Ensemble: an OER aggregator

I’ve been working with the Steeple podcasting folks to build an aggregator for searching feeds of Open Education Resources. The result – Ensemble – reuses various parts of Jopml plus some of the code found inside FeedForward for handling Opml processing and for feed autodiscovery.

You can try out Ensemble here.

So far I’ve just aggregated the feeds available from a few places, such as Oxford and the Open University, but it shows how the concept works. Next up I’m tweaking the indexing a bit, and putting in place some code to handle creative commons licenses more intelligently. The idea is that this would sit as a piece of middleware, and that widgets or other web apps can dynamically grab OPML or JSON for user queries.

Hey, that makes 3 spin-offs from one app, not bad going…

FeedForward: A Personal Information Application

I just submitted a paper that summarises the findings of the FeedForward project; it covers a lot of aspects of the project such as the design, the underlying approach to workflow, and the issues we had in implementation and also with interoperability. Worth a read if you’re interested in developing similar applications, or are thinking about the future of the information environment and the sort of standards we can use for scholarly communications.

Its CC-BY licensed so feel free to take sections and reuse them.

While this paper marks sort-of-the-end of the formal project, much more is happening in the world of FeedForward…

New release (M2.2) – on the final stretch to 1.0

We’re nearly there! The new release is now available to download. This is the last planned beta release before we make it to the 1.0.

We didn’t add many new features this time around, but we do have an About box. Not the most attractive ever, but at least you now know what version you’re trying out (thanks to Tore for pointing that out!).

Probably more significantly we’ve added a new Preferences dialog, with options to enter your web proxy settings (for those poor souls at Strathclyde Uni and elsewhere that need them – thanks John!) You can also tune how quickly FeedForward scans your feeds:

(Note that as we use a feedcache it won’t actually impact the sites too much if you crank it up.)

Hopefully if there are no new bugs, we can have the 1.0 launch in the next couple of months.

FeedForward: 2nd place for “Most Innovative Project”

I arrived at the JISC Repositories and Preservation programme meeting to find out that FeedForward had been voted 2nd in a competition for the title of “Most Innovative Project”. Nice!

Anyway, I had to improvise a “poster” and demo in the corridor by the coffee machine. Thankfully Dave Flanders was on hand with a room full of big pens and flipchart paper for me to borrow:


FeedForward posters

FeedForward M2.1 released

The releases are coming thick and fast as we get close to – finally – the 1.0 launch. We only plan one more beta build – M2.2, next month possibly – before we reach that target.

Milestone 2.1 adds the ability to locate and add feeds directly from Jopml (see the previous blog post), improved OPML import, and the ability to export collections in RSS, Atom, BibTex and RIS (EndNote) formats. It also closes a number of bugs. (Amongst other things we’ve used a zip format for the download that even Windows Vista understands!)

You can download the new release here.

The M2.2 release will add web proxy support, a preferences dialog, and an About box, and close the remaining small bugs ready for the 1.0 launch.

Jopml & FeedForward

One of the things we really wanted to do in FeedForward from the outset was to try to make use of “feed registry” services to get rid of some of the hassle of locating and subscribing to feeds. This was particularly as we wanted new users – possibly also new to using a dedicated RSS reader – to be able to just specify their area of research and get going straight away.

However, our initial choice, IESR, turned out to be problematic. Basically, while there was a goood set of machine APIs we could make use of, there just wasn’t enough usable content. While a lot of repositories had information in IESR, not many had their RSS registered, even where it was supported.

Our second choice came about through getting involved with the JISC Users & Innovation Programme: the tictocs project created a repository of table-of-contents feeds for academic journals. Tictocs collected loads of useful feeds, and worked with the publishers to improve their support. This was just the content we needed. Sadly, there was no API.

Eventually, I just gave in and built my own service – Jopml (Journals into OPML)!


Jopml is a Ruby on Rails application built using the data exported from TicTocs (as a .csv dump), but makes it easy to write a machine interface by building OPML output directly from the query. So http://jopml.org/feeds?q=polymers gets you the HTML search results, while http://jopml.org/feeds.opml?q=polymers gets you the same but in OPML, ready to import into a feedreader.

With this service in place I could add new functionality into FeedForward. First, you select the “Import journal feeds from Jopml.org” menu option:


Jopml menu option in FeedForward

Next, you enter some search terms and click “Search”. The dialog box updates to show the number of hits. When you’re happy, you click OK to add them to your application.


Jopml dialog box

(Note that as the service uses Ferret you can use booleans, phrases and all the usual Ferret query options)

And you’re done!


Feeds added

This functionality is currently in the source code in subversion, but should make it to the next release.

While I developed Jopml to add this feature to FeedForward, there is also nothing to stop anyone from making use of it for their own applications. If you make something interesting, let me know!

New release looms closer…

We’re steadily working closer to the next release. Its been a long time coming, but I think is worth it. We’ve actually removed a lot of features for this release, as we get a better idea of what works and what’s useful, and focus on improving the usability core features – we’ll put features back in if anyone requests them, but I think a “less is more” approach is better at this point.

We’ve moved to a new version of Eclipse (Ganymede) which is a lot faster and supports more platform specific features which caused some headaches in M1 (drag and drop in particular had some quirks). We’ve also been improving the UI in various ways, as you can see in this screenshot:


We’ve got a bit more testing to do before the release, but feel free to check out the code from Subversion if you’re keen!

New website!

Well, not really a “new” website, but its had a bit of an overhaul, and we’ve put in a feedback widget so you can tell us what you’d like to see in FeedForward. You can also now access the site from a slightly less naff URL – its now http://getfeedforward.org

The M2 release is finally nearing completion after a pretty chaotic last half of 2008, and should be up on the site as soon as we’re sure that this time it really, really will work on XP and Vista!