Re: [CODE4LIB] Twitter annotations and library software
On 07.06.2010 16:15, Jay Luker wrote: Hi all, I found this thread rather interesting and figured I'd try and revive the convo since apparently some things have been happening in the twitter annotation space in the past month. I just read on techcrunch that testing of the annotation features will commence next week [1]. Also it appears that an initial schema for a book type has been defined [2]. [1] http://techcrunch.com/2010/06/02/twitter-annotations-testing/ [2] http://apiwiki.twitter.com/Annotations-Overview#RecommendedTypes Have any code4libbers gotten involved in this beyond just opining on list? I don't this so - the discussion slipped to general data modelling questions. For the specific, limited use case of twitter annotations I bet the recommended format from [2] will be fine (title is implied as common attribute, url is optional): {book:{ title: ..., author: ..., isbn: ..., year: , url: ... }} I only miss an article type with a doi field for non-books. Cheers, Jako -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
Hi all, I found this thread rather interesting and figured I'd try and revive the convo since apparently some things have been happening in the twitter annotation space in the past month. I just read on techcrunch that testing of the annotation features will commence next week [1]. Also it appears that an initial schema for a book type has been defined [2]. Have any code4libbers gotten involved in this beyond just opining on list? --jay [1] http://techcrunch.com/2010/06/02/twitter-annotations-testing/ [2] http://apiwiki.twitter.com/Annotations-Overview#RecommendedTypes
Re: [CODE4LIB] Twitter annotations and library software
Alex, Could you expand on how you think the problem that OpenURL tackles would have been better approached with existing mechanisms? I'm not debating this necessarily, but from my perspective when OpenURL was first introduced it solved a real problem that I hadn't seen solved before. Owen On Thu, Apr 29, 2010 at 11:55 PM, Alexander Johannesen alexander.johanne...@gmail.com wrote: Hi, On Thu, Apr 29, 2010 at 22:47, Walker, David dwal...@calstate.edu wrote: I would suggest it's more because, once you step outside of the primary use case for OpenURL, you end-up bumping into *other* standards. These issues were raised all the back when it was created, as well. I guess it's easy to be clever in hindsight. :) Here's what I wrote about it 5 years ago (http://shelter.nu/blog-159.html) ; So let's talk about 'Not invented here' first, because surely, we're all guilty of this one from time to time. For example, lately I dug into the ANSI/NISO Z39.88 -2004 standard, better known as OpenURL. I was looking at it critically, I have to admit, comparing it to what I already knew about Web Services, SOA, http, Google/Amazon/Flickr/Del.icio.us API's, and various Topic Maps and semantic web technologies (I was the technical editor of Explorers Guide to the Semantic Web) I think I can sum up my experiences with OpenURL as such; why? Why have the library world invented a new way of doing things that already can be done quite well already? Now, there is absolutely nothing wrong with the standard per se (except a pretty darn awful choice of name!!), so I'm not here criticising the technical merits and the work put into it. No, it's a simple 'why' that I have yet to get a decent answer to, even after talking to the OpenURL bigwigs about it. I mean, come on; convince me! I'm not unreasonable, no truly, really, I just want to be convinced that we need this over anything else. Regards, Alex -- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps --- http://shelter.nu/blog/ -- -- http://www.google.com/profiles/alexander.johannesen --- -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
Tim, I'd vote for adopting the same approach as COinS on the basis it already has some level of adoption, and we know covers at least some of the stuff libraries and academic users (as used by both libraries and consumer tools such as Zotero) might want to do. We are talking Books (from what you've said), so we don't have to worry about other formats. (although it does mean we can do journal articles and some other stuff as well for no effort) Mendeley and Zotero already speak COinS, it is pretty simple, and there are already several code libraries to deal with it. It isn't where I hope we end up in the longterm but if we talk about this happening tomorrow, why not use something that is relatively simple, already has a good set of implementations, and we know works for several cases of embedding book metadata in a web environment Owen On Thu, Apr 29, 2010 at 7:01 PM, Jakob Voss jakob.v...@gbv.de wrote: Dear Tim, you wrote: So this is my recommended framework for proceeding. Tim, I'm afraid you'll actually have to do the hard work yourself. No, I don't. Because the work isn't fundamentally that hard. A complex standard might be, but I never for a moment considered anything like that. We have *512 bytes*, and it needs to be usable by anyone. Library technology is usually fatally over-engineered, but this is a case where that approach isn't even possible. Jonathan did a very well summary - you just have to pick what you main focus of embedding bibliographic data is. A) I favour using the CSL-Record format which I summarized at http://wiki.code4lib.org/index.php/Citation_Style_Language because I had in mind that people want to have a nice looking citation of the publication that someone tweeted about. The drawback is that CSL is less adopted and will not always fit in 512 bytes B) If you main focus is to link Tweets about the same publication (and other stuff about this publication) than you must embed identifiers. LibraryThing is mainly based on two identifiers 1) ISBN to identify editions 2) LT work ids to identify works I wonder why LT work ids have not picked up more although you thankfully provide a full mapping to ISBN at http://www.librarything.com/feeds/thingISBN.xml.gz but nevermind. I thought that some LT records also contain other identifiers such as OCLC number, LOC number etc. but maybe I am wrong. The best way to specify identifiers is to use an URI (all relevant identifiers that I know have an URI form). For ISBN it is uri:isbn:{ISBN13} For LT Work-ID you can use the URL with your .com top level domain: http://www.librarything.com/work/{LTWORKID}http://www.librarything.com/work/%7BLTWORKID%7D That would fit for tweets about books with an ISBN and for tweets about a work which will make 99.9% of tweets from LT about single publications anyway. C) If your focus is to let people search for a publication in libraries than and to copy bibliographic data in reference management software then COinS is a way to go. COinS is based on OpenURL which I and others ranted about because it is a crapy library standard like MARC. But unlike other metadata formats COinS usually fits in less then 512 bytes. Furthermore you may have to deal with it for LibraryThing for libraries anyway. Although I strongly favour CSL as a practising library scientist and developer I must admit that for LibraryThing the best way is to embed identifiers (ISBN and LT Work-ID) and maybe COinS. As long as LibraryThing does not open up to more complex publications like preprints of proceeding-articles in series etc. but mainly deals with books and works this will make LibraryThing users happy. Then, three years from now, we can all conference-tweet about a CIL talk, about all the cool ways libraries are using Twitter, and how it's such a shame that the annotations standard wasn't designed with libraries in mind. How about a bet instead of voting. In three years will there be: a) No relevant Twitter annotations anyway b) Twitter annotations but not used much for bibliographic data c) A rich variety of incompatible bibliographic annotation standards d) Semantic Web will have solved every problem anyway .. Cheers Jakob -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
On Fri, Apr 30, 2010 at 18:47, Owen Stephens o...@ostephens.com wrote: Could you expand on how you think the problem that OpenURL tackles would have been better approached with existing mechanisms? As we all know, it's pretty much a spec for a way to template incoming and outgoing URLs, defining some functionality along the way. As such, URLs with basic URI templates and rewriting have been around for a long time. Even longer than that is just the basics of HTTP which have status codes and functionality to do exactly the same. We've been doing link resolving since mid 90's, either as CGI scripts, or as Apache modules, so none of this were new. URI comes in, you look it up in a database, you cross-check with other REQUEST parameters (or sessions, if you must, as well as IP addresses) and pop out a 303 (with some possible rewriting of the outgoing URL) (with the hack we needed at the time to also create dummy pages with META tags *shudder*). So the idea was to standardize on a way to do this, and it was a good idea as such. OpenURL *could* have had a great potential if it actually defined something tangible, something concrete like a model of interaction or basic rules for fishing and catching tokens and the like, and as someone else mentioned, the 0.1 version was quite a good start. But by the time when 1.0 came out, all the goodness had turned so generic and flexible in such a complex way that handling it turned you right off it. The standard also had a very difficult language, and more specifically didn't use enough of the normal geeky language used by sysadmins around. The more I tried to wrap my head around it, the more I felt like just going back to CGI scripts that looked stuff up in a database. It was easier to hack legacy code, which, well, defeats the purpose, no? Also, forgive me if I've forgotten important details; I've suppressed this part of my life. :) Kind regards, Alex -- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps --- http://shelter.nu/blog/ -- -- http://www.google.com/profiles/alexander.johannesen ---
Re: [CODE4LIB] Twitter annotations and library software
Thanks Alex, This makes sense, and yes I see what your saying - and yes, if you end up going back to custom coding because it's easier it does seem to defeat the purpose. However I'd argue that actually OpenURL 'succeeded' because it did manage to get some level of acceptance (ignoring the question of whether it is v0.1 or v1.0) - the cost of developing 'link resolvers' would have been much higher if we'd been doing something different for each publisher/platform. In this sense (I'd argue) sometimes crappy standards are better than none. We've used OpenURL v1.0 in a recent project and because we were able to simply pick up code already done for Zotero, and we already had an OpenURL resolver, the amount of new code we needed for this was minimal. I think the point about Link Resolvers doing stuff that Apache and CGI scripts were already doing is a good one - and I've argued before that what we actually should do is separate some of this out (a bit like Johnathan did with Umlaut) into an application that can answer questions about location (what is generally called the KnowledgeBase in link resolvers) and the applications that deal with analysing the context and the redirection (To introduce another tangent in a tangential thread, interestingly (I think!) I'm having a not dissimilar debate about Linked Data at the moment - there are many who argue that it is too complex and that as long as you have a nice RESTful interface you don't need to get bogged down in ontologies and RDF etc. I'm still struggling with this one - my instinct is that it will pay to standardise but so far I've not managed to convince even myself this is more than wishful thinking at the moment) Owen On Fri, Apr 30, 2010 at 10:33 AM, Alexander Johannesen alexander.johanne...@gmail.com wrote: On Fri, Apr 30, 2010 at 18:47, Owen Stephens o...@ostephens.com wrote: Could you expand on how you think the problem that OpenURL tackles would have been better approached with existing mechanisms? As we all know, it's pretty much a spec for a way to template incoming and outgoing URLs, defining some functionality along the way. As such, URLs with basic URI templates and rewriting have been around for a long time. Even longer than that is just the basics of HTTP which have status codes and functionality to do exactly the same. We've been doing link resolving since mid 90's, either as CGI scripts, or as Apache modules, so none of this were new. URI comes in, you look it up in a database, you cross-check with other REQUEST parameters (or sessions, if you must, as well as IP addresses) and pop out a 303 (with some possible rewriting of the outgoing URL) (with the hack we needed at the time to also create dummy pages with META tags *shudder*). So the idea was to standardize on a way to do this, and it was a good idea as such. OpenURL *could* have had a great potential if it actually defined something tangible, something concrete like a model of interaction or basic rules for fishing and catching tokens and the like, and as someone else mentioned, the 0.1 version was quite a good start. But by the time when 1.0 came out, all the goodness had turned so generic and flexible in such a complex way that handling it turned you right off it. The standard also had a very difficult language, and more specifically didn't use enough of the normal geeky language used by sysadmins around. The more I tried to wrap my head around it, the more I felt like just going back to CGI scripts that looked stuff up in a database. It was easier to hack legacy code, which, well, defeats the purpose, no? Also, forgive me if I've forgotten important details; I've suppressed this part of my life. :) Kind regards, Alex -- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps --- http://shelter.nu/blog/ -- -- http://www.google.com/profiles/alexander.johannesen --- -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
On Fri, Apr 30, 2010 at 20:29, Owen Stephens o...@ostephens.com wrote: However I'd argue that actually OpenURL 'succeeded' because it did manage to get some level of acceptance (ignoring the question of whether it is v0.1 or v1.0) - the cost of developing 'link resolvers' would have been much higher if we'd been doing something different for each publisher/platform. In this sense (I'd argue) sometimes crappy standards are better than none. Well, perhaps. I see OpenURL as the natural progression from PURL, in which both have their degree of success, however I'm careful using that word as I live on the outside of the library world. It may well be a success on the inside. :) I think the point about Link Resolvers doing stuff that Apache and CGI scripts were already doing is a good one - and I've argued before that what we actually should do is separate some of this out (a bit like Johnathan did with Umlaut) into an application that can answer questions about location (what is generally called the KnowledgeBase in link resolvers) and the applications that deal with analysing the context and the redirection Yes, split it into smaller chunks is always smart, especially with complex issues. For example, in the Topic Maps world, the who standard (reference model, data model, query language, constraint language, XML exchange language, various notational languages) is wrapped up with a guide in the middle. Make them into smaller parcels, and make your flexible point there. If you pop it all into one, no one will read it and fully understand it. (And don't get me started on the WS-* set of standards on the same issues ...) (To introduce another tangent in a tangential thread, interestingly (I think!) I'm having a not dissimilar debate about Linked Data at the moment - there are many who argue that it is too complex and that as long as you have a nice RESTful interface you don't need to get bogged down in ontologies and RDF etc. I'm still struggling with this one - my instinct is that it will pay to standardise but so far I've not managed to convince even myself this is more than wishful thinking at the moment) Ah, now this is certainly up my alley. As you might have seen, I'm a Topic Maps guy, and we have in our model a distinction between three different kinds of identities; internal, external indicators and published subject identifiers. The RDF world only had rdf:about, so when you used www.somewhere.org, are you talking about that thing, or does that thing represent something you're talking about? Tricky stuff which has these days become a *huge* problem with Linked Data. And yes, they're trying to solve that by issuing a HTTP 303 status code as a means of declaring the identifiers imperative, which is a *lot* of resolving to do on any substantial set of data, and in my eyes a huge ugly hack. (And what if your Internet falls down? Tough.) Anyway, here's more on these identity problems ; http://www.ontopia.net/topicmaps/materials/identitycrisis.html As to the RESTful notions, they only take you as far as content-types can take you. Sure, you can gleam semantics from it, but I reckon there's an impedance mismatch between just the things librarians how got down pat ; meta data vs. data. CRUD or, in this example, GPPD (get/post/put/delete), who aren't in a dichotomy btw, can only determine behavior that enables certain semantic paradigms, but cannot speak about more complex relationships or even modest models. (Very often models aren't actionable :) The funny thing is that after all these years of working with Topic Maps I find that these hard issues have been solved years ago, and the rest of the world is slowly catching up to it. I blame the lame DAML+OIL background of RDF and OWL, to be honest; a model too simple to be elegantly advanced and too complex to be easily useful. Kind regards, Alex -- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps --- http://shelter.nu/blog/ -- -- http://www.google.com/profiles/alexander.johannesen ---
Re: [CODE4LIB] Twitter annotations and library software
Jonathan Rochkind wrote: Call me pedantic but if you do not have an identifier than there is no hope to identity the publication by means of metadata. You only *describe* it with metadata and use additional heuristics (mostly search engines) to hopefully identify the publication based on the description. But the entire OpenURL infrastructure DOES this, and does it without using search engines. It's a real world use case that has a solution in production! So, yeah, I call you pedantic for wanting to pretend the use case and the real world solution doesn't exist. :) As you said OpenURL is an *infrastructure*. It only makes sense if you have resolvers that map an OpenURL to a unique publications. This resolvers do the identification while OpenURL only describes (as long as you do not put a unique publication identifier into an OpenURL). In contrast an identifier can be used to compare and search publications without any additional infrastructure. You can call it description rather than identification if you like, that is a question of terminology. But it's description that is meant to uniquely identify a particular publication, and that a whole bunch of software in use every day succesfully uses to identify a particular publication. It's not just terminology if you can either just compare two strings for equalness (identifcation) or you need an infrastructure with knowledge bases and specific software (to make use of a description). OpenURL is of no use if you seperate it from the existing infrastructure which is mainly held by companies. No sane person will try to build an open alternative infrastructure because OpenURL is a crapy library-standard like MARC etc. This rant on OpenURL summarizes it well: http://cavlec.yarinareth.net/2006/10/13/i-hate-library-standards/ The OpenURL specification is a 119 page PDF - that alone is a reason to run away as fast as you can. If a twitter annotation setup wants to be able to identify publications that don't have standard identifiers, then you don't want to ignore this use case and how actually in production software currently deals with it. You can perhaps find a better way to deal with it -- I'm certainly not arguing for OpenURL as the be all end all, I rather hate OpenURL actually. But dismissing it as impossible is indeed pedantic, since it's being done! If a twitter annotation setup wants to get adopted than it should not be build on a crapy complex library standard like OpenURL. It IS a hacky and error-prone solution, to be sure. But it's the best solution we've got, because it's simply a fact that we have many publications we want to identify that lack standard identifiers. Ok, back to serious: Bibliographic Twitter annotations should be designed in a way that libraries (or whoever provides that knowledge bases aka OpenURL resolvers) can use it to look up a publication by its metadata. So there should be a transformation Twitter annotation = OpenURL If you choose CSL as bibliographic input format you can hopefully create a CSL style that does not produce a citation but an OpenURL - Voilà! I must admit that this solution is based on the open assumption that CSL record format contains all information needed for OpenURL which may not the case. A good point to start from is the function createContextObject in https://www.zotero.org/svn/extension/trunk/chrome/content/zotero/xpcom/ingester.js which is used by Zotero to create OpenURLs. Cheers Jakob -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
Okay, I know it's cool to hate on OpenURL, but I feel I have to clarify a few points: OpenURL is of no use if you seperate it from the existing infrastructure which is mainly held by companies. No sane person will try to build an open alternative infrastructure because OpenURL is a crapy library-standard like MARC etc. OpenURL is mostly implemented by libraries, yes, but it isn't necessarily *just* a library standard - this is akin to saying that Dublin Core is a library standard. Only sort of. The other issue I have is that — although Jonathan used the term to make a point — OpenURL is *not* an infrastructure, it is a protocol. Condemning the current OpenURL infrastructure (which is mostly a vendor-driven oligopoly) is akin to saying in 2004 that HTTP and HTML sucks because Firefox hadn't been released yet and all we had was IE6. Don't condemn the standard because of the implementation. The OpenURL specification is a 119 page PDF - that alone is a reason to run away as fast as you can. The main reason for this is because OpenURL can do much, much, much more than the simple resolve a unique copy use case that libraries use it for. We're using maybe 1% of the spec for 99% of our practice, probably because librarians weren't imaginative (as Jim Weinheimer would say) enough to think of other use cases beyond that most pressing one. I'd contend that OpenURL, like other technologies (cough XML) is greatly misunderstood, and therefore abused, and therefore discredited. I think there is also often confusion between the KEV schemas and OpenURL itself (which is really what Dorothea's blog rant is about); I'm certainly guilty of this myself, as Jonathan can attest. You don't *have* to use the KEVs with OpenURL, you can use anything, including eg. Dublin Core. If a twitter annotation setup wants to get adopted than it should not be build on a crapy complex library standard like OpenURL. I don't quite understand this (but I think I agree) — twitter annotation should be built on a data model, and then serialized via whatever protocols make sense (which may or may not include OpenURL). I must admit that this solution is based on the open assumption that CSL record format contains all information needed for OpenURL which may not the case. … A good example. And this is where you're exactly right that we need better tools, namely OpenURL resolvers which can do much more than they do now. I've had the idea for a number of years now that OpenURL functionality should be merged into aggregation / discovery layer (eg. OAI harvester)-type systems, because, like OAI-PMH, OpenURL can *transport metadata*, we just don't use it for that in practice. A ContextObject is just a triple that makes a single assertion about two entities (resources): that A references B. Just like an RDF statement using http://purl.org/dc/terms/references, but with more focus on describing the entities rather than the assertion. Maybe if I put it that way, OpenURL sounds a little less crappy. MJ
Re: [CODE4LIB] Twitter annotations and library software
On Thu, Apr 29, 2010 at 8:17 AM, MJ Suhonos m...@suhonos.ca wrote: Okay, I know it's cool to hate on OpenURL, but I feel I have to clarify a few points: It's not that it's cool to hate on OpenURL, but if you've really worked with it it's easy to grow bitter. snip Maybe if I put it that way, OpenURL sounds a little less crappy. No, OpenURL is still crappy and it will always be crappy, I'm afraid, because it's tremendously complicated, mainly from the fact that it tries to do too much. The reason that context-sensitive services based on bibliographic citations comprise 99% of all OpenURL activity is because: A) that was the problem it was originally designed to solve B) it's the only thing it really does well (and OpenURL 1.0's insistence on being able to solve any problem almost takes that strength away from it) The barriers to entry + the complexity of implementation almost guarantee that there's a better or, at any rate, easier alternative to any problem. The difference between OpenURL and DublinCore is that the RDF community picked up on DC because it was simple and did exactly what they needed (and nothing more). A better analogy would be Z39.50 or SRU: two non-library-specific protocols that, for their own reasons, haven't seen much uptake outside of the library community. -Ross.
Re: [CODE4LIB] Twitter annotations and library software
On 29 April 2010 13:17, MJ Suhonos m...@suhonos.ca wrote: The OpenURL specification is a 119 page PDF - that alone is a reason to run away as fast as you can. The main reason for this is because OpenURL can do much, much, much more than the simple resolve a unique copy use case that libraries use it for. We're using maybe 1% of the spec for 99% of our practice, probably because librarians weren't imaginative (as Jim Weinheimer would say) enough to think of other use cases beyond that most pressing one. It's worth contrasting this with the original OpenURL specification, now retro-numbered as v0.1: http://www.openurl.info/registry/docs/pdf/openurl-01.pdf This is the one that everyone implemented in a burst of enthusiasm earlier this decade. You know, in the way, almost on-one's implemented v1.0. That document is TEN pages long. Eight, really, since the total count includes a page containing the foreword written after the event and a page of acknowledgements consisting of a single 11-word sentence. Can we be surprised that this specification attracted more interest than the one fifteen times longer? OpenURL 1.0 took that simple, comprehensible spec -- one that you could read over lunch and fully understand -- and blew it up into a super-generalised exercise in architecture astronautics. And then provided ANOTHER big document explaining how you can profile OpenURL 1.0 to make it do the stuff that v0.1 does (i.e. what you actually WANT it to do) -- except, of course, that it expresses the same concepts in a different way, so that v0.1 and v1.0 OpenURLs are mutually incomprehensible. All of this to support vapour use-cases that no-one has taken advantage of because no-one ever needed to do that stuff. So the sum achievement of OpenURL 1.0 has been (A) to fill people with fear of what used to be a very useful and perfectly straightforward specification, and (B) where implemented at all, to balkanise implementations. I'd contend that OpenURL, like other technologies (cough XML) is greatly misunderstood, and therefore abused, and therefore discredited. I think there is also often confusion between the KEV schemas and OpenURL itself (which is really what Dorothea's blog rant is about); I'm certainly guilty of this myself, as Jonathan can attest. You don't *have* to use the KEVs with OpenURL, you can use anything, including eg. Dublin Core. Yeah. So long as you don't mind that only 0.01% of the world's OpenURL resolvers will know what to do with them.
Re: [CODE4LIB] Twitter annotations and library software
We're using maybe 1% of the spec for 99% of our practice, probably because librarians weren't imaginative (as Jim Weinheimer would say) enough to think of other use cases beyond that most pressing one. I would suggest it's more because, once you step outside of the primary use case for OpenURL, you end-up bumping into *other* standards. Dorthea'sblog post that Jakob referenced in his message is a good example of that. She was trying to use OpenURL (via COINS) to get data into Zotero. Mid-way through the post she wonders if maybe she should have gone with unAPI instead. And, in fact, I think that would have been a better approach. unAPI is better at doing that particular task than OpenURL. And I think that may explain why OpenURL hasn't become the One Standard to Rule Them All, even though it kind of presents itself that way. --Dave == David Walker Library Web Services Manager California State University http://xerxes.calstate.edu From: Code for Libraries [code4...@listserv.nd.edu] On Behalf Of MJ Suhonos [...@suhonos.ca] Sent: Thursday, April 29, 2010 5:17 AM To: CODE4LIB@LISTSERV.ND.EDU Subject: Re: [CODE4LIB] Twitter annotations and library software Okay, I know it's cool to hate on OpenURL, but I feel I have to clarify a few points: OpenURL is of no use if you seperate it from the existing infrastructure which is mainly held by companies. No sane person will try to build an open alternative infrastructure because OpenURL is a crapy library-standard like MARC etc. OpenURL is mostly implemented by libraries, yes, but it isn't necessarily *just* a library standard - this is akin to saying that Dublin Core is a library standard. Only sort of. The other issue I have is that — although Jonathan used the term to make a point — OpenURL is *not* an infrastructure, it is a protocol. Condemning the current OpenURL infrastructure (which is mostly a vendor-driven oligopoly) is akin to saying in 2004 that HTTP and HTML sucks because Firefox hadn't been released yet and all we had was IE6. Don't condemn the standard because of the implementation. The OpenURL specification is a 119 page PDF - that alone is a reason to run away as fast as you can. The main reason for this is because OpenURL can do much, much, much more than the simple resolve a unique copy use case that libraries use it for. We're using maybe 1% of the spec for 99% of our practice, probably because librarians weren't imaginative (as Jim Weinheimer would say) enough to think of other use cases beyond that most pressing one. I'd contend that OpenURL, like other technologies (cough XML) is greatly misunderstood, and therefore abused, and therefore discredited. I think there is also often confusion between the KEV schemas and OpenURL itself (which is really what Dorothea's blog rant is about); I'm certainly guilty of this myself, as Jonathan can attest. You don't *have* to use the KEVs with OpenURL, you can use anything, including eg. Dublin Core. If a twitter annotation setup wants to get adopted than it should not be build on a crapy complex library standard like OpenURL. I don't quite understand this (but I think I agree) — twitter annotation should be built on a data model, and then serialized via whatever protocols make sense (which may or may not include OpenURL). I must admit that this solution is based on the open assumption that CSL record format contains all information needed for OpenURL which may not the case. … A good example. And this is where you're exactly right that we need better tools, namely OpenURL resolvers which can do much more than they do now. I've had the idea for a number of years now that OpenURL functionality should be merged into aggregation / discovery layer (eg. OAI harvester)-type systems, because, like OAI-PMH, OpenURL can *transport metadata*, we just don't use it for that in practice. A ContextObject is just a triple that makes a single assertion about two entities (resources): that A references B. Just like an RDF statement using http://purl.org/dc/terms/references, but with more focus on describing the entities rather than the assertion. Maybe if I put it that way, OpenURL sounds a little less crappy. MJ
Re: [CODE4LIB] Twitter annotations and library software
It's not that it's cool to hate on OpenURL, but if you've really worked with it it's easy to grow bitter. Well, fair enough. Perhaps what I'm defending isn't OpenURL per se, but rather the concept of being able to transport descriptive assertions the way the 1.0 spec proposes. The reason that context-sensitive services based on bibliographic citations comprise 99% of all OpenURL activity is because: A) that was the problem it was originally designed to solve Yes, right. And neither libraries nor vendors moved past this when 1.0 came out for the reasons described (too complex, no immediate use cases). The barriers to entry + the complexity of implementation almost guarantee that there's a better or, at any rate, easier alternative to any problem. Let me be clear: I am *all* for a better system — even the first RSS specs were fragmented and crappy, which led to Atom. But for the time they were around, they were useful, if kludgy. My only point (and I think, Jonathan's) is that OpenURL, for better or worse, *exists* and *works* now, if not ideally. If it sucks, the onus is on us, I think, to improve it or produce something better. The difference between OpenURL and DublinCore is that the RDF community picked up on DC because it was simple and did exactly what they needed (and nothing more). Actually the difference between OpenURL and DC is that one is a transport protocol and one is a metadata schema. :-) But I get your and Mike's point about OpenURL 1.0 being too complicated for librarians to bother with. All of this to support vapour use-cases that no-one has taken advantage of because no-one ever needed to do that stuff. So the sum achievement of OpenURL 1.0 has been (A) to fill people with fear of what used to be a very useful and perfectly straightforward specification, and (B) where implemented at all, to balkanise implementations. Sounds a lot like Z39.50, to me, actually. I guess I just see this as a classic example of librarians (and of course I'm generalizing) sitting with a tool-in-hand and saying this isn't good enough, tossing it in the trash, and then lamenting the lack of tools for doing useful things. Sort of like MODS (for those on the NGC4lib list). I know we're supposed to be pragmatists on C4L, but do we just relegate ourselves to doing stuff we need to do, or pushing our existing tools to experiment? You don't *have* to use the KEVs with OpenURL, you can use anything, including eg. Dublin Core. Yeah. So long as you don't mind that only 0.01% of the world's OpenURL resolvers will know what to do with them. Absolutely. So how about we build some better resolvers and do useful and interesting new things with them? Like, Twitter annotations. :-) MJ
Re: [CODE4LIB] Twitter annotations and library software
Let me correct myself (for the detail-oriented among us): Actually the difference between OpenURL and DC is that one is a transport protocol and one is a metadata schema. :-) OpenURL is a *serialization format* which happens to be actionable by a transport protocol (HTTP), which is its main benefit.
Re: [CODE4LIB] Twitter annotations and library software
I agree that OpenURL is crappy. My point was that the problem case -- 'identifying' (or describing an element sufficient for identification, if you like to call it that) publications that do not have standard identifiers -- is a real one. OpenURL _does_ solve it. You _probably_ don't want to ignore this problem case in a twitter annotation scenario. If you can solve it _better_ than OpenURL, than all the better. Or if you decide intentionally to exclude it from your scenario, that's fine, you know your intended domain. But OpenURL, despite it's crappiness, _does_ address this problem case reasonably effectively, and it is really in use. I'm certainly not trying to be an OpenURL booster. But it works, and until/unless we have something better, is is addressing a problem case that is really important in many scenarios (like getting users to licensed full text, naturally). Jonathan Ross Singer wrote: On Thu, Apr 29, 2010 at 8:17 AM, MJ Suhonos m...@suhonos.ca wrote: Okay, I know it's cool to hate on OpenURL, but I feel I have to clarify a few points: It's not that it's cool to hate on OpenURL, but if you've really worked with it it's easy to grow bitter. snip Maybe if I put it that way, OpenURL sounds a little less crappy. No, OpenURL is still crappy and it will always be crappy, I'm afraid, because it's tremendously complicated, mainly from the fact that it tries to do too much. The reason that context-sensitive services based on bibliographic citations comprise 99% of all OpenURL activity is because: A) that was the problem it was originally designed to solve B) it's the only thing it really does well (and OpenURL 1.0's insistence on being able to solve any problem almost takes that strength away from it) The barriers to entry + the complexity of implementation almost guarantee that there's a better or, at any rate, easier alternative to any problem. The difference between OpenURL and DublinCore is that the RDF community picked up on DC because it was simple and did exactly what they needed (and nothing more). A better analogy would be Z39.50 or SRU: two non-library-specific protocols that, for their own reasons, haven't seen much uptake outside of the library community. -Ross.
Re: [CODE4LIB] Twitter annotations and library software
Yes, what MJ said is indeed exactly my perspective as well. MJ Suhonos wrote: It's not that it's cool to hate on OpenURL, but if you've really worked with it it's easy to grow bitter. Well, fair enough. Perhaps what I'm defending isn't OpenURL per se, but rather the concept of being able to transport descriptive assertions the way the 1.0 spec proposes. The reason that context-sensitive services based on bibliographic citations comprise 99% of all OpenURL activity is because: A) that was the problem it was originally designed to solve Yes, right. And neither libraries nor vendors moved past this when 1.0 came out for the reasons described (too complex, no immediate use cases). The barriers to entry + the complexity of implementation almost guarantee that there's a better or, at any rate, easier alternative to any problem. Let me be clear: I am *all* for a better system — even the first RSS specs were fragmented and crappy, which led to Atom. But for the time they were around, they were useful, if kludgy. My only point (and I think, Jonathan's) is that OpenURL, for better or worse, *exists* and *works* now, if not ideally. If it sucks, the onus is on us, I think, to improve it or produce something better. The difference between OpenURL and DublinCore is that the RDF community picked up on DC because it was simple and did exactly what they needed (and nothing more). Actually the difference between OpenURL and DC is that one is a transport protocol and one is a metadata schema. :-) But I get your and Mike's point about OpenURL 1.0 being too complicated for librarians to bother with. All of this to support vapour use-cases that no-one has taken advantage of because no-one ever needed to do that stuff. So the sum achievement of OpenURL 1.0 has been (A) to fill people with fear of what used to be a very useful and perfectly straightforward specification, and (B) where implemented at all, to balkanise implementations. Sounds a lot like Z39.50, to me, actually. I guess I just see this as a classic example of librarians (and of course I'm generalizing) sitting with a tool-in-hand and saying this isn't good enough, tossing it in the trash, and then lamenting the lack of tools for doing useful things. Sort of like MODS (for those on the NGC4lib list). I know we're supposed to be pragmatists on C4L, but do we just relegate ourselves to doing stuff we need to do, or pushing our existing tools to experiment? You don't *have* to use the KEVs with OpenURL, you can use anything, including eg. Dublin Core. Yeah. So long as you don't mind that only 0.01% of the world's OpenURL resolvers will know what to do with them. Absolutely. So how about we build some better resolvers and do useful and interesting new things with them? Like, Twitter annotations. :-) MJ
Re: [CODE4LIB] Twitter annotations and library software
Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
I wouldn't count on the community using anything, just because random people on the listserv voted on it. If you're coding it, you should take account of the feedback, and then go on and create something that YOU will use, and makes sense to you. And then hope other people do too. That's pretty much the best you can do. Vote by random people on a listserv is hardly a guarantee of getting a standard that actually works, or that people actually use -- just look at OpenURL 1.0! Tim Spalding wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
I'm going to throw in my two cents. I dont think (and correct me if i'm wrong) we have mentioned once what a user might actually put in a twitter annotation. a book title? an article title? a link? i think creating some super complicated thing for a twitter annotation dooms it to failure. after all, its twitter...make it short and sweet. also the 1.0 document for OpenURL isn't really that bad (yes I have read it). a good portion of it is a chart with the different metadata elements. also open url could conceivably refer to an animal and then link to a bunch of resources on that animal, but no one has done that. i don't think that's a problem with OpenURL i think thats a problem with the metadata sent by vendors to link resolvers and librarians lack of creativity (yes i did make a ridiculous generalization that was not intended to offend anyone but inevitably it will). having been a vendor who has worked with openurl, i know that the informaiton databases send seriously effects (affects?) what you can actually do in a link resolver. On Thu, Apr 29, 2010 at 10:23 AM, Tim Spalding t...@librarything.com wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
On Thu, Apr 29, 2010 at 10:32 AM, Rosalyn Metz rosalynm...@gmail.com wrote: I'm going to throw in my two cents. I dont think (and correct me if i'm wrong) we have mentioned once what a user might actually put in a twitter annotation. a book title? an article title? a link? I think the idea is these would be machine generated from an application. So, imagine LT, Amazon, Delicious Library or SFX having a Tweet this! button and *that* provides the annotation (not the user). i think creating some super complicated thing for a twitter annotation dooms it to failure. after all, its twitter...make it short and sweet. Indeed, it's limited. also the 1.0 document for OpenURL isn't really that bad (yes I have read it). a good portion of it is a chart with the different metadata elements. also open url could conceivably refer to an animal and then link to a bunch of resources on that animal, but no one has done that. i don't think that's a problem with OpenURL i think thats a problem with the metadata sent by vendors to link resolvers and librarians lack of creativity (yes i did make a ridiculous generalization that was not intended to offend anyone but inevitably it will). having been a vendor who has worked with openurl, i know that the informaiton databases send seriously effects (affects?) what you can actually do in a link resolver. No, this is the mythical promise of 1.0, but delivery is, frankly, much more complicated than that. It is impractical to expect an OpenURL link resolver to make sense of any old thing you throw at it and return sensible results. This is the point of the community profiles, to narrow the infinite possibilities a bit. None of our current profiles would support the scenario you speak of and I would be surprised if such a service were to be devised, that it would be built on OpenURL. I think it's very easy to underestimate how complicated it is to actually build something using OpenURL since in the abstract it seems like a very logical solution to any problem. -Ross. On Thu, Apr 29, 2010 at 10:23 AM, Tim Spalding t...@librarything.com wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
At #ldow2010 on Tuesday there was a presentation on semantic Twitter via TwitLogic: http://twitlogic.fortytwo.net/ You can download the full paper if you're really curious: http://events.linkeddata.org/ldow2010/papers/ldow2010_paper16.pdf Twitter Annotations system was mentioned at the end as a possible side option. There's bound to be a good bit of talk in the Linked Data community on strapping RDF/RDFa into Twitter Annotations, but I believe that's still beginning. Additionally (as someone outside of the library community proper), OpenURL's dependence on resolvers would be the largest concern. Anyone could build similar real thing URL's and use 303 See Other redirects to return one or more digital resources about that real thing. See this for more information: http://lists.w3.org/Archives/Public/www-tag/2005Jun/0039 Enjoy the reads, Benjamin -- President BigBlueHat P: 864.232.9553 W: http://www.bigbluehat.com/ http://www.linkedin.com/in/benjaminyoung On 4/29/10 10:32 AM, Rosalyn Metz wrote: I'm going to throw in my two cents. I dont think (and correct me if i'm wrong) we have mentioned once what a user might actually put in a twitter annotation. a book title? an article title? a link? i think creating some super complicated thing for a twitter annotation dooms it to failure. after all, its twitter...make it short and sweet. also the 1.0 document for OpenURL isn't really that bad (yes I have read it). a good portion of it is a chart with the different metadata elements. also open url could conceivably refer to an animal and then link to a bunch of resources on that animal, but no one has done that. i don't think that's a problem with OpenURL i think thats a problem with the metadata sent by vendors to link resolvers and librarians lack of creativity (yes i did make a ridiculous generalization that was not intended to offend anyone but inevitably it will). having been a vendor who has worked with openurl, i know that the informaiton databases send seriously effects (affects?) what you can actually do in a link resolver. On Thu, Apr 29, 2010 at 10:23 AM, Tim Spaldingt...@librarything.com wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
Benjamin Young wrote: Additionally (as someone outside of the library community proper), OpenURL's dependence on resolvers would be the largest concern. This is a misconception. An OpenURL context object can be created to provide structured semantic citation information, without any dependence on a resolver. Just as a way of serializing structured semantic citation information in a standard way. This is basically what COinS does. Now, the largest concern with OpenURL to me is actually just that it's way harder to understand and work with than it should be to meet it's primary use cases, and that means trying to use it as a standard for a new use case is probably asking for trouble in adoption curve. So here are the questions, my own summary analysis of this thread: 1. What are the citations you think users would want to attach to a tweet? a. Will they ALL have standard identifiers that can be expressed as some form of URI (ISBN, DOI, etc). b. Or are there an important enough subset of citations that will NOT have standard identifiers that you still want to support? If you choose 'a' above, then the solution to me seems clear: Simply attach a URI as your 'citation metadata' -- be willing to use info: URIs for ISBNs, ISSNs, LCCNs, OCLCnums, DOIs. It should be clearly identified as identifier for thing cited by this tweet somehow, but the 'payload' is just a URI. [ I know some people don't like non-resolvable info: URIs. I like em, and THIS use case shows why. It allows you to attach an ISBN to a tweet as a URI right now today, keeping your metadata schema simple just a URI while still allowing ISBNs ]. And then we're done if we choose 'a' above, it's pretty simple. If you choose 'b' above, then you need a way to identify (or describe sufficient for identification) publications that do not have standard identifiers. An OpenURL context object using the standard scholarly formats (the only ones actually being used much in the real world) is ONE such way that is _actually_ being used today for _just_ this purpose. So it would be worth looking at. You could try to use it whole cloth, or you could just take the element schema from the scholarly formats and re-purpose it. You could try to fix some of it's problems. (There are many). Or you could ignore OpenURL (or rather than ignore, review it briefly for ideas) and use one of the other formats that haven't really yet caught on yet, but might be designed a lot better than OpenURL. Examples brought up in this thread include something by Jakob Voss (that I don't have the URL handy for), some kind of citation-in-json format (that I don't have the url handy for), and Bibo in RDF (that I don't have the url handy for). If you decide to go with any of these, it's probably worth _comparing_ them to OpenURL to make sure what can be expressed in OpenURL with standard scholarly formats can _also_ be expressed in the format you chose. (Last time I looked at Bibo, I recall there was no place to put a standard identifier like a DOI. So maybe using Bibo + URI for standard identifier would suffice. etc.) So this is my recommended framework for proceeding. Tim, I'm afraid you'll actually have to do the hard work yourself. Standards creation is hard. You aren't going to get something good just by getting some listserv to vote. Many of us involved in this discussion may find this intellectually interesting, but may have no actual use _ourselves_ for such a format anyway. If Amazon or someone like that comes up with something, it will end up becoming the 'de facto' standard, so I recommend trying to talk to Amazon to see if they're thinking about this -- or just wait to see if/what Amazon comes up with, and use that. Jonathan
Re: [CODE4LIB] Twitter annotations and library software
ok right now exlibris has a recommender service for sfx that stores metadata from an openurl. lets say a vendor bothered to pass an element like rft.subject=hippo (which is most likely unlikely to happen since they can't even pass an issn half the time). that subject got stored in the recommender service. next time a child saw something in ebsco animals about hippos they could click the find this button (or whatever it says) and the recommender service could bring up everything on hippos. the openurl that would be passed would be something like http://your.linkresolver.com/name?rft.subject=hippo yes this is simplistic, but its more creative then say doing something boring like just bringing up the full text or doing something half ass creative like bringing up articles that are cited in the footnotes. and to say something like rft.subject (or whatever it might be called) is out of the scope of group profiles is a little absurd since we are talking about things that already have subjects attached to them (see any database or other library related system). of course you'll probably want to talk about next how subjects aren't standardized and that makes it possible. that is true, but that isn't openurl's fault or the link resolvers fault, thats the database vendors who refuse to get with the program. On Thu, Apr 29, 2010 at 11:02 AM, Ross Singer rossfsin...@gmail.com wrote: On Thu, Apr 29, 2010 at 10:32 AM, Rosalyn Metz rosalynm...@gmail.com wrote: I'm going to throw in my two cents. I dont think (and correct me if i'm wrong) we have mentioned once what a user might actually put in a twitter annotation. a book title? an article title? a link? I think the idea is these would be machine generated from an application. So, imagine LT, Amazon, Delicious Library or SFX having a Tweet this! button and *that* provides the annotation (not the user). i think creating some super complicated thing for a twitter annotation dooms it to failure. after all, its twitter...make it short and sweet. Indeed, it's limited. also the 1.0 document for OpenURL isn't really that bad (yes I have read it). a good portion of it is a chart with the different metadata elements. also open url could conceivably refer to an animal and then link to a bunch of resources on that animal, but no one has done that. i don't think that's a problem with OpenURL i think thats a problem with the metadata sent by vendors to link resolvers and librarians lack of creativity (yes i did make a ridiculous generalization that was not intended to offend anyone but inevitably it will). having been a vendor who has worked with openurl, i know that the informaiton databases send seriously effects (affects?) what you can actually do in a link resolver. No, this is the mythical promise of 1.0, but delivery is, frankly, much more complicated than that. It is impractical to expect an OpenURL link resolver to make sense of any old thing you throw at it and return sensible results. This is the point of the community profiles, to narrow the infinite possibilities a bit. None of our current profiles would support the scenario you speak of and I would be surprised if such a service were to be devised, that it would be built on OpenURL. I think it's very easy to underestimate how complicated it is to actually build something using OpenURL since in the abstract it seems like a very logical solution to any problem. -Ross. On Thu, Apr 29, 2010 at 10:23 AM, Tim Spalding t...@librarything.com wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
On Thu, Apr 29, 2010 at 11:21 AM, Jonathan Rochkind rochk...@jhu.edu wrote: (Last time I looked at Bibo, I recall there was no place to put a standard identifier like a DOI. So maybe using Bibo + URI for standard identifier would suffice. etc.) BIBO has all sorts of identifiers (including DOI): http://bibotools.googlecode.com/svn/bibo-ontology/trunk/doc/dataproperties/doi___1125128004.html As well as ISBN (10 and 13), ISSN/e-issn, LCCN, EAN, OCLCNUM, and more. -Ross.
Re: [CODE4LIB] Twitter annotations and library software
So this is my recommended framework for proceeding. Tim, I'm afraid you'll actually have to do the hard work yourself. No, I don't. Because the work isn't fundamentally that hard. A complex standard might be, but I never for a moment considered anything like that. We have *512 bytes*, and it needs to be usable by anyone. Library technology is usually fatally over-engineered, but this is a case where that approach isn't even possible. You aren't going to get something good just by getting some listserv to vote. My suggestion was to have people interested in actually using it vote. Many of us involved in this discussion may find this intellectually interesting, but may have no actual use _ourselves_ for such a format anyway. Oh, I bet half of you guys have sharing buttons on your OPAC or elsewhere. And many of you are on Twitter and, at least occasionally, discuss a book. If Amazon or someone like that comes up with something, it will end up becoming the 'de facto' standard, so I recommend trying to talk to Amazon to see if they're thinking about this -- or just wait to see if/what Amazon comes up with, and use that. You're right. It's a thankless task to get even a subset of library technologists to agree on something like this. It'd be less important if I didn't know the Amazon solution will leave off key pieces libraries need. Then, three years from now, we can all conference-tweet about a CIL talk, about all the cool ways libraries are using Twitter, and how it's such a shame that the annotations standard wasn't designed with libraries in mind. Best, Tim
Re: [CODE4LIB] Twitter annotations and library software
I still don't really see how what you're talking about would practically be accomplished. For one, to have rft.subject, like you mention, would require using the dublincore context set. Since that wouldn't be useful on its own for the services that link resolvers currently offer, OpenURL sources (i.e. AI database providers) would have to support SAP 2 (XML) context objects so they can pass the book/journal/patent/etc. referent metadata along with the Dublin Core referent metadata. It also becomes a POST rather than a simple link (GET). What I'm saying is it ups the requirements on all ends of the ecosystem, for what? What you're talking about would be *much* more easily implemented via SRU and CQL (or OpenSearch), anyway, since your example is really performing a search. Since OpenURL doesn't have any semblance of standardized response format, a client wouldn't know what to do with the response, anyway. -Ross. On Thu, Apr 29, 2010 at 11:29 AM, Rosalyn Metz rosalynm...@gmail.com wrote: ok right now exlibris has a recommender service for sfx that stores metadata from an openurl. lets say a vendor bothered to pass an element like rft.subject=hippo (which is most likely unlikely to happen since they can't even pass an issn half the time). that subject got stored in the recommender service. next time a child saw something in ebsco animals about hippos they could click the find this button (or whatever it says) and the recommender service could bring up everything on hippos. the openurl that would be passed would be something like http://your.linkresolver.com/name?rft.subject=hippo yes this is simplistic, but its more creative then say doing something boring like just bringing up the full text or doing something half ass creative like bringing up articles that are cited in the footnotes. and to say something like rft.subject (or whatever it might be called) is out of the scope of group profiles is a little absurd since we are talking about things that already have subjects attached to them (see any database or other library related system). of course you'll probably want to talk about next how subjects aren't standardized and that makes it possible. that is true, but that isn't openurl's fault or the link resolvers fault, thats the database vendors who refuse to get with the program. On Thu, Apr 29, 2010 at 11:02 AM, Ross Singer rossfsin...@gmail.com wrote: On Thu, Apr 29, 2010 at 10:32 AM, Rosalyn Metz rosalynm...@gmail.com wrote: I'm going to throw in my two cents. I dont think (and correct me if i'm wrong) we have mentioned once what a user might actually put in a twitter annotation. a book title? an article title? a link? I think the idea is these would be machine generated from an application. So, imagine LT, Amazon, Delicious Library or SFX having a Tweet this! button and *that* provides the annotation (not the user). i think creating some super complicated thing for a twitter annotation dooms it to failure. after all, its twitter...make it short and sweet. Indeed, it's limited. also the 1.0 document for OpenURL isn't really that bad (yes I have read it). a good portion of it is a chart with the different metadata elements. also open url could conceivably refer to an animal and then link to a bunch of resources on that animal, but no one has done that. i don't think that's a problem with OpenURL i think thats a problem with the metadata sent by vendors to link resolvers and librarians lack of creativity (yes i did make a ridiculous generalization that was not intended to offend anyone but inevitably it will). having been a vendor who has worked with openurl, i know that the informaiton databases send seriously effects (affects?) what you can actually do in a link resolver. No, this is the mythical promise of 1.0, but delivery is, frankly, much more complicated than that. It is impractical to expect an OpenURL link resolver to make sense of any old thing you throw at it and return sensible results. This is the point of the community profiles, to narrow the infinite possibilities a bit. None of our current profiles would support the scenario you speak of and I would be surprised if such a service were to be devised, that it would be built on OpenURL. I think it's very easy to underestimate how complicated it is to actually build something using OpenURL since in the abstract it seems like a very logical solution to any problem. -Ross. On Thu, Apr 29, 2010 at 10:23 AM, Tim Spalding t...@librarything.com wrote: Can we just hold a vote or something? I'm happy to do whatever the community here wants and will actually use. I want to do something that will be usable by others. I also favor something dead simple, so it will be implemented. If we don't reach some sort of conclusion, this is an interesting waste of time. I propose only people engaged in doing something along these lines get to vote? Tim
Re: [CODE4LIB] Twitter annotations and library software
OK, back to Tim's specific question. I'm not sure why you want to put bib data in a tweet at all for your application. Why not just use a shortened URL pointing at your page of metadata? That page could offer metadata via BIBO, Open Graph and FOAF in RDFa, COinS, RIS, etc. using established methods to serve multiple applications at once. When Twitter annotations come along, the URL can be put in the annotation field. Eric On Apr 21, 2010, at 6:08 AM, Tim Spalding wrote: Have C4Lers looked at the new Twitter annotations feature? http://www.sitepoint.com/blogs/2010/04/19/twitter-introduces-annotations-hash-tags-become-obsolete/ I'd love to get some people together to agree on a standard book annotation format, so two people can tweet about the same book or other library item, and they or someone else can pull that together. I'm inclined to start adding it to the I'm talking about and I'm adding links on LibraryThing. I imagine it could be easily added to many library applications too—anywhere there is or could be a share this on Twitter link, including OPACs, citation managers, library event feeds, etc. Also, wouldn't it be great to show the world another interesting, useful and cool use of library data that OCLC's rules would prohibit? So the question is the format. Only a maniac would suggest MARC. For size and other reasons, even MODS is too much. But perhaps we can borrow the barest of field names from MODS, COinS, or from the most commonly used bibliographic format, Amazon XML. Thoughts? Tim -- Check out my library at http://www.librarything.com/profile/timspalding Eric Hellman President, Gluejar, Inc. 41 Watchung Plaza, #132 Montclair, NJ 07042 USA e...@hellman.net http://go-to-hellman.blogspot.com/
Re: [CODE4LIB] Twitter annotations and library software
Dear Tim, you wrote: So this is my recommended framework for proceeding. Tim, I'm afraid you'll actually have to do the hard work yourself. No, I don't. Because the work isn't fundamentally that hard. A complex standard might be, but I never for a moment considered anything like that. We have *512 bytes*, and it needs to be usable by anyone. Library technology is usually fatally over-engineered, but this is a case where that approach isn't even possible. Jonathan did a very well summary - you just have to pick what you main focus of embedding bibliographic data is. A) I favour using the CSL-Record format which I summarized at http://wiki.code4lib.org/index.php/Citation_Style_Language because I had in mind that people want to have a nice looking citation of the publication that someone tweeted about. The drawback is that CSL is less adopted and will not always fit in 512 bytes B) If you main focus is to link Tweets about the same publication (and other stuff about this publication) than you must embed identifiers. LibraryThing is mainly based on two identifiers 1) ISBN to identify editions 2) LT work ids to identify works I wonder why LT work ids have not picked up more although you thankfully provide a full mapping to ISBN at http://www.librarything.com/feeds/thingISBN.xml.gz but nevermind. I thought that some LT records also contain other identifiers such as OCLC number, LOC number etc. but maybe I am wrong. The best way to specify identifiers is to use an URI (all relevant identifiers that I know have an URI form). For ISBN it is uri:isbn:{ISBN13} For LT Work-ID you can use the URL with your .com top level domain: http://www.librarything.com/work/{LTWORKID} That would fit for tweets about books with an ISBN and for tweets about a work which will make 99.9% of tweets from LT about single publications anyway. C) If your focus is to let people search for a publication in libraries than and to copy bibliographic data in reference management software then COinS is a way to go. COinS is based on OpenURL which I and others ranted about because it is a crapy library standard like MARC. But unlike other metadata formats COinS usually fits in less then 512 bytes. Furthermore you may have to deal with it for LibraryThing for libraries anyway. Although I strongly favour CSL as a practising library scientist and developer I must admit that for LibraryThing the best way is to embed identifiers (ISBN and LT Work-ID) and maybe COinS. As long as LibraryThing does not open up to more complex publications like preprints of proceeding-articles in series etc. but mainly deals with books and works this will make LibraryThing users happy. Then, three years from now, we can all conference-tweet about a CIL talk, about all the cool ways libraries are using Twitter, and how it's such a shame that the annotations standard wasn't designed with libraries in mind. How about a bet instead of voting. In three years will there be: a) No relevant Twitter annotations anyway b) Twitter annotations but not used much for bibliographic data c) A rich variety of incompatible bibliographic annotation standards d) Semantic Web will have solved every problem anyway .. Cheers Jakob -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
I vote (heh) for d which will look a lot like c anyway, but with smatterings of owl:sameAs and Range-14 style 303's to keep things interesting. :) -- President BigBlueHat P: 864.232.9553 W: http://www.bigbluehat.com/ http://www.linkedin.com/in/benjaminyoung On 4/29/10 2:01 PM, Jakob Voss wrote: How about a bet instead of voting. In three years will there be: a) No relevant Twitter annotations anyway b) Twitter annotations but not used much for bibliographic data c) A rich variety of incompatible bibliographic annotation standards d) Semantic Web will have solved every problem anyway
Re: [CODE4LIB] Twitter annotations and library software
Hi, On Thu, Apr 29, 2010 at 22:47, Walker, David dwal...@calstate.edu wrote: I would suggest it's more because, once you step outside of the primary use case for OpenURL, you end-up bumping into *other* standards. These issues were raised all the back when it was created, as well. I guess it's easy to be clever in hindsight. :) Here's what I wrote about it 5 years ago (http://shelter.nu/blog-159.html) ; So let's talk about 'Not invented here' first, because surely, we're all guilty of this one from time to time. For example, lately I dug into the ANSI/NISO Z39.88 -2004 standard, better known as OpenURL. I was looking at it critically, I have to admit, comparing it to what I already knew about Web Services, SOA, http, Google/Amazon/Flickr/Del.icio.us API's, and various Topic Maps and semantic web technologies (I was the technical editor of Explorers Guide to the Semantic Web) I think I can sum up my experiences with OpenURL as such; why? Why have the library world invented a new way of doing things that already can be done quite well already? Now, there is absolutely nothing wrong with the standard per se (except a pretty darn awful choice of name!!), so I'm not here criticising the technical merits and the work put into it. No, it's a simple 'why' that I have yet to get a decent answer to, even after talking to the OpenURL bigwigs about it. I mean, come on; convince me! I'm not unreasonable, no truly, really, I just want to be convinced that we need this over anything else. Regards, Alex -- Project Wrangler, SOA, Information Alchemist, UX, RESTafarian, Topic Maps --- http://shelter.nu/blog/ -- -- http://www.google.com/profiles/alexander.johannesen ---
Re: [CODE4LIB] Twitter annotations and library software
On Tue, Apr 27, 2010 at 7:02 AM, Jakob Voss jakob.v...@gbv.de wrote: If you want to put bibliographic metadata into twitter annotations (good idea) you first need to clarify the basic purpose of embedding this information. I see two of them: I. Identification: To identify other tweets and resources that refer to the same publication II. Description: To nicely show which publication someone refers to. I think this is right. I wonder, would you consider a potential use case for Description to also provide machine readable data for a resource when a standard identifier is not known? It would be interesting to explore what identifiers + csl (and other options) would look like in a twitter annotation if you had time to mock something up in a wiki somewhere :-) //Ed [1] http://citationstyles.org/citation-style-language/schema/
Re: [CODE4LIB] Twitter annotations and library software
Hi it's funny how quickly you vote against BibTeX, but at least it is a format that is frequently used in the wild to create citations. If you call BibTeX undocumented and garbage then how do you call MARC which is far more difficult to make use of? My assumption was that there is a specific use case for bibliographic data in twitter annotations: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. II. Deliver a citation = use a citation-oriented format (BibTeX, CSL, RIS) I was not voting explicitly for BibTeX but at least there is a large community that can make use of it. I strongly favour CSL (http://citationstyles.org/) because: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations - there are dozens of CSL citation styles so you can display a citation in any way you want As Ross pointed out RIS would be an option too, but I miss the easy open source tools that use RIS to create citations from RIS data. Any other relevant format that I know (Bibont, MODS, MARC etc.) does not aim at identification or citation at the first place but tries to model the full variety of bibliographic metadata. If your use case is III. Provide semantic properties and connections of a publication Then you should look at the Bibliographic Ontology. But III does *not* just subsume usecase II. - it is a different story that is not beeing told by normal people but only but metadata experts, semantic web gurus, library system developers etc. (I would count me to this groups). If you want such complex data then you should use other systems but Twitter for data exchange anyway. A list of CSL metadata fields can be found at http://citationstyles.org/downloads/specification.html#appendices and the JavaScript-Processor (which is also used in Zotero) provides more information for developers: http://groups.google.com/group/citeproc-js Cheers Jakob P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
On Wed, Apr 28, 2010 at 4:17 AM, Jakob Voss jakob.v...@gbv.de wrote: P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } This looks really nice for the Description side. Has the JSON serialization for CSL been detailed anywhere yet? //Ed
Re: [CODE4LIB] Twitter annotations and library software
We've had problems with RIS on a recent project. Although there is a specification (http://www.refman.com/support/risformat_intro.asp), it is (I feel) lacking enough rigour to ever be implemented consistently. The most common issue in the wild that I've seen is use of different tags for the same information (which the specification does not nail down enough to know when each should be used): Use of TI or T1 for primary title Use of AU or A1 for primary author Use of UR, L1 or L2 to link to 'full text' Perhaps more significantly the specification doesn't include any field specifically for a DOI, but despite this EndNote (owned by ISI ResearchSoft, who are also responsible for the RIS format specification) includes the DOI in a DO field in its RIS output - not to specification. Owen On Wed, Apr 28, 2010 at 9:17 AM, Jakob Voss jakob.v...@gbv.de wrote: Hi it's funny how quickly you vote against BibTeX, but at least it is a format that is frequently used in the wild to create citations. If you call BibTeX undocumented and garbage then how do you call MARC which is far more difficult to make use of? My assumption was that there is a specific use case for bibliographic data in twitter annotations: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. II. Deliver a citation = use a citation-oriented format (BibTeX, CSL, RIS) I was not voting explicitly for BibTeX but at least there is a large community that can make use of it. I strongly favour CSL ( http://citationstyles.org/) because: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations - there are dozens of CSL citation styles so you can display a citation in any way you want As Ross pointed out RIS would be an option too, but I miss the easy open source tools that use RIS to create citations from RIS data. Any other relevant format that I know (Bibont, MODS, MARC etc.) does not aim at identification or citation at the first place but tries to model the full variety of bibliographic metadata. If your use case is III. Provide semantic properties and connections of a publication Then you should look at the Bibliographic Ontology. But III does *not* just subsume usecase II. - it is a different story that is not beeing told by normal people but only but metadata experts, semantic web gurus, library system developers etc. (I would count me to this groups). If you want such complex data then you should use other systems but Twitter for data exchange anyway. A list of CSL metadata fields can be found at http://citationstyles.org/downloads/specification.html#appendices and the JavaScript-Processor (which is also used in Zotero) provides more information for developers: http://groups.google.com/group/citeproc-js Cheers Jakob P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
Ed Summers wrote: II. Description: To nicely show which publication someone refers to. I think this is right. I wonder, would you consider a potential use case for Description to also provide machine readable data for a resource when a standard identifier is not known? There are lookup services to get a standard identifier when only some bibliographic data is known - mainly OpenURL. I have not investigated whether you can easily map CSL format to OpenURL or if you need to also embed the OpenURL as twitter annotation. However all lookup services that I know are either crapy or proprietary or both. This is not a technical issue but just based on a lack of data (hopefully to get better with more linked open data). Given enough open bibliographic data anyone can create a lookup service where you throw in some title, author and this stuff and get back an identified record. I think there also are some services called library catalog for this purpose. Anyway this os nothing that can be solved with a bibliographic data format alone. Either you have a standard identifier or you have not. If you have not, you must rely on third party services that run independent of your bibliographic data. It would be interesting to explore what identifiers + csl (and other options) would look like in a twitter annotation if you had time to mock something up in a wiki somewhere :-) I summarized my findings on CSL at http://wiki.code4lib.org/index.php/Citation_Style_Language and included some ideas of CSL and other data in twitter annotations. Feel free to modify! Cheers Jakob -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
I was also just working on DOI with RIS. It looks like both Endnote and Refworks recognize 'DO' for DOIs. But apparently Zotero does not. If Zotero supported it, I'd say we'd have a de facto standard on our hands. In fact, I couldn't figure out how to pass a DOI to Zotero using RIS. Or, at least, in my testing I never saw the DOI show-up in Zotero. I don't really use Zotero, so I may have missed it. --Dave == David Walker Library Web Services Manager California State University http://xerxes.calstate.edu From: Code for Libraries [code4...@listserv.nd.edu] On Behalf Of Owen Stephens [o...@ostephens.com] Sent: Wednesday, April 28, 2010 2:26 AM To: CODE4LIB@LISTSERV.ND.EDU Subject: Re: [CODE4LIB] Twitter annotations and library software We've had problems with RIS on a recent project. Although there is a specification (http://www.refman.com/support/risformat_intro.asp), it is (I feel) lacking enough rigour to ever be implemented consistently. The most common issue in the wild that I've seen is use of different tags for the same information (which the specification does not nail down enough to know when each should be used): Use of TI or T1 for primary title Use of AU or A1 for primary author Use of UR, L1 or L2 to link to 'full text' Perhaps more significantly the specification doesn't include any field specifically for a DOI, but despite this EndNote (owned by ISI ResearchSoft, who are also responsible for the RIS format specification) includes the DOI in a DO field in its RIS output - not to specification. Owen On Wed, Apr 28, 2010 at 9:17 AM, Jakob Voss jakob.v...@gbv.de wrote: Hi it's funny how quickly you vote against BibTeX, but at least it is a format that is frequently used in the wild to create citations. If you call BibTeX undocumented and garbage then how do you call MARC which is far more difficult to make use of? My assumption was that there is a specific use case for bibliographic data in twitter annotations: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. II. Deliver a citation = use a citation-oriented format (BibTeX, CSL, RIS) I was not voting explicitly for BibTeX but at least there is a large community that can make use of it. I strongly favour CSL ( http://citationstyles.org/) because: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations - there are dozens of CSL citation styles so you can display a citation in any way you want As Ross pointed out RIS would be an option too, but I miss the easy open source tools that use RIS to create citations from RIS data. Any other relevant format that I know (Bibont, MODS, MARC etc.) does not aim at identification or citation at the first place but tries to model the full variety of bibliographic metadata. If your use case is III. Provide semantic properties and connections of a publication Then you should look at the Bibliographic Ontology. But III does *not* just subsume usecase II. - it is a different story that is not beeing told by normal people but only but metadata experts, semantic web gurus, library system developers etc. (I would count me to this groups). If you want such complex data then you should use other systems but Twitter for data exchange anyway. A list of CSL metadata fields can be found at http://citationstyles.org/downloads/specification.html#appendices and the JavaScript-Processor (which is also used in Zotero) provides more information for developers: http://groups.google.com/group/citeproc-js Cheers Jakob P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
Unfortunately RefWorks only imports DO - not exports! We now recommend using RefWorks XML when exporting (for our project) - which is fine, but not publicly documented as far as I know :( Zotero recommend using BibTex for importing from RefWorks I think Owen On Wed, Apr 28, 2010 at 2:05 PM, Walker, David dwal...@calstate.edu wrote: I was also just working on DOI with RIS. It looks like both Endnote and Refworks recognize 'DO' for DOIs. But apparently Zotero does not. If Zotero supported it, I'd say we'd have a de facto standard on our hands. In fact, I couldn't figure out how to pass a DOI to Zotero using RIS. Or, at least, in my testing I never saw the DOI show-up in Zotero. I don't really use Zotero, so I may have missed it. --Dave == David Walker Library Web Services Manager California State University http://xerxes.calstate.edu From: Code for Libraries [code4...@listserv.nd.edu] On Behalf Of Owen Stephens [o...@ostephens.com] Sent: Wednesday, April 28, 2010 2:26 AM To: CODE4LIB@LISTSERV.ND.EDU Subject: Re: [CODE4LIB] Twitter annotations and library software We've had problems with RIS on a recent project. Although there is a specification (http://www.refman.com/support/risformat_intro.asp), it is (I feel) lacking enough rigour to ever be implemented consistently. The most common issue in the wild that I've seen is use of different tags for the same information (which the specification does not nail down enough to know when each should be used): Use of TI or T1 for primary title Use of AU or A1 for primary author Use of UR, L1 or L2 to link to 'full text' Perhaps more significantly the specification doesn't include any field specifically for a DOI, but despite this EndNote (owned by ISI ResearchSoft, who are also responsible for the RIS format specification) includes the DOI in a DO field in its RIS output - not to specification. Owen On Wed, Apr 28, 2010 at 9:17 AM, Jakob Voss jakob.v...@gbv.de wrote: Hi it's funny how quickly you vote against BibTeX, but at least it is a format that is frequently used in the wild to create citations. If you call BibTeX undocumented and garbage then how do you call MARC which is far more difficult to make use of? My assumption was that there is a specific use case for bibliographic data in twitter annotations: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. II. Deliver a citation = use a citation-oriented format (BibTeX, CSL, RIS) I was not voting explicitly for BibTeX but at least there is a large community that can make use of it. I strongly favour CSL ( http://citationstyles.org/) because: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations - there are dozens of CSL citation styles so you can display a citation in any way you want As Ross pointed out RIS would be an option too, but I miss the easy open source tools that use RIS to create citations from RIS data. Any other relevant format that I know (Bibont, MODS, MARC etc.) does not aim at identification or citation at the first place but tries to model the full variety of bibliographic metadata. If your use case is III. Provide semantic properties and connections of a publication Then you should look at the Bibliographic Ontology. But III does *not* just subsume usecase II. - it is a different story that is not beeing told by normal people but only but metadata experts, semantic web gurus, library system developers etc. (I would count me to this groups). If you want such complex data then you should use other systems but Twitter for data exchange anyway. A list of CSL metadata fields can be found at http://citationstyles.org/downloads/specification.html#appendices and the JavaScript-Processor (which is also used in Zotero) provides more information for developers: http://groups.google.com/group/citeproc-js Cheers Jakob P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com Email: o...@ostephens.com -- Owen Stephens Owen Stephens Consulting Web: http://www.ostephens.com
Re: [CODE4LIB] Twitter annotations and library software
- there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations A quick word of caution here; we got excited about citeproc-js until learning that it actually requires a specific extension compiled into the Javascript interpreter, E4X: http://gsl-nagoya-u.net/http/pub/citeproc-doc.html#javascript-interpreters This is fine and cool, but is not as widely supported as Javascript itself; eg. Internet Explorer, Chrome, Safari, and a number of server-side Javascript engines do not have E4X support: http://en.wikipedia.org/wiki/E4x That said, I'm very excited about CSL in general and this thread in particular — structured citation parsing is what I dream about at night. Great stuff. MJ
Re: [CODE4LIB] Twitter annotations and library software
Jakob Voss wrote: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. Ah, but for better or for worse, that's not the world we live in. We have LOTS of publications that either lack such identifiers altogether, or where information about identifiers is not available. (Mostly the former). That we need to identify. This is an actual use case, you can't just dismiss it by saying it can't be done! The biggest example is pretty much every scholarly journal article. (A significant _minority_ have DOI or pmid; the majority have neither). And we DO identify these articles, by a description meant to serve as identification, often by using OpenURL.Maybe we're not doing it seriously, but it's a real use case, and it's being done in the wild in production. Jonathan
Re: [CODE4LIB] Twitter annotations and library software
Jakob Voss wrote: There are lookup services to get a standard identifier when only some bibliographic data is known - mainly OpenURL. A standard identifier is not always _available_ -- even if you have access to a service to look up standard identifiers ( a not neccesarily realistic expectation for real world use cases) , not every publication HAS a standard identifier. Jonathan
Re: [CODE4LIB] Twitter annotations and library software
Has anyone actually gotten up a _server-side_ process that uses CSL to produce formatted citations? Using the citeproc-js with a certain custom compiled js interpreter, or anything else? This is what I'm interested in -- I'm not concerned with making it run in a browser, so custom compiled JS interpreter isn't a showstopper. But is still something that I'm not familiar with doing, so is going to take me a while to figure out how to set up. If anyone has already set anything up (using citeproc-js or anything else we may not know about), can you let us know, and maybe share your tips/instructions/code? Jonathan MJ Suhonos wrote: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations A quick word of caution here; we got excited about citeproc-js until learning that it actually requires a specific extension compiled into the Javascript interpreter, E4X: http://gsl-nagoya-u.net/http/pub/citeproc-doc.html#javascript-interpreters This is fine and cool, but is not as widely supported as Javascript itself; eg. Internet Explorer, Chrome, Safari, and a number of server-side Javascript engines do not have E4X support: http://en.wikipedia.org/wiki/E4x That said, I'm very excited about CSL in general and this thread in particular — structured citation parsing is what I dream about at night. Great stuff. MJ
Re: [CODE4LIB] Twitter annotations and library software
Jonathan Rochkind wrote: Jakob Voss wrote: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. Ah, but for better or for worse, that's not the world we live in. We have LOTS of publications that either lack such identifiers altogether, or where information about identifiers is not available. (Mostly the former). That we need to identify. This is an actual use case, you can't just dismiss it by saying it can't be done! Call me pedantic but if you do not have an identifier than there is no hope to identity the publication by means of metadata. You only *describe* it with metadata and use additional heuristics (mostly search engines) to hopefully identify the publication based on the description. But these additional heuristics are not part of the metadta while a well-defined identifier implies a standard of how the identifier had been created and how it can be looked up. The last hope if there is no identifier is to create one. For instance our library system creates internal record numbers (such as OCLC numbers) which can be reused. You can also define an algorithm that creates a hash as identifier like the bibkey I mentioned. But as long as there is no identifier there is no identification independent from a bibliographic database that already contains the record to search in. Jakob -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
Jakob Voss wrote: Call me pedantic but if you do not have an identifier than there is no hope to identity the publication by means of metadata. You only *describe* it with metadata and use additional heuristics (mostly search engines) to hopefully identify the publication based on the description. But the entire OpenURL infrastructure DOES this, and does it without using search engines. It's a real world use case that has a solution in production! So, yeah, I call you pedantic for wanting to pretend the use case and the real world solution doesn't exist. :) You can call it description rather than identification if you like, that is a question of terminology. But it's description that is meant to uniquely identify a particular publication, and that a whole bunch of software in use every day succesfully uses to identify a particular publication. It IS a hacky and error-prone solution, to be sure. But it's the best solution we've got, because it's simply a fact that we have many publications we want to identify that lack standard identifiers. If a twitter annotation setup wants to be able to identify publications that don't have standard identifiers, then you don't want to ignore this use case and how actually in production software currently deals with it. You can perhaps find a better way to deal with it -- I'm certainly not arguing for OpenURL as the be all end all, I rather hate OpenURL actually. But dismissing it as impossible is indeed pedantic, since it's being done! Jonathan
Re: [CODE4LIB] Twitter annotations and library software
I mean, really, if the folks at RefWorks, EndNote, Papers, Zotero and LibX don't have crash programs underway to integrate Twitter clients into their software to send and receive reference metadata payloads they can use in the Twitter annotation field, they really ought to hire me to come and bash some sense into them. Really. I still think by-reference payloads would got the farthest, as described at http://go-to-hellman.blogspot.com/2010/04/when-shall-we-link.html would go the farthest, but surely these folks know very well what they can send and receive. Eric On Apr 28, 2010, at 4:17 AM, Jakob Voss wrote: Hi it's funny how quickly you vote against BibTeX, but at least it is a format that is frequently used in the wild to create citations. If you call BibTeX undocumented and garbage then how do you call MARC which is far more difficult to make use of? My assumption was that there is a specific use case for bibliographic data in twitter annotations: I. Identifiy publication = this can *only* be done seriously with identifiers like ISBN, DOI, OCLCNum, LCCN etc. II. Deliver a citation = use a citation-oriented format (BibTeX, CSL, RIS) I was not voting explicitly for BibTeX but at least there is a large community that can make use of it. I strongly favour CSL (http://citationstyles.org/) because: - there is a JavaScript CSL-Processor. JavaScript is kind of a punishment but it is the natural environment for the Web 2.0 Mashup crowd that is going to implement applications that use Twitter annotations - there are dozens of CSL citation styles so you can display a citation in any way you want As Ross pointed out RIS would be an option too, but I miss the easy open source tools that use RIS to create citations from RIS data. Any other relevant format that I know (Bibont, MODS, MARC etc.) does not aim at identification or citation at the first place but tries to model the full variety of bibliographic metadata. If your use case is III. Provide semantic properties and connections of a publication Then you should look at the Bibliographic Ontology. But III does *not* just subsume usecase II. - it is a different story that is not beeing told by normal people but only but metadata experts, semantic web gurus, library system developers etc. (I would count me to this groups). If you want such complex data then you should use other systems but Twitter for data exchange anyway. A list of CSL metadata fields can be found at http://citationstyles.org/downloads/specification.html#appendices and the JavaScript-Processor (which is also used in Zotero) provides more information for developers: http://groups.google.com/group/citeproc-js Cheers Jakob P.S: An example of a CSL record from the JavaScript client: { title: True Crime Radio and Listener Disenchantment with Network Broadcasting, 1935-1946, author: [ { family: Razlogova, given: Elena } ], container-title: American Quarterly, volume: 58, page: 137-158, issued: { date-parts: [ [2006, 3] ] }, type: article-journal } -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de Eric Hellman President, Gluejar, Inc. 41 Watchung Plaza, #132 Montclair, NJ 07042 USA e...@hellman.net http://go-to-hellman.blogspot.com/
Re: [CODE4LIB] Twitter annotations and library software
Hi Tim, you wrote: Unless someone can come up with a perfect pre-cooked format—one that not only covers what we need but is also super easy and space-efficient (we have only 1/2k to use!)—Why don't we just decide on: 'simplebib' : { } and start filling in fields. I don't think it makes sense to externalize the information under another URL, at least in the first instance. That at least doubles the calls involved, and makes whatever you build dependent on lots of external services that may or may not work. Oh yeah, let's create just another ad-hoc metadata format, because obviously there are not enough different formats around there! To be honest: I admire your multitude of good ideas and efforts but this is one of the rare counterexamples. If you want to put bibliographic metadata into twitter annotations (good idea) you first need to clarify the basic purpose of embedding this information. I see two of them: I. Identification: To identify other tweets and resources that refer to the same publication II. Description: To nicely show which publication someone refers to. The purpose of identification can be served by the following means: a). standard identifiers b). standard identifiers c). standard identifiers Examples of standard identifiers include ISBN, OCLC Number, ASIN, LibraryThing Work-ID, well-defined bibliographic hash keys [*] etc. The purpose of description can best be served by a format that can easily be displayed for human beeings. You can either use a simple string or a well-known format. A string can be displayed but people will put all different citation formats in there. Right now there are only two established metadata formats that aim at creating a citation: a) BibTeX b) The input format of the Citation Style Language (CSL) I bet that CSL is the easier way to go. See http://citationstyles.org/ for details and examples. Cheers Jakob [*] See http://www.gbv.de/wikis/cls/Bibliographic_Hash_Key for a description of the mapping mechanism that is also used in BibSonomy to match BibTeX records. -- Jakob Voß jakob.v...@gbv.de, skype: nichtich Verbundzentrale des GBV (VZG) / Common Library Network Platz der Goettinger Sieben 1, 37073 Göttingen, Germany +49 (0)551 39-10242, http://www.gbv.de
Re: [CODE4LIB] Twitter annotations and library software
On Tue, Apr 27, 2010 at 7:02 AM, Jakob Voss jakob.v...@gbv.de wrote: The purpose of description can best be served by a format that can easily be displayed for human beeings. You can either use a simple string or a well-known format. A string can be displayed but people will put all different citation formats in there. Right now there are only two established metadata formats that aim at creating a citation: a) BibTeX b) The input format of the Citation Style Language (CSL) This isn't entirely true. There's RIS (http://en.wikipedia.org/wiki/RIS_%28file_format%29) and BIBO (http://bibliontology.com/) is starting to become quite common in the linked data sphere. There's also BibJSON (http://www.bibkn.org/bibjson/index.html) which I've had a browser tab open for months with the intention of actually looking at and actually seems quite well suited for how Twitter will store annotations. My opinion of it all along, however, has been very similar to yours -- why another citation format and why bind it so closely to a particular serialization? -Ross.
Re: [CODE4LIB] Twitter annotations and library software
-1 for BibTex! It can be hard to comprehensively parse without inadvertently creating garbage. Tom On Wed, Apr 28, 2010 at 1:00 AM, Ross Singer rossfsin...@gmail.com wrote: On Tue, Apr 27, 2010 at 7:02 AM, Jakob Voss jakob.v...@gbv.de wrote: The purpose of description can best be served by a format that can easily be displayed for human beeings. You can either use a simple string or a well-known format. A string can be displayed but people will put all different citation formats in there. Right now there are only two established metadata formats that aim at creating a citation: a) BibTeX b) The input format of the Citation Style Language (CSL) This isn't entirely true. There's RIS (http://en.wikipedia.org/wiki/RIS_%28file_format%29) and BIBO (http://bibliontology.com/) is starting to become quite common in the linked data sphere. There's also BibJSON (http://www.bibkn.org/bibjson/index.html) which I've had a browser tab open for months with the intention of actually looking at and actually seems quite well suited for how Twitter will store annotations. My opinion of it all along, however, has been very similar to yours -- why another citation format and why bind it so closely to a particular serialization? -Ross.
Re: [CODE4LIB] Twitter annotations and library software
Jakob Voss wrote: a) BibTeX Can I vote against BibTex, please? At the core of BibTeX is a language called 'BST' or that's the file extension used, which is as close as it comes to a name. This is an entirely undocumented language written to work on a patchily documented format. It's stack-based (not unlike PostScript), with special operation(s) to manipulate names based on deep assumptions about names and the ways they are formatted. These assumptions, by and large, hold for the personal names of North American English speakers (but I seem to recall is unable to correctly format the name of the President of the USA due to his title). The further you move from names of North American English speakers, the more they break (non-ASCII characters, eastern order names, complex titles, non-standard capitalisation, etc, etc, etc). BST is non-recursive, attempting to execute recursive functions gives the error Curse on you, wizard, before you recurse on me. Yes, the BST interpreter does refer to users as wizards, which seems less cool after the first 12 hours of debugging. Users have adapted to BibTeX by using an experimental approach---tinkering with the BibTeX entries until they 'look right,' which in most cases involves cramming everything into what BibTeX thinks of as the surname, because BibTex never omits or initialises the surname. If we're going to use a bibliographic framework, please, please, please don't make it BibTeX. cheers stuart -- Stuart Yeates http://www.nzetc.org/ New Zealand Electronic Text Centre http://researcharchive.vuw.ac.nz/ Institutional Repository
Re: [CODE4LIB] Twitter annotations and library software
On Wed, Apr 21, 2010 at 6:08 AM, Tim Spalding t...@librarything.com wrote: I'd love to get some people together to agree on a standard book annotation format, so two people can tweet about the same book or other library item, and they or someone else can pull that together. I'm inclined to start adding it to the I'm talking about and I'm adding links on LibraryThing. I imagine it could be easily added to many library applications too—anywhere there is or could be a share this on Twitter link, including OPACs, citation managers, library event feeds, etc. By this description alone it seems to me that OpenURL, perhaps implemented as some variation on COinS, would make the most sense. With OpenURL, the fields have already been defined. Perhaps the underlying JSON for the annotation could look something like the following: { 'annotations': { 'z3988': { 'contextobject': 'ctx_ver=Z39.88-2004amp;rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournalamp;rft.issn=1045-4438'} } } Additionally, one could specify an optional resolver parameter if so desired. Mark A. Matienzo Digital Archivist, Manuscripts and Archives Yale University Library
Re: [CODE4LIB] Twitter annotations and library software
Have you looked the the citation microformat ( http://microformats.org/wiki/citation) ? Don't know where work with this stands but it seems pretty interesting to me. Karen On Wed, Apr 21, 2010 at 8:21 AM, Mark A. Matienzo m...@matienzo.org wrote: On Wed, Apr 21, 2010 at 6:08 AM, Tim Spalding t...@librarything.com wrote: I'd love to get some people together to agree on a standard book annotation format, so two people can tweet about the same book or other library item, and they or someone else can pull that together. I'm inclined to start adding it to the I'm talking about and I'm adding links on LibraryThing. I imagine it could be easily added to many library applications too—anywhere there is or could be a share this on Twitter link, including OPACs, citation managers, library event feeds, etc. By this description alone it seems to me that OpenURL, perhaps implemented as some variation on COinS, would make the most sense. With OpenURL, the fields have already been defined. Perhaps the underlying JSON for the annotation could look something like the following: { 'annotations': { 'z3988': { 'contextobject': 'ctx_ver=Z39.88-2004amp;rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournalamp;rft.issn=1045-4438'} } } Additionally, one could specify an optional resolver parameter if so desired. Mark A. Matienzo Digital Archivist, Manuscripts and Archives Yale University Library
Re: [CODE4LIB] Twitter annotations and library software
On Wed, Apr 21, 2010 at 6:08 AM, Tim Spalding t...@librarything.com wrote: I'm inclined to start adding it to the I'm talking about and I'm adding links on LibraryThing. I imagine it could be easily added to many library applications too—anywhere there is or could be a share this on Twitter link, including OPACs, citation managers, library event feeds, etc. You might want to add it now, but I don't think annotations are available yet in Twitter. If you haven't seen it already Marcel Molina of Twitter outlined how annotations might work in a post he made to the twitter-api discussion list [1]. So the question is the format. Only a maniac would suggest MARC. For size and other reasons, even MODS is too much. But perhaps we can borrow the barest of field names from MODS, COinS, or from the most commonly used bibliographic format, Amazon XML. Thoughts? It sounds like a good idea to have a common pattern. I could definitely see a use case for wanting to aggregate conversations around books and such. //Ed
Re: [CODE4LIB] Twitter annotations and library software
whoops, forgot my footnote :-) [1] http://groups.google.com/group/twitter-api-announce/browse_thread/thread/fa5da2608865453
Re: [CODE4LIB] Twitter annotations and library software
I was wondering if there was a good microformat. The trick is that the citation format is very much about stuff that gets displayed, and lacks the critical linking ids you'd want—ISBN, SSN, LCCN, OCLC, ASIN, EAN, etc. If people know of others that would work, maybe that's the answer. On Wed, Apr 21, 2010 at 8:38 AM, Karen Coombs librarywebc...@gmail.com wrote: Have you looked the the citation microformat ( http://microformats.org/wiki/citation) ? Don't know where work with this stands but it seems pretty interesting to me. Karen
Re: [CODE4LIB] Twitter annotations and library software
So almost all of those identifiers can be formatted as a URI. Although sometimes it takes an info: uri, which some people don't like, but I like, for reasons relevant to their usefulness here. ISBN, ISSN, LCCN, and OCLCnum all have registered info: URI sub-schemes. I once tried to figure out how to express an EAN as a URI, and I think I _did_ eventually find _something_, but it was kind of confusing and hard to track down (The EAN/UPC/etc people have some info URI subschemes registered too, I think, but it's hard to figure out what it all means). For ASIN, I have been in the habit of using an Amazon http URI, the problem is that Amazon really offers several http URIs for the same ASIN, so you kind of just have to pick one format. Oh, and you can do DOI as an info: URI too. So your annotation _could_ simply be a URI. And get a lot of stuff. But this leaves out a lot of things that don't really have good identifiers at all: Articles in popular (not scholarly) newspapers/journals; most daily newspapers as titles themselves (don't usually have an ISSN); Movies; books too old (or for other odd reasons lacking) an ISBN (or lccn or oclcnum). Scholarly articles that don't have a DOI (the majority of them). Maybe you could use the citation microformat extended to take arbitrary URI identifiers? So for stuff without an identifier, you've got the citation details, but you can still stick identifiers in with URIs? And as someone else mentioned, this _is_ pretty much the use-case of traditional OpenURL, and it does handle it well enough: allowing you put enough structured citation in to identify the referent for things without identifiers, allowing you to put arbitrary URIs in rft_id. But OpenURL is kind of a monster to work with. And doesn't deal too well with certain kinds of citations like movies or music either, it's really focused on published textual materials. Jonathan Tim Spalding wrote: I was wondering if there was a good microformat. The trick is that the citation format is very much about stuff that gets displayed, and lacks the critical linking ids you'd want—ISBN, SSN, LCCN, OCLC, ASIN, EAN, etc. If people know of others that would work, maybe that's the answer. On Wed, Apr 21, 2010 at 8:38 AM, Karen Coombs librarywebc...@gmail.com wrote: Have you looked the the citation microformat ( http://microformats.org/wiki/citation) ? Don't know where work with this stands but it seems pretty interesting to me. Karen
Re: [CODE4LIB] Twitter annotations and library software
Unless someone can come up with a perfect pre-cooked format—one that not only covers what we need but is also super easy and space-efficient (we have only 1/2k to use!)—Why don't we just decide on: 'simplebib' : { } and start filling in fields. I don't think it makes sense to externalize the information under another URL, at least in the first instance. That at least doubles the calls involved, and makes whatever you build dependent on lots of external services that may or may not work. Best, Tim On Wed, Apr 21, 2010 at 10:45 AM, Jonathan Rochkind rochk...@jhu.edu wrote: So almost all of those identifiers can be formatted as a URI. Although sometimes it takes an info: uri, which some people don't like, but I like, for reasons relevant to their usefulness here. ISBN, ISSN, LCCN, and OCLCnum all have registered info: URI sub-schemes. I once tried to figure out how to express an EAN as a URI, and I think I _did_ eventually find _something_, but it was kind of confusing and hard to track down (The EAN/UPC/etc people have some info URI subschemes registered too, I think, but it's hard to figure out what it all means). For ASIN, I have been in the habit of using an Amazon http URI, the problem is that Amazon really offers several http URIs for the same ASIN, so you kind of just have to pick one format. Oh, and you can do DOI as an info: URI too. So your annotation _could_ simply be a URI. And get a lot of stuff. But this leaves out a lot of things that don't really have good identifiers at all: Articles in popular (not scholarly) newspapers/journals; most daily newspapers as titles themselves (don't usually have an ISSN); Movies; books too old (or for other odd reasons lacking) an ISBN (or lccn or oclcnum). Scholarly articles that don't have a DOI (the majority of them). Maybe you could use the citation microformat extended to take arbitrary URI identifiers? So for stuff without an identifier, you've got the citation details, but you can still stick identifiers in with URIs? And as someone else mentioned, this _is_ pretty much the use-case of traditional OpenURL, and it does handle it well enough: allowing you put enough structured citation in to identify the referent for things without identifiers, allowing you to put arbitrary URIs in rft_id. But OpenURL is kind of a monster to work with. And doesn't deal too well with certain kinds of citations like movies or music either, it's really focused on published textual materials. Jonathan Tim Spalding wrote: I was wondering if there was a good microformat. The trick is that the citation format is very much about stuff that gets displayed, and lacks the critical linking ids you'd want—ISBN, SSN, LCCN, OCLC, ASIN, EAN, etc. If people know of others that would work, maybe that's the answer. On Wed, Apr 21, 2010 at 8:38 AM, Karen Coombs librarywebc...@gmail.com wrote: Have you looked the the citation microformat ( http://microformats.org/wiki/citation) ? Don't know where work with this stands but it seems pretty interesting to me. Karen -- Check out my library at http://www.librarything.com/profile/timspalding
Re: [CODE4LIB] Twitter annotations and library software
On Wed, Apr 21, 2010 at 10:58 AM, Tim Spalding t...@librarything.com wrote: Unless someone can come up with a perfect pre-cooked format—one that not only covers what we need but is also super easy and space-efficient (we have only 1/2k to use!)—Why don't we just decide on: 'simplebib' : { } and start filling in fields. Because I don't think we've decided anything. I for one don't think we should have yet another arbitrary citation format floating around the Web. Mark A. Matienzo Digital Archivist, Manuscripts and Archives Yale University Library
Re: [CODE4LIB] Twitter annotations and library software
Just to clarify, encoding identifiers as URI's, my suggestion, is NOT externalizing the information under another URL. It is just picking a standard format for identifiers, the identifier format of the web, to re-use standards and cut down on custom vocabulary. If your 'simplebib' idea made sense, it could look like: 'simplebib' : { identifier: info:isbn:1234556X } or identifier: info:oclcnum:whatever etc. Note that info URIs not only don't need to be looked up from another URL to resolve -- info URIs are actually un-resolvable! While the ASIN http URI is (sort of) resolvable, it still doesn't _need_ to be looked up to resolve. Nothing is externalized. 'simplebib' : { identifier: http://amazon.com/asin/whatever or whatever. Likewise for OpenURL. Despite the name, OpenURL is, in practice, a standard vocabulary/encoding for citation details, it is not a method of 'externalizing the information'.This is an OpenURL context object in KEV format that identifies a particular book: rft.title=Manufacturing Consentrft.au=Noam Chomskyrft_id=info:isbn:whatever Etc. If you want to make up your own brand new citation format, then of course that is within your capabilities. It seems to me that trying to re-use as much infrastructure that already exists is good. Even if that's just re-using URI infrastructure (including info: URIs). Especially if you expect anyone other than you to 'adopt' this. Jonathan Tim Spalding wrote: Unless someone can come up with a perfect pre-cooked format—one that not only covers what we need but is also super easy and space-efficient (we have only 1/2k to use!)—Why don't we just decide on: 'simplebib' : { } and start filling in fields. I don't think it makes sense to externalize the information under another URL, at least in the first instance. That at least doubles the calls involved, and makes whatever you build dependent on lots of external services that may or may not work. Best, Tim On Wed, Apr 21, 2010 at 10:45 AM, Jonathan Rochkind rochk...@jhu.edu wrote: So almost all of those identifiers can be formatted as a URI. Although sometimes it takes an info: uri, which some people don't like, but I like, for reasons relevant to their usefulness here. ISBN, ISSN, LCCN, and OCLCnum all have registered info: URI sub-schemes. I once tried to figure out how to express an EAN as a URI, and I think I _did_ eventually find _something_, but it was kind of confusing and hard to track down (The EAN/UPC/etc people have some info URI subschemes registered too, I think, but it's hard to figure out what it all means). For ASIN, I have been in the habit of using an Amazon http URI, the problem is that Amazon really offers several http URIs for the same ASIN, so you kind of just have to pick one format. Oh, and you can do DOI as an info: URI too. So your annotation _could_ simply be a URI. And get a lot of stuff. But this leaves out a lot of things that don't really have good identifiers at all: Articles in popular (not scholarly) newspapers/journals; most daily newspapers as titles themselves (don't usually have an ISSN); Movies; books too old (or for other odd reasons lacking) an ISBN (or lccn or oclcnum). Scholarly articles that don't have a DOI (the majority of them). Maybe you could use the citation microformat extended to take arbitrary URI identifiers? So for stuff without an identifier, you've got the citation details, but you can still stick identifiers in with URIs? And as someone else mentioned, this _is_ pretty much the use-case of traditional OpenURL, and it does handle it well enough: allowing you put enough structured citation in to identify the referent for things without identifiers, allowing you to put arbitrary URIs in rft_id. But OpenURL is kind of a monster to work with. And doesn't deal too well with certain kinds of citations like movies or music either, it's really focused on published textual materials. Jonathan Tim Spalding wrote: I was wondering if there was a good microformat. The trick is that the citation format is very much about stuff that gets displayed, and lacks the critical linking ids you'd want—ISBN, SSN, LCCN, OCLC, ASIN, EAN, etc. If people know of others that would work, maybe that's the answer. On Wed, Apr 21, 2010 at 8:38 AM, Karen Coombs librarywebc...@gmail.com wrote: Have you looked the the citation microformat ( http://microformats.org/wiki/citation) ? Don't know where work with this stands but it seems pretty interesting to me. Karen
Re: [CODE4LIB] Twitter annotations and library software
I think Twitter annotations would be a good use for http://thing-described-by.org/ or a functional equivalent. The payload of the annotation would simply be a description URI and a namespace and value for descriptions by reference 1. the mechanism would be completely generic, usable for any sort of reference, not siloed in libraryland. In other words, we might actually get people to adopt it. 2. libraryland descriptions could use BIBO or RDA or both or whatever, and could be concise or verbose 3. descriptions could be easily reused I'll write this up a bit more and would be interested in comment, but it's where this post was going: http://go-to-hellman.blogspot.com/2010/04/when-shall-we-link.html Eric Hellman President, Gluejar, Inc. 41 Watchung Plaza, #132 Montclair, NJ 07042 USA e...@hellman.net http://go-to-hellman.blogspot.com/