lots happens while I'm away. Number of thoughts
1. This looks great. I would suggest some sort of way to prevent the capitalization and mispelling errors by way of drop-down menus, or sets of links or something.

Typos will inevitably happen, but careful monitoring of the wanted categories page will enable them to be picked up quickly and fixed. I'm hopeful though that the basics are straightforward and our volunteers will do all this reliably, and that teething troubles can be ironed out quickly at the start. it would actually be easier to start with a smaller group of volunteers (i.e. no more than a dozen), and we can learn from that process when recruiting more volunteers. Thank you for putting yourself forward BTW!

I know we've all got other commitments, and tagging 20,000 works is going to be a marathon rather than a sprint. There won't be any artificial deadlines, so people will be able to work at their own pace.

[vinteuil]

Pearlnerd, I'm more concerned that this be set up in a manner that will be self-sustaining and future-oriented, and not just suit somebody's desire to have the word "Lieder" to look up.
I would love to see drop-down menus, if that can be implemented on the work pages. This will require a finite and well-thought-out list of Genres/Work Types and Instrumentations (as well as Languages).
The problem we've had, both with the categories and with this tagging proposal is that that the Work Types have not been well-thought-out.

Steltz, I must ask you, have you ever searched for any vocal music other than German? How did you manage that, seeing that it was probably not labeled a Lied in any way.

You know, I've never had any problems searching for music on IMSLP of any sort, possibly because we have such great manually maintained lists. If you're interested in doing one for Art Songs or Opera, that might help a lot of what you're thinking of.

And, P.Davydov, I'll be away again next week, so I guess I'm going to be seeing some hits on my annual performance review . But seriously, if you could do the featured scores, that would be great,

[imslp]

This seems like a hot topic tonight!

I haven't read Davydov's post extremely closely, but I do have one rather fundamental question about it. I was wondering: instead of having a rather small set of users doing all the work, would it be beneficial to make the system such that anyone can use it? The reason I ask this is because if anyone can use it, people may come up with tags that are useful, but are not done by that small group of users.

And a related question is, does such "other" tags exist at all? Or does a genre + instrumentation tag cover everything?

A third question is, what is the benefit of having one detailed tag (e.g. "Sonata for Violin and Piano") versus having two tags ("Sonata" and "Violin + Piano") and using category intersect on it? I mean, the detailed tag will need to be embedded into the larger categories anyway...

A fourth question is, what is the difference between this, and simply having the current category system blown open (i.e. letting the small group of people add categories where appropriate on a case by case basis, without changing anything else technically). Except for the entire recategorization of all the pages, the open category system, and the limited number of people who can change it, is there another difference?

Again, I reaffirm my support for this, but it is my duty to raise all possible questions to have them thought through before any major effort is made, so that such is not wasted. These questions are not to dismiss the usefulness of this idea.

@Davydov: I got your PM. I'm entirely for the idea, but I want to make sure that we know precisely what we are getting into (and the costs/benefits) before we actually start this major process.

[steltz]

Steltz, I must ask you, have you ever searched for any vocal music other than German? How did you manage that, seeing that it was probably not labeled a Lied in any way.

I have searched for many things, not just Lieder. Its just that it's the word that, under your system, would be deleted, when it's an incredibly common one.

I am very involved in my university library, and issues of searchability, browsability (which = usability) are things I deal with a lot. I am also extremely aware of what extremely knowledgeable users know to look for, as opposed to less knowledgeable students, as opposed to amateurs.

You simply cannot have one word be the only word that will bring up a successful search. People get frustrated that they aren't getting enough hits with the words they think of using, and they give up.

A third question is, what is the benefit of having one detailed tag (e.g. "Sonata for Violin and Piano") versus having two tags ("Sonata" and "Violin + Piano") and using category intersect on it? I mean, the detailed tag will need to be embedded into the larger categories anyway...

At the moment, the intersect looks to me to be more efficient in terms of fewer tags, and (off the top of my head), it might solve another problem we have currently. Right now, some of the solo sonatas with continuo are in the "Sonata (piano with 1 instrument)" category. While generally correct (I was eventually going to suggest "keyboard" rather (but this idea's better)), the intersect idea would mean that a Baroque specialist could look for "Sonata" + "Violin" + "Continuo" and not have to wade through the Franck or Brahms sonatas to get to the ones he wants. We will eventually have potentially thousands in each category, so, thinking to the future, people need to be able hone their searches.

From an earlier post, there was something about not putting the number of instruments in so the tags work properly. I feel the numbers have to go somewhere, because people may not want to download lots of scores to find the chamber works that use, for example 2 flutes instead of 1 or 3.

Am I correct that the instrumentation field would still give instrument numbers, e.g.:

tag: sonata
tag: violin
tag: piano
tag: 20th Century
instrumentation: 2 violins, piano

for Milhaud's Sonata for 2 violins and piano
(Bad example, since it's not PD)

[Melodia]

I haven't been following this thread much, but a proper tagging system would allow for searches with multiple tags includive as needed.
It would also allow for easy single click listing of everything under than tag.
It's also a VERY VERY good idea.

[Davydov]

Hello Feldmahler,

Thanks for your latest questions, which I'll do my best to answer:

1. Should anyone be able to do the tagging?. In an ideal world it would be, but because it's so important that the syntax/capitalisation of the tags is correct, we need to make sure this is done consistently. Drop-down lists would not be a solution because certain headings should only be used in certain circumstances (as will become clear when I've finished the detailed guidance for project members). A comparison would be like letting the users of a library move books to any shelves they liked, and you'd inevitably get arguments and idiosyncratic decisions based on personal preferences. So I'm advocating that a small group of us as "librarians", taking care of the cataloguing and shelving (while still listening to feedback from our readers).

2. The benefit of the "bottom-up" approach is that all the existing "other" categories will no longer be needed. If we have work(s) for a particular grouping of instruments, then there will automatically be a category for it. If not, there won't.

3. As for having one tag instead of two (reflecting the work type and instrumentation), this is actually crucial. To take your example, using a separate tag for "Sonatas" would produce a list of hundreds of works in that category, sorted alphabetically by title, which isn't particularly useful. But using the one-tag system, all works in the Sonatas category will be conveniently grouped under their sub-categories of "Sonatas for violin and piano", "Sonatas for piano", "Sonatas for violin and continuo", etc. (Because of the 'bottom-up' principle only the categories for which works exist will be used, so there will be no redundancy.) The same thing would apply on the instrumentation field, so that "Compositions for violin and piano" would have a sub-category for "Sonatas for violin and piano", differentiating them from other types of work for the same instruments.

The intersects would then become a much more powerful tool, as users could search for "Beethoven"+"Sonatas for piano", rather than just "Beethoven"+"Sonatas"; or (taking Steltz's example), "Sonatas for 2 violins and piano"+"20th century".

4. While there could be a limited number of special cases where individual tags could be added at the discretion of the project team member (e.g. "National anthems"), for the reasons stated above it's essential that all the IMSLP "librarians" follow the same "one tag" cataloging system.

I hope that all makes sense, and feel free to get back to me if you have any more questions.

[Melodia]

Er, maybe I'm missing something, but if you're using TAGS (as opposed to the catagories that MeidaWiki uses) then you WOULD want a "sonata" tag that lists everything. But you could ALSO search for all three tags of "sonata" "violin" and "beethoven" all at once, and would get a nice small list. This is the ideal way of doing things, as it allows for things that may be a bit on the odd side, rather than forcing everything to be shoehorned in (which is why the the whole 'genre' concept can often get annoying)

[Davydov]

Ah, sorry, that's not quite what we're talking about here. The "tags" in this case map to wiki categories, and I don't think the tagging system you described is possible with the current Mediawiki software. I'd definitely agree that the problem with the present genre system is that it's tries to artificially fit works into categories where they don't belong, which is why this different approach should be a huge improvement.