We will be taking Connotea down for some much needed care an attention to the DB on Monday the 7th of September at 14.00 GMT (that’s 10 am New York time, 3pm London time).
We will be rebuilding the database indices and doing some other performance tuning on the database, so hopefully Connotea will be snappier when we come back up later on the 7th.
Hi everyone.
We moved Connotea to our new data centre about two weeks ago. There were a few days when the service was running off of a virtual machine while we shipped the dedicated hardware. This would have let to some service issues on around the 20th.
The new hardware was shipped, but we are having some ongoing issues with the re-setup of the system.
We are looking into the problems and when I have a better idea of what exactly is causing the current issues I’ll let you know along with what our planned fixes are.
Sorry for the problems,
- Ian
We are looking for students, undergrad, master’s or PhD students, who are working in the life sciences, to help us with a survey.
If you are based in the UK the survey is at http://www.insightexpress.com/s/Conn151043.
If you are based in the US the url of the survey is at http://www.insightexpress.com/s/Conn151021.
In fact anyone who is involved in any of the following subjects would be great: Agricultural sciences, Biological sciences, Psychology, "Clinical medicine or Health sciences.
We are trying to figure out what to build for the scientific community, and want to get some direction.
The survey is a little detailed, it could take up to about 15 mins to complete, so if you have any time to spare we would really appreciate it.
Earlier this year PLOS one rolled out support for Connotea. Pete Binfield, the managing editor wrote to point out that they have a blog post
describing the features, which include supporting bookmarking of items as well as reporting on whether an article has been bookmarked in Connotea.
Its great to see publishers starting to adopt integration with tools such as Connotea and CiteULike. PLOS has really been at the forefront of introducing new technologies to the sphere of academic publishing. As more and more information on the real time reading habits of academics becomes available we will move more towards a situation in which it will be possible to get an accurate picture of the current state of science.
Of course that’s not to say that the instant view is the one that should be the most important one, given the requirement for science to measure itself against the harshest peer-reviewer, stern reality herself. It still has to be better than the coarse tools that are currently predominantly used to peer into the trends of scientific thought.
OK, so you have grabbed a copy of the source for Connotea and you want to install it on your nice shiny new mac book pro. How do you do it? I’ve just written up a preliminary set of instructions for installing Connotea on OS X. It’s not trivial, but it is possible. I had to reinstall the stack recently after a software update from Apple broke my machine :(.
I’ll leave these instructions out in the wild for a while, and any feedback would be great. I’ll roll them into the source code in a little while when I get a chance!
I recently place the public snapshot of Connotea onto GitHub. We continue to use darcs internally for managing the flow of code from dev to live machines, however as the codebase has become larger the speed of transfers across ssh has become a real pain under darcs. I had been playing around with making a snapshot of the public repo in Git and a request from some groups interested in the most recent version of the code prompted my to place the Git copy that I had been working on out into the open. You can now go over and have a look at the code at the connotea-public page on github.
Our old snapshots still live in sourceforge, however I’ve moved away from using sourceforge for a whole bunch of reasons.
At the same time that I started playing around with GitHub the main NPG codebase is moving over to mercurial, so a move to mercurial is not to be ruled out at some point in the future for Conntoea, but for the time being it’s going to remain with git. Thankfully there are lots of utilities that preserve change history and allow you to move from darcs to git and from git to mercurial so any such changes should be painless.
Peter Suber has made a call for people in the OA community to tag items on Connotea as a part of the Open Access Tracking Project. The aim of the project is to get people to converge on using a common tag for tracking open access news. The proposed tag is oa.new.
Great! I’m a fan of open data, open source and open knowledge and I’m thrilled that people can find Connotea useful in organising their information. He raises a number of points around using Connotea for this, and just wanted to quickly comment on some of them.
By default we deliver the 10 most recent RSS items, however you can call a larger number by specifying the amount in the url like so: feed://www.connotea.org/rss/user/IanMulvany?num=100. I believe that in principle there is a limit of 1000 items that can be retrieved, but we would appreciate it if you restrict calls to a reasonable number.
One thing I’d like to call out. The links in our RSS feeds link to the bookmarked resources, and not to the Connotea page. We used to link to the Connotea page, but that didn’t seem right, so we changes this last year.
Something that is not well know, but might be of interest, the parsing engine can also display a library page in plain text using the following uri structure:
http://www.connotea.org/txt/user/IanMulvany
This can be handy if you want to quickly get data into a text document.
I’m not sure I see the problem here. The RSS feeds provide a time ordered list of when items were tagged, the the newest items on top. Over time one would expect the same item to appear in a feed if it gets tagged again by a new person.
If you go to the item page in Connotea this is a different matter. We do display one instance per person who bookmarked an item. There are two reasons for this. The first is that you see which tags different use has used for the same resource. The second reason is just one of code efficiency, it was just easier to reuse the template for a single bookmark in this way. It may indeed not be optimal, and if people are having a really hard time with this I’m willing to consider a change.
If Connotea users would like to automatically add the oa.new tag to items that they have already tagged with some other related tag then they can do this with the “Rename tag” feature. It’s a bit more powerful that just renaming. You can also use it to split a tag into multiple new tags. I’ve just renamed my “openaccess” tag to “openaccess” and “oa.new”, effectively rolling in the new tag to my previously tagged items, without losing the tag I had been using. We have discussed tag relations, and that is something I would like to introduce at some point in the future, but other commitments mean that we are unable to introduce this feature right now.
Well, last night we switched over to the new hardware. There are still one or two little niggles to iron out, but I hope this will bring considerably more stability to the service for the next few months.
Have a great christmas everyone!.
Well, after installing Connotea on our new hardware and testing it this week, we are going to transition the main site over to the new machines on Sunday.
We will be putting the service into read only mode for a while, and then the site will be down for a few moments as we switch over.
If the transition goes smoothly we will be back up with a much improved performance.
I know we have been having major problems over the last two months, but the new hardware has significantly more memory and so the query bottlenecks that we have been experiencing should ease off. It may take us a little while in the new year to get the most optimal server settings, but the new horsepower should help from the get go.
Markus Heckner from the university of Regensburg contacted me looking for volunteers to participate in a survey on the use of social bookmarking systems. They are looking for people who have at least 20 papers tagged in their Connotea library, if you are interested in helping out then head on over to http://pc51010.uni-r.de/limesurvey/index.php?sid=63888&lang=en. The full call for participation is below.
© 2009 Nature Publishing Group
