Image directories by RSS aggregation

A report published last year on Community Led Image Collections suggested that:

  • visual directories showing a selection of images from each relevant collection might be a useful means of disseminating information about those collections; and
  • while sophisticated repository interoperability specs might be a long term aim, such a directory could be kick-started using more lightweight technologies, specifically by aggregating RSS feeds (with images as enclosures) supplied by the image collections.

The JISC wanted to test the viability of this idea and so provided, through the Engineering Subject Centre, some money that allowed a colleague (Lisa) and me to try to build and populate an image thumbnail RSS aggregator.
We [and by “we” I mean Lisa] found that building one was technically feasible, but perhaps not quite as easy as might be hoped. More importantly we found cultural or organizational barriers within the image collections that prevented us receiving the necessary feeds to populate it.

One of the reasons for suggesting RSS as a means of aggregation was the large number of tools that support it. We had no problem in finding tools that could help image collection owners create an RSS feed for us, we suggested they use Feed for All if they couldn’t figure out how to do it themselves. On the aggregation side, however, we had more problems. Many of the tools that displayed enclosures seemed to presuppose that the enclosure was audio/video content (i.e. that the feed was a podcast). So for example, FEED2JS displays a “play” icon instead of the contents of the enclosure. Not what you need for a visual directory. We had better results with RSS2HTML though you need to use iFrames to allow more than one feed to be displayed per page. The problem here was that the pages tend to load very slowly: it relies on real-time connexions to multiple repositories downloading several images (which aren’t always thumbnail sized) from each in order to display just the one page. A more AJAXy sort of approach might help get round this, so we tried Grazr. That version loads faster, but lacks the immediate visual appeal. Perhaps a better approach would be to cache the RSS feeds on the image directory server, but that required a little more effort than we could put into a very short project.

This general approach does seem to offer some interesting possibilities. I would agree that a visual directory makes sense when trying to show what image collections are available in a given discipline area. Rather than try to describe the contents of an image collection show some thumbnails. Focusing the sample of images on the interests of the user also makes sense: having a directory of selected engineering image collections on a site for engineers is a start, but it would be better if the user could enter a search term and the directory be built by generating an RSS feed from image records in the collections that match that term. This is all technically feasible.

We did come a across two non-technical snags. Firstly, the image collections didn’t have the type of images that engineering academics wanted to use in their teaching materials. Secondly, none of the collection owners we approached implemented any sort of RSS feed (even though several expressed a keen interest). There’s not really much of a technical angle on the first problem, but the second was revealing. It relates to various organizational issues in how these collections were managed. Lack of continued funding was cited by some as a problem, and that doesn’t help but I don’t think it is the whole problem. Separation of technical expertise from academic input seems to me to be closer to the root of the issue; this has two manifestations. Firstly is that in the real small collections, typically run by an individual enthusiast, the view seemed to be that if something was available via a web page then that was enough. If anyone wanted to provide a service based on those images they were available (scrape away!). Often the technical implementation behind such image collections didn’t support much else. For the larger collections, the case for providing an RSS feed was often successfully made to the manager of the image collection, but frequently they had no control over the technical effort. For example the manager might be in an academic department and the technical effort provided by an IT service department; one might be interested but the other will be overstretched with different priorities.

If you’re interested, full report is available from our website.

3 Responses to “Image directories by RSS aggregation”

  1. Martin Poulter Says:

    Thanks a lot for this, Phil. It backs up the point that more academic sites need to get into the spirit of embeddability.

  2. psychemedia Says:

    Hi Phil

    Your post raises some interesting issues relating to something I’ve been mulling around – how we in the OU can better (re)discover media assets that have been produced for previous Open University courses:

    Course Content Image Search:

    An advantage of getting images out via feeds, is that they can be embedded in a page via a slideshow widget. There are lots of third party ones around, and a couple that are built maybe to (presumably) industry strength – e.g. the slideshow widget:
    or the carousel widget:
    built on top of the Yahoo/YUI libraries;

    Or the Googe AJAX Feed API slideshow widget:

    The SeaDragon tool from Microsoft looks like it could provide an interesting interface over large image collections:


  3. Phil Barker Says:

    Thanks Tony, they all look like they merit further investigation. And the SeaDragon demo is very impressive. Phil.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: