Pages

Monday, April 21, 2008

How much do I hate Connotea? Let me count the ways...

Logo When I started writing this post, hate was definitely the right word, but maybe now it's a bit strong. My current opinion of Connotea would be closer to detest:

  • Cheesy interface, looks like it was coloured in by a five year old. I suppose this could be a matter of taste (as in, you don't have any).
  • Tiny user community! No amplification of knowledge through a non-existent social network - sucks compared to del.icio.us.
  • Link spam! Doesn't get removed by site admins very quickly, if at all. (Connotea has an ongoing spam problem).
  • Constant worry in the back of my mind that it is owned by Nature Publishing Group. What effect does this have, and to what extent is Connotea a walled garden? (Is Connotea controlled in the same way as Nature's From the Blogosphere, i.e. the bit of the blogosphere that Nature publishes?).

Connotea sucks

Grudging acknowledgment of the good points:
  • Handling of doi's works well. Most times I try to click on a doi, it breaks, but Connotea seems to make doi's work as they should.
  • Browser bookmarklet works well, interfaces with PubMed nicely.
  • Interfaces with a range of bibiographic software (Endnote, Reference Manager and has a plain text export option for RefWorks).

Sigh, see what I go through for you in my self-appointed role to walk the walk? Next week, I will be mostly using CiteULike, but if you're a Conntea user and you have a different opinion, I'd be very interested to hear from you.

Update: Neil has the figures.


22 comments:

  1. Connotea did not changed much since it was created although many users suggested some points that could be added or improved: ( http://www.connotea.org/wiki/RequestedFeatures ) however i keep using it just because I started my library with it. However I guess soon or later it will be merged with Nature Networks and it could become a more interesting tool. Moreover, I think that the 'network' option from del.icio.us is really powerful : that is why, each monthers , I transfer my bookmarks from connotea to delicious.

    Pierre

    ReplyDelete
  2. The overlapping nature of these resources is the really annoying part. If Connotea and CiteUlike mergered, then they would gain real traction with scientists.

    ReplyDelete
  3. @ajc This is always the way with online tools; multiple choices, pick the one that annoys you least - none of them are perfect.

    I want to like Connotea, but the implementation is just too poor for it to be usable for me. I find it ugly, slow, often broken or buggy and lacking basic usability features (batch operations, anyone?) I think it also suffers from an identity crisis - is it a reference manager or a general bookmarking tool? Other tools handle references better, del.icio.us handles bookmarks better. Especially annoying: abstracts not displayed at all for references scraped from web pages, but displayed in huge ugliness as "notes" if imported from bibtex/ris.

    CiteULike: not perfect either, but cleaner, more usable, more functional, better as a reference manager and a more responsive community, in my experience.

    Connotea needs more developers (it went from a team of one to a new team of - one) and I think they should have ripped it up and started again. The spaghetti of Perl CPAN modules in the backend goes a way to explaining the problems.

    Well, I've ranted enough.

    ReplyDelete
  4. Thanks Neil. The multiple sites are a real problem here as they subdivide the potential user base and suck the potential out of the resource discovery possibilities of a social network.

    ReplyDelete
  5. I've been using citeulike and I have to say one of the things I like about it is that it is NOT owned by nature. :-)

    ReplyDelete
  6. Getting votes for CiteUlike here. Question is, should that be our student exemplar, or should we just go for del.icio.us? I'm leaning towards del.icio.us for undregradate programs, CiteUlike for PG/PD.

    ReplyDelete
  7. It is definitely a post grad thing at this point.

    ReplyDelete
  8. Does Leicester have a resolver that can work out the url for a doi'd object, and then pass an authenticated you through your library to that resource?

    e.g. in the OU, we can do this - http://blogs.open.ac.uk/Maths/ajh59/014020.html - if you have OU credentials...

    ReplyDelete
  9. Yes, he said confidently, but I'm not sure how deeply it's buried (and I'll get that confirmed :-)

    ReplyDelete
  10. I looked at Connotea at one point last year but never returned. Don't know how - just didn't help in the the way that I hoped it would.

    As I read somewhere the other day,with new appl's spouting up every day, I think Web 3.0 will be a much more streamlined all inclusive on/off line experience.

    ReplyDelete
  11. I haven't tried citeUlike but I have noticed that I've started tagging papers using delicious which would have been obvious connotea candidates so I think I've subconsciously abandoned connotea

    ReplyDelete
  12. I'll be curious to see if you find CiteULike any better. I don't see that much difference between the two. I'd also be curious to hear how easy you find it to go back and forth between the two. One of the promises of both sites (and Elsevier's 2Collab) is that you can import and export your links and tabs effortlessly. I've found it to be a painful process, with lots of lost information as one goes back and forth. Good luck!

    ReplyDelete
  13. Hi Ajc,

    First disclosure, I'm the product development manager for Connotea, and I joined NPG about a year ago to take on the role.

    I'll just start by saying that I'm probably the least satisfied person out there with Connotea, in that I know all of her faults in detail and yearn to see her become what we all know she, and services like her, can become. That aside the tool does provide use for many researcers out there, and sometime I am deeply encouraced by how much it seems to help some people. It encourages especially as the kinds of things we want to do with Connotea will benefit these users even more.

    I am also deeply aware of how annoying a site can be to someone when things just don't work the way you expect them too. I totally accept that we need to improve responsivness of the API, and the interface to managment and search of one's libraries, amonst other things, but I'll get to that later. I'll address your concerns directly.

    About the style, there have been calls for switchable stylesheets, and that would be cool, but I think it is the least of the current issues, and when appropriate we will get around to dealing with it. In the mean time you can use a different style if you are using firefox: minimalist style.

    With regards to amplification of knowledge. All scientific specific sites will always be minor compared to a general site like delicious. The truth is that at the moment most scientists don't know about delicious, let alone more scientific related sites. We have had a steady growth in what I would call genuine users since launch, and that's probably to be expected. A steady growth over the next couple of years will lead to collaborative filtering working well, but you are going to have to pull in data not just from sites like connotea, but from doing things like mapping the relationship between usage and citations in the entire graph. The knid of work these guy's are doing with mesur. At npg we are thinking about how to bring all of this data together, but we are still in the early stages. In the meantime I understand that lack of support for bulk editing make Connotea a pain, and we are working on improving this.

    With regards to Spam, this was a much bigger issue last year than it is now. This has also been resposible for our drop in preformance as a fair portion of requests that come by are for spurious rss hits, and every few months we get really toasted by some bots, but we are getting better and understanding this and dealing with it. We are moving to get another machine up in order to handle rss and api requests specifically. If we didn't have to deal with the spam issue, then I think there would be something wrong with our site, however it has detracted from improving outward facing services in the past year, to our detriment, and we are chaging that focus now.

    About Connotea being an NPG site. I think that for at least as many people that worry about it an equal number choose to use Connotea as they feel that by being backed by a publisher that has been around for some time there is an implicit gurantee that the service will be around for a while. In the year that I have been involved this is the 3rd time that I have explicitly addressed this question, and it is a fair one. Part of our answer to that is that our code is open source, and we have an open API into our data. One advantage to a product like this coming out of a publishing company is that we are interested in technologies around the full spectrum of scientific communication. Working with people who develop Nature Network, or who look at content matching will feedback into Connotea and make it a stronger product, at least I hope ;). I'm clearly biased when it comes specifically to responding to this question, so I won't say any more.

    About bad matching of ads. That's pretty experimental at the moment and there are a few problems. When we used google ad-sense we mostly got links to tea companies, owing the the name of the site. The new system we are trying out allows us to populate specific buckets with content we pick. One of those buckets is Amazon content and right now we are getting a lot of matches to wrestling sites. I believe that algorithim is mathing "tag" to tag-team, but I think that should be fixed by the end of this week. My aim with this is to have specific informational content, like cross matching to other usrs in conntoea, in addition to more commercial links.

    I'll touch briefly on the codebase. We had hoped to encourage a larger developer community around the code, actully if anyone has any suggestions on what we can do with the current processes to make that happen then I'm all ears, however having been deployed in late 2004 has left us with a bit of a Heiz can of CPAN modules. I have been hearing calls from many quaters to do a re-write in an agile language, but there are a lot of dangers inherrent in doing a rewrite when you have a system that is sufficiently complex. The complexity, by the way, is mainly through evolution and not design, as some of the early transparently designed query structure was desperately slow. Since then one of my colleagues has been toying around with Google App engine just to see what one could do and he has produced pycite, which makes me think that this might not be totally suicidal, but I'm still very much on the fence over this at the moment.

    - Ian

    ReplyDelete
  14. @Ian - I always feel bad when Connotea is exposed to the harsh world of blog comments around the web; and you always pop up with good grace and humour to keep us informed! Good on you for putting up with it and best of luck with the continued development.

    ReplyDelete
  15. Ian, thank you for your thoughtful and detailed comment, and for contributing to the ongoing debate over the role of Web 2.0 tools for research scientists. I'll respond to some of your points in a follow-up post.

    ReplyDelete
  16. Maybe silly question but I can't figure out how to import bookmarks from Connotea to Delicious. Can please someone explain? Ty, Chris

    ReplyDelete
  17. I'm not sure you can - Connotea is a walled garden?

    ReplyDelete
  18. Connotea is not a walled Garden. We have a full open public API that provides more power and flexibility in terms of the queries that you can make against it than the equivalent delicious API. You can, for instance, use the API to query any public bookmark, not just your own, the results come back as RDF triples using some pretty standard markup such as foaf, and DC elements. The parse tree is reasonable to walk, I use the python wrapper for the API, there are wrappers in perl, ruby and javascript.
    (some links here:
    http://network.nature.com/blogs/user/ianmulvany/2007/08/14/java-wrapper-for-the-connotea-api-now-available)

    We don't output in JSON, that wasn't popular when the system was built, but we might do so in the future as it is clear that is is becoming pretty much a de facto standard.

    Any transaction that you can do on the site you can do through the API, from writing, editing, deleting, marking private. Documentation here:
    http://www.connotea.org/wiki/WebAPI.

    There is one very big issue with the API, response times are slow, There is a throttle on usage. We are pretty close to getting a second machine up and running that will be a dedicated machine for read only requests against the Conntoea DB, and when this comes online I hope to lift the throttle.
    http://www.connotea.org/wiki/Throttling

    Opneness and portability of the data is really important to us. I think that having an "export to delicious" button in the web app, and an "import from delicious" button in the app would be cool, but I'm working on some other things right now.

    Other people (a big nod to pierre here) have used the API for a variety of tools:
    http://www.connotea.org/wiki/ConnoteaTools

    You can use the API to get your bookmarks from Connotea to delicious, but of course why would you possibly want to do that ;) (that was a joke)

    ReplyDelete
  19. Thanks for responding Ian. Just to clarify, there's no way for us mere mortals to move libraries between delicious and Connotea without going into the API?

    ReplyDelete
  20. Hi AJC,


    conotea 2 delicious
    http://benlund.info/2007/01/30/contodel

    delicious 2 connotea
    http://ghastlyfop.com/blog/del2con.php

    Though I've not tested either for a while, so your milage may vary.

    Making an attractive tool that could do both is left as an exercise to the reader!

    Actually, in fairness, this raises one very interesting meme, or weakness, of web 2.0, if you will. The way that we have built our interfaces in web 2.0 have been developer centric, and not user centric. If you look at slide 68 of the following presentation:
    http://www.slideshare.net/mickstravellin/universal-mccann-international-social-media-research-wave-3
    you see that in a recent survey RSS uptake is less than 40% for regular Internet users. OpenID delegation is a case in point, possible if you run your own domain, but who does?

    Being a climber and a juggler when it comes to the numbers game then it becomes a pissing contest, and so I'm turned off by the term Web 3.0, it's all just web N.0 if you will, but I think that the architecture of collaboration will have to start disappearing waaay behind the UI's and will have to become implicit to use. A global way of communicating citations would be a good start.

    ReplyDelete
  21. Well I really do like Connotea for what I use it and I use it for strictly academic sources while delicious for absolutely everything, but as it is now, Connotea is just too small, not enough people use it to get the full picture, larger scale of opinions. I would like to search for users who tagged the same resources as I did also in delicious to find additional materials on the specific topic.

    @Ian: I have zero clue how to use API and Ruby script is like Chinese to me.

    Pierre wrote: Moreover, I think that the 'network' option from del.icio.us is really powerful : that is why, each monthers , I transfer my bookmarks from connotea to delicious.
    I was searching for this network option like insane the other day but wasn't able to find it. I was hoping someone knows this slick trick. :)

    ReplyDelete
  22. please never mind 'me searching the network option' comment

    ReplyDelete