[edit] Wikimedia should definitely encourage building this

Main point is that we already need such a tool for the existing Wikimedia projects and no other site is currently able to provide freely licensed catalogue data. Some are trying or making promises to a smaller extend. I would rather no recommend a classic "formal" catalog structure but rather experimentation with the Semantic MediaWiki extension in order to make that content machine readable.

Mathias

By "formal" I mean to integrate existing standards and notations established by dead-tree libaries, for interoperability. I see this as complementing the tagging/folksonomy approach. And/or this could be linked up to wikipedia categories; actually, thinking about it, this is also a candidate for the omegawiki based multilingual categorization scheman proposed form commons[1].

Anyway, not sure if semantic mediawiki would be the right choice, wikidata seems to be to better candidate. Wikicat already proposes that, but it is (as far as I see) missing other aspects, such as being usable as a personal research tool, that is, allow me to collect suff I need for a specific purpose in one place. Which would be a good motivation for people to use this thing.

Anyway, I think it would be good to make more people aware of the Wikicat proposal (which is big and pretty technical - it's a bit intimidating, I have only skimmed it so far myself). Perhaps it can be revived (the page history shows little activity since mid-2006). -- Daniel 15:31, 20 January 2008 (CET)

[edit] Zotero?

It is the closest thing I can think of. AT least it's free and fairly portable and personalizable. Now if it had that citation markup done and released, it would make MANY people happy. 24.201.115.9 01:40, 21 January 2008 (CET)

but that does not have any collaborative features, does it? Combining it with wp:Bibster could be interesting, but it's still not a collaboratively edited library catalogue. Also, I'm looking for something web-centric, something useful not only as a specialized tool for the dedicated user, but also as The Place to casually look for bibliography info. -- Daniel 14:20, 21 January 2008 (CET)

Zotero does not offer many of the features but is a high candidate from my point of view. Collaborative support is on the list of top features to be implemented (myself is also working on this). Of the existing Systems LibraryThing is the best match. I'd also count on BibSonomy - they are very quick in adding new features and open for collaborations. Anyway: without additional coding there is nothing that really suits your (any my!) needs so far. -- JakobVoss

right... maybe you know someone who would pay me to do it :) -- Daniel 11:23, 8 February 2008 (CET)

that depends on what "it" exactely includes ;-) -- JakobVoss 18:51, 12 February 2008 (CET)

Wenn ich mir die anderen Lösung in Sachen Bibliographieren anschaue, kann ich Jakob nur zustimmen, dass Zotero die Alternative mit den besten Möglichkeiten ist. Ausbaufähig scheint sie ja auch zu sein, insbesondere in Sachen Kompatibilität zu anderen Systemen wie LibraryThing und CiteULike. Und die Erweiterung um die sog. web 2.0-Komponenten ist offenbar in den Startlöchern. Ralf.

I guess I'll give Zotero another try - though I couldn't find out how to import my stuff from citeulike right away (only individual entries). I guess I'll have download it as bibtex to my computer first. To be honest, working with an "unconnected" local application ofr this feels... strange. The UI possibilities are much better than on the web, of course. But it feels so... solitary.

Anyway - I created an Bibsonomy account yesterday. I'm in the process of moving my CiteULike and Del.icio.us stuff there, it seems to have the potential to replace both, we will see. The UI is much better than at citeulike, anyway. But still, I am missing citation graphs, and the wiki aspect. Bibsonomy with a dash of Citeseer and a bit of WikiData would be nice - and the whole thing tied in neatly with Zotero. -- Daniel 10:46, 10 February 2008 (CET)

What citation graphs do you refer to? Can one somehow access the citation data that CiteULike uses? Maybe we can start with data from CiteSeer?

I don't think citeulike uses any citation data (despite the name). The only large repository of that kind of data i know of is indeed citeseer. No idea how they feel about giving out their data for others to use...

Scraping citation data from PDF should be quite doable, and from tex it should be trivial. It would also be really cool if a microformat for bibliographical data would emerge. -- Daniel 00:58, 13 February 2008 (CET)

addendum wrt microformats: http://microformats.org/wiki/citation -- Daniel 00:50, 14 February 2008 (CET)

Nun gut ... ich bin deinem Beispiel gefolgt und habe ebenfalls einen Bibsonomy-Account erstellen lassen. Der Import der Zotero-BibTeX-Daten war nicht 100%ig befriedigend. Ein paar Felder sind nicht übernommen worden und die ursprünglich durch Kommata getrennten Tags wurden als ein kompletter Tag (inkl. der Kommata innerhalb der Zeichenkette) importiert. Da die Bibsonomy-Funktion, Tags zu editieren, sich offenbar nur auf den Ersatz des gesamten Tags bezieht (folglich der fehlerhaft importierte Tag "enzyklopädie,wörterbuch" entweder nur in "enzyklopädie" oder "wörterbuch" geändert werden kann) ... etwas mühsam. Natürlich ist das Ablegen privater PDFs nicht zu unterschätzen ... aber ich verharre erstmal in abwartender Haltung bei Zotero und hoffe, dass sich dort bald etwas tut. -- Ralf.

Den Delimiter für tags kann man einstellen (ein bisschen versteckt: "options" meben dem "upload" knopf) - das ist aber auf "," voreingestellt. Hat bei mir auch für den import von citeulike gut geklappt (aber nicht alle eintgräge wurden importiert - das muss ich mir noch genau angucken).

Aha ... hab ihn. Danke für den hint ... ein paar Einträge gingen auch bei mir hops. -- Ralf.

Öh, jetzt hab ich das nochmal angeguckt, un wie's aussieht, geht das doch nicht: ich hatte "," eingestellt, getrennt hat er aber anscheinend an " " (was citeulike hinter dem komma einfügt). Jetzt habe ich einen haufen tags die mit Komma enden :/

Der Import der Zotero-BibTeX-Datei hat mit dem "," und dem "_"-Ersatz von Leerzeichen recht gut geklappt. Zumindest kamen so keine Tags mehr zustande, die Kommata angehängt haben. -- Ralf.

Scheint leider noch etwas buggy zu sein... schade... und der bug tracker ist komisch. -- Daniel 16:49, 11 February 2008 (CET)

Interessant finde ich die "Relations" von tags (sowas wie Kategorisierung, glaube ich - muss noch damit spielen). -- Daniel 15:59, 11 February 2008 (CET)

Die Relations sind in der Tat so etwas wie eine Kategorisierung. Sie werden dann nützlich, wenn man sich alle Einträge anschaut, die einem Tag zugewiesen sind, um diverse zugehörige Einträge zu finden. -- Ralf.

Hey-ho ... offenbar wurden unsere Gebete erhört. Seit vorgestern (21.2.) scheint es die Möglichkeit mit Omeka zu geben, die zotero-Daten zu teilen. Allerdings habe ich das Tool noch nicht zum Laufen bringen können und momentan wenig Zeit dafür. Wenn also jemand damit Erfahrung macht und sie mit mir teilen möchte ... ich bin ganz Ohr. --Colman 20:37, 23 February 2008 (CET)

[edit] Zotero to BibSonomy

Last week I finished the prototyp of a simple Zotero extension that can be used to directly send the exported data to an HTML form. Instead of saving BibTeX and uploading to BibSonomy you could "copy to BibSonomy" one click. A general problem are the limits of BibTeX and how to avoid duplicating entries. Good ideas are welcome. I am also unsure about the reverse way (scrap your whole BibSonomy collection and import it into Zotero). BibSonomy provides an API, maybe you could use it. Zotero has announced to implement community features but my impression on the developer mailing list is that they still argue about the basics (whether to use AtomPub or not etc.) so it could take some time. Greetings -- JakobVoss 18:34, 12 February 2008 (CET)

Wow, that sounds good! Are the bobsonomy people aware of what you are doing? IMHO, it would be ideal to develop a "frontend/backend" relation, out of the current "import/export" situation. Or, to put it differently: what is needed is a protocol cum API, not just a "dead" format like bibtext. -- Daniel 01:00, 13 February 2008 (CET)

[edit] Zotero and Omeka

Ich habe da so ein Gefühl, dass ein eigener Abstract für dieses Thema notwendig sein wird. Omeka als neue Zotero-ready web publishing platform.--Colman 20:40, 23 February 2008 (CET)

[edit] Software Architecture

Building from scratch always makes more fun, but to get a tool that can and will be used "in the wild", we should think about which modules make sense (the buzzword is "Service Oriented Architecture") and then see what existing systems could take which role and which part must be newly created. Here are some more quick thoughts (by Jakob, not logged in):

• bibliographical records for works (monographs, papers, articles, etc)
• bibliographical records for publications (books, magazines, etc, with edition)
• bibtex import/export

That's a storage system for bibliographical records. There are many such systems with different limits. For abstraction treat it as a simple database with insert/update/delete and define the import/export data format. BibTeX is very limited but kind of an esperanto of bibliographic formats. MODS is more powerfull but I am not sure about it.

Yes, indeed. BibTeX is the smallest common denominator. Something that should definitly be supported, but which should not hold us back. See also my comment wrt a protocol in the previous section

• explicite license info, promote open content / open access

In many cases the whole bibliographic entity is not open but there is a an open preprint or such. We need to link licensed copies to a record (n:1). The information "where can I get this book/paper and what's the license" should be seperated from the basic bibliographic data (author, title, date..) in my opinion

• allow document upload, just for analysis, or for personal use, or publically

Ok, a document storage that can be linked to records the same way you can link online versions.

• wiki style collaborative updates, summaries, reviews like OpenLibrary or Bibliowiki
• tagging, like at bibsonomy

A very interesting part, but editing in form of the bag-model (we all edit one record like Wikipedia) or set model (everyone edits his own record)?

Actually, i would like to have both. I'm just not sure how to keep them separate, or make the difference clear to users (since they would work almost the same way). Maybe naming is all: Categorization is global, shelves are personal? Something like that? While I really want collaborative organization of entries (as opposed to x copies, as most bib sites use currently), I walso want to be able to organize stuff by wether i've already read it, or for which project I need it.

• meta-pages for authors, publishers, universities, conferences, etc - each with an automatic listing of books

You need URIs or a system that resolved canonical names to meta-pages. Maybe a Wiki with namespaces for author, publisher etc. and included automatic listings which dynamically come from a search in the record store. It's all about identifiers and canonical names. You need to clarify where to store the identifiers on import/export and/or how to control names.

identifiers and canonical names, yes. And properites of entities that can be used to list stuff.

• annotated reading lists
• personal shelves, tags, and comments (public or private), like the library thing or citeUlike

A system to group and tag bibliographic records. If each record has a URL, we can use any social tagging system.

yes - again, the tricky part is getting the public and personal bits to integrate nicely, and not get confusing.

• citation graph, like citeseer

First you need citation data which is not easily available. In theory a simple "cites"-relation to link records' is everything you need, in practise it all depends on the user interface. Looks like a seperated module.

make it flexible: allow arbitrary relations between records. And modules can just add them from wherever they find them. For example from scraping citation data from pdfs.

• gleaning info from well-known sites, and from unknown sites using rdf meta-info, microformats, or heuristics
• gleaning info from pdf/tex files, online or uploaded

A bunch of importers/translators/scrapers... - Zotero has a lot of them (see scaffold) but we need more. The target format needs to be clear!

The format is largely a matter of taste, but what we really need is a clear data model, which is nither too restrictive nor too arbitrary.

• formal library catalog structure, in addition to, or integrated with tagging (like the proposed Wikicat)

keep it simple. If you have the parts above (storage system for bibliographical records, editing records, system to group and tag record, system to link copies to a record, a document storage to manage such copies, citation database, import routines), you have everything you need.

ideally we would have a simple and flexible core spec and engine, which can be extended by modules/plugins as needed. In practice, the extension you need is never the one you anticipated when designing the system :) -- Daniel 01:10, 13 February 2008 (CET)

[edit] W3C: Semantic blogging for bibliographies

This sounds rather interesting: http://www.citeulike.org/profile/brightbyte/import_go

I have only skimmed it so far, and it's a bit old (2001), but it seems to touch on many points of my initial rant.

I guess I should read it carefully, and keep collecting some more idea here, and then write blog about it again. Hm... this makes me wonder... on a wiki blog (a wog?), would it make sense to rewrite the current page, or make a new entry? -- Daniel 22:40, 26 February 2008 (CET)

[edit] bibsonomy

Just for reference, I have started to collect my experiences with bibsonomy. -- Daniel 01:48, 27 February 2008 (CET)