Talk:WikiAuthors

From Meta, a Wikimedia project coordination wiki

In my vew, a major function of Wikiversity should be to help Wikimedia projects deal with sources/references. Maybe this project should merge with the existing Wikiversity proposal. --JWSurf 15:46, 25 January 2006 (UTC)[reply]

When this proposal is undertaken, it will function for whereever sources/references are needed. The Wikipedia project is one that is in need for a great sources functionality.. When WikiAuthors provides this, it is infrastructure for all projects that need it. GerardM 19:28, 25 January 2006 (UTC)[reply]

It is too narrow to limit such a project only to MEDLINE. Author name disambiguation is just as big a problem in any large bibliographic database, such as that of the Library of Congress [1] or Die Deutsche Bibliothek [2] (Germany). Instead of starting a new wiki for every catalog, you should include all catalogs in one big wiki project (if a wiki project is a good idea at all, perhaps with the support of Wikidata?). One benefit is that the same author will be represented in many of these catalogs. This project proposal is now in the category:books together with WikiAbstracts, WikiBib, WikiReviews, Wikicite, Wikireview, and Wiki Research Bibliography. --LA2 19:47, 25 January 2006 (UTC)[reply]

Wikicat[edit]

Just to second Lars's recommendation- this project should probably be folded into a more general one- namely Wikicat. From the research I've done so far it seems bibliographic catalogs intentionally use the author name as written, and then normalize all variants to point to a canonical representation through other mechanisms. My idea is for Wikicat to also record the author name as written, but then allow each such entry to be joined to the "actual" person as represented in an authors or persons Wikidata dataset.

Jleybov 00:13, 26 January 2006 (UTC)[reply]

All your base are belong to us[edit]

This All your base are belong to us is fine. It proves only one thing. The thing is that collaboration is possible because otherwise nothing comes of it. We already have four people discussing things, all claiming that there is this bigger framework and it is theirs.

I can make an equally good case to have this project be part of Wikispecies.. (Hey, I did my research :) ) And it is therefore important not to club each other on the head with more of one's TLA's. If only not to get a headache. There are several "standards" for persons and their publications, Brummitt and Powell have in their Authors of Plant Names (1992) one such standard.

At issue is that names are ambiguous. So the application that will serve us will be in need of associating people with names in many ways. It will be necessary to know who is meant when a certain standard is a given. One option is to have people sign on to the database and identify THEIR publication be it scientific or otherwise. Given the collaborative nature of such a project, it will bring us also an impressive list of publications.

A publications / authors / sources project is worth having. Important is that it means different things to different people. For an author it means that he wants to be associated with HIS publications. For a wikipedian it means that the article is sourced and the point can be researched. Important is the realisation that by thinking of a project like this with communities in mind will bring you multiple approaches. All equally valid, all contributing to a rich tapestry of data. GerardM 09:28, 26 January 2006 (UTC)[reply]

Well, if I'm not mistaken, Wikispecies is not currently considered a successful project[3] and is being recommended for complete re-implementation within Wikidata [4]. Even in the initial stages of a project's design decisions are made which subsequently define its course: "One option is to have people sign on to the database". What database? The "database" of wiki pages and their text (along the lines of Wikipedia's "List of" pages) or a Wikidata database? If the latter, will this database exist and reference datasets from other projects or be isolated?
It seems you agree that this should probably be implemented in Wikidata because a relational database is a known, proven solution for the storage and query requirements of WikiAuthors. The problem of disambiguating multiple forms of an author's name is also a known one with a standard solution called Authority control [5] and I see no reason why we should disregard it, or those projects which will be addressing this same issue.
Jleybov 20:26, 26 January 2006 (UTC)[reply]
Wikispecies would in my opinion be better served by using a relational database. I used it to point out that author / publication matching is indeed something that is NOT specific to any one project.
I would have it in no other way but in a Wikidata project. I am happy to consider any reasonable approach. When "authority control" works, fine. When we have to amend it to fit our needs, we should. When the data resulting from the authority control system is Freely available, we are crazy when we do not use it. When it is not Free, we may have no choise. GerardM 20:48, 26 January 2006 (UTC)[reply]

no merge with WikiCat[edit]

In my opinion there is a fundamental difference between WikiCat and Wiki-Authors as it was re-conceived by a recent workshop including the original proposer Miguel Andrade. Especially when we wish that the major publishers share their author datasets, we have to stay away from content storage. WikiCat can be very useful to collect documents not provided by publishers.

WikiCat and WikiAuthors start from a different premisse. WikiAuthors starts of with the notion that many publications have a problem in identifying it's author. There are massive programs underway to disambiguate authors, but these efforts are often limited in focus; fi only biomedical publications. WikiAuthors will start with biomedical publications but is open to any and all publications. WikiCat is really good in providing a framework which is more inclusive that the original WikiAuthors scheme as it ties in with the systematic ways of the world of librarians.
When WikiAuthors is provided with information to seed it's content, this will have a big difference on the up-take of the project. This information can be both be provided by the librarians world as by the publisher world. It is in my opinion better for the publishers when they provide this information. When they do not, they will be marginalised in this project and consequently their relevance will be deminished in the eyes of many.
Both WikiCat and WikiAuthors do not intent to replace the information that is in the book. When there is good information about a book available, this information can be referenced to. However, it must be clear that the data design that is needed for both projects is highly complementary and that it would be foolish not to prevent the duplication that arises from seperate projects. WikiAuthors starts of with the inventarisation of publications and the disambiguation of authors but all the information that WikiCat includes is relevant.
WikiAuthors is going to be build. WikiCat's functionality can be and should be made part of one project. Publishers will preferably adopt this project and when they don't the librarians will. The argument is on the timeline. GerardM 06:57, 2 September 2006 (UTC)[reply]


I would just like to add that Wikicat is designed to contain only bibliographic meta-data and not actual content, and thus there should be no legal issues with publishers (legal issues with open catalogs themselves is an entirely different matter). Indeed it would be superb if both projects could use publisher-generated bibliograhic meta-data, and I would be surprised if such a system does not already exist in one form or another.
And as Gerard said, both projects can benefit one another. The data imported into Wikicat can, for example, be used by the heuristic algorithms of WikiAuthors. Author disambiguation is a known issue in library science called authority control. Thus bibliographic data from competent catalog sources will typically provide enough data about an author to disambiguate him from others with similar names- i.e. J. Smith will be recorded as: John B. Smith, b. 1953- . Detailed journal/periodical data is typically not collected by any catalog, so we unforuntately cannot use authority-controlled records to disambiguate authors within most research literature. However, if John B. Smith co-authored a book on immunology with Jane L. Doe in 1997 and there is a paper on immunology with J. Smith and J. Doe as its principal authors published in 1995, chances are the authors of the paper and the book are identical and so we can disambiguate the journal paper authors by linking them to the unambigious/authority-controlled book authors. I'm sure there are other areas in which Wikicat and WikiAuthors will complement one another. Jleybov 23:57, 5 September 2006 (UTC)[reply]

I think we all agree then. Wikicat is related to WikiAuthors. They are different projects though. Accordingly, I have added a section on related projects at the bottom of the page with a link to Wikicat and removed the merge banner. I would not mind having a section explaining the relation between the projects. Of course, anybody feel free to add it. Miguel Andrade 19:09, 21 September 2006 (UTC)[reply]

Yes, this is fine. My original merge suggestion was based upon my understanding of the WikiAuthors project as detailed on this page. WikiAuthors is quite more extensive than currently indicated here, though, and it would be nice if the project page were updated to reflect its latest features.
Jleybov 00:17, 22 September 2006 (UTC)[reply]