IMSLP Forums

Hello,

Today we are announcing the launch of the IMSLP Red Links Project.

We currently require all contributors to works (arrangers, librettists, etc.) to be properly linked on pages when new submissions are made. But many pages on IMSLP have older edits where this is not the case, and where people have been linked without a profile being created, or where they were not linked properly at all. The goal of this project is to gradually correct all of this by providing a central list of people whose profiles need to be created/fixed and pages where these problems occur.

As the head of this project, it is my goal to make this task as easy as possible by providing a centralized coordination method, but ultimately we will need your help to add this information over time. Every bit helps. This initiative is very important as IMSLP moves towards using more structured data.

Please see the link above and we hope you can help!

Thanks!
DBM

A necessary project, indeed, DB. If you think IMSLP has today about 40,000 composers/arrangers/librettists/translators categories, only creating categories in this first phase of the project will increase that number by 25%. Its second phase seems to be much more laborious as each workpage would need to be edited manually (or maybe a bot could be designed to add {{LinkArr}}, {{LinkLib}}, {{LinkDed}}, etc, templates automatically to the respective lines when that information is filled in the pages). A (very long) third phase (why not) could be a complete revision of the scores to find out the information about personalities which have not been assigned. A single vocal collection can be linked to dozens of categories.

If we are talking about adding more than 10,000 new categories to the site, I think that the contributions to this project should be as homogeneous as possible. For instance, how we will use links to external sources, spelling dates of flourishing, adding pictures (20th century photos may still be protected), adding miscellaneous information (with sources!), etc. I'd also say that {{MoreInfo}} template should be always added to the categories when we don't have much information about a personality (that could help in the future to identify them). With that said, maybe a revision of the new categories would be necessary if less experienced users start to contribute more extensively.

I'd divide two kinks of new categories to be created: one of those people who are still alive (compositions/editions published through a CC licence) and another of those who are already in the public domain. Regarding this first group, what about an automatic email to be sent to them if they would like to provide the information needed to the pages?

Regarding the second group, I call the attention to the value of IMSLP as an authority control reference worldwide and how the quality and completeness of the pages are important to this recognition. It's fundamental that the categories (whenever possible) have cross-references with other authority projects: VIAF (above all!), RISM, Worldcat, DNB, ISNI, MusicSack, LOC, BnF, etc. We already have templates for all of them.

As RISM database has already been incorporated into VIAF, I don't see in the near future the same thing happening with IMSLP, but I'm sure that one day it'll happen. (Just out of curiosity, RISM has 144,670 personalities indexed - and that taking into account that they receive much of the information already processed from a lot of libraries/archives).

I lost count of how many categories I created and edited in the last years, but I must say that what takes more time is finding the correct VIAF, RISM, Worldcat, etc, identifiers. I need to mention that all data source from VIAF is licensed through Open Data Commons Attribution License (https://viaf.org/viaf/data/) and I'd be fabulous if this database could be used in the process of creating new categories, not only to link the page to VIAF but also to automatically extract the biographical information when available. The first step to create a new category would be finding a specific name in a list or autocomplete field based on the VIAF database, only if the match was not found we'd create the category manually. Musical references of the VIAF database could be extracted based only on those entries with cross-reference with RISM.

Thanks,

Thanks Feduol.

Phase 2 will be helped a lot by automation, don't worry. The details of the bot are something I'll get into more in the future. There will definitely be manual work to do, but a lot can be done at least semi-automatically by matching plain text with existing categories (and asking for a simple confirmation).

For VIAF/authority links, what should be done is adding the category to Wikidata (which IMSLP will check at certain intervals to link to all these sites). I also have a script I will plan to use to do this en masse by checking for categories not linked to Wikidata and seeing if there are correspondingly-named Wikidata entries lacking an IMSLP entry.

For the categories, it is difficult to actually sort them automatically into categories, as the point of the program to generate this list was to parse the list of broken links (whereas that would require a lot more decisive analysis of pages). But when you find a user who is an IMSLP contributor, you can email them to ask for biographical details.

Also, please note the existence of the following new templates (optional, not required, which are used to provide structured data to the new tool APIs):

- {{LinkPerson|FirstName|Lastname|born|died|disambig}} (3, 4, 5 optional). Links to a person just like {{LinkLib}} etc. but does not add the page to a category, just applying the same formatting. (also note the 5th option is a disambiguator, which we now use for categories of people with the same names as others; the disambiguator is shown correctly in all formatting and the new TS API understands it as separate from the name itself; an example of one of these is "Shakespeare, William @tenor^" (in internal MediaWiki page name), which is displayed as "William Shakespeare (tenor)" or "Shakespeare, William (tenor)" depending on the context in IMSLP.

- {{AltNames}} - used to provide alternate names in a structured form; displays as a lit. Used in the form {{Altnames|<name>|<name>|<...>}}

- {{FN}} - used to format names in {{AltNames}} and indicate properly what is the first and last name, in the form {{FN|First|Last|disambig tag}}.

See for example Mozart's page for a proper use of AltNames and FN.

Thanks for your explanation.

I'll keep in mind the new templates you mentioned. Just two more issues:

. {{LinkPerson}} seems to be similar to {{LinkName}}. I saw your editions on {{LinkName}} page but I must say I didn't understand everything completely. Now I’m a little confused when which one should be used. I wonder if this would not be the case to have just one template for the same purpose.

. Great idea with {{AltNames}} and {{FN}} to have alternative names structured and readable to the system. I'll use them from now on. Just to mention that over the years how the names have been inserted in |Alternate Names= lines is almost an aleatory process (family name before the name, after, commas, semicolons, dashes, etc...).