2010-10-12 Content model revisions
Following lengthy discussions we have, we think, much improved our use of Fedora content modelling by making our Hydra content models complementary and additive. This process is explained on the
Samvera objects, content models (cModels) and disseminators page.
2010-08-16 Revised Hydra specs
Following a very successful Hydra team meeting at Stanford last week we shall shortly be publishing on this wiki some revisions to the original Hydra specs. These build on the experience of creating Hydrangea and should form the basis of Hydrangea 1.0, the Hydra starter pack, which will be released later in the fall. Unlike beta 1, it is our firm intention that Hydrangea 1.0 should produce sustainable objects/content. We anticipate at least a beta 2 and an RC in between times...
2010-08-05 Hydrangea beta 1
The beta 1 release of
Hydrangea, a demonstration application of the Hydra stack, is now available.
The Hydra team will be presenting a paper in the main conference proceedings at Open Repositories 2010. This presentation will also signal the 'launch' of an alpha version of
. Attending will be Tom Cramer from Stanford, Richard Green and Chris Awre from Hull, and a representative (tba) from Virginia. In addition, Matt Zumwalt and Eddie Shin from MediaShelf LLC (who are working with us on Hydra/ngea) will be there.
Open Repositories 2010 (OR10 - Madrid, Spain, July 2010)
The Hydra team have submitted a formal paper for consideration by the conference organisers. We hope to contribute in various ways to both the main conference and to the Fedora specialist strand. Further, a workshop around Hydra is being discussed.
LibDevConX (23-25 March) will bring together some 30 invited participants at Stanford. One of the major topics will be Hydra. (Likely more of which anon...)
2010-01-20 New Year update
Happy New Year to all our readers... Apologies that the news has been a little sparse of late; we've actually been very busy and I am guilty of not updating this site as often as I should probably have done. So, let's catch up.
Fedora 3.3 and FeSL
Security for Hydra is a topic that has been greatly exercising us and we awaited the release of Fedora 3.3 including FeSL in mid-December with great anticipation. Late in December, working with the Islandora group at UPEI, we tested the first release and it seems to work 'as advertised'. FeSL will be needed, along with some inbuilt magic, to provide the necessary security options for a Hydra installation.
The revised Hydra content model approach agreed last autumn sees many of the metadata datastreams expressed as Fedora 'managed content' rather than 'in-line XML': this in order to keep our Fedora objects small and nimble. We now have these redesigned objects indexing and displaying correctly in our search and discovery interface.
Secured repository delivery
The Hydra team is meeting at the very beginning of February to discuss the next tasks, two of which relate to security: how to index the repository so that users are not aware of objects that they are not going to able to have, and how to deal with delivering content from a secured repository. Some proof of concept work has already been done.
Work is under way to provide management functionality through a Hydra client application ('Hydrangea' as a working name...) with a target date of Easter for an internal test release.
Hull proof-of-concept site
The proof-of-concept site described below is well past its 'take-down-by' date! However, we shall leave it in place for now until we can demonstrate the secured delivery facilities mentioned above.
2009-10-06 Proof-of-concept site at Hull
Today Hull has exposed its 'proof-of-concept' site to the web. As it says on the home page, this is not an implementation of the current Hydra specification (rather, the preceding one) but it will give you some idea of where Hull is going in implementing Hydra's work. We'd be grateful if you wouldn't spread the URL widely until later in the year when we have a site up to the latest spec.
2009-09-2 thru 4 Hydra meeting, Charlottesville VA
The team met for three days at UVa to review progress and to plan the next year's work. Early tests to implement Hydra structures have taught us a lot and as a result we are slightly changing the way we define Hydra objects and how this is represented in Fedora content models. The team felt that our work in this area was now in a state where we could go public with it and so our documents on content models and disseminators will now be maintained here on this wiki 'in public view'. We have discovered a compelling need for Fedora to allow 'optional' datastreams to be defined in a content model and we are pressing for this enhancement.
2009-08-18 Update: Hydra search and discovery interface
Hull is pleased to announce that it has a prototype Hydra search and discovery interface running. The system is based on a Fedora 3.2.1 repository containing Hydra-compliant objects (compound and complex) with a customised Blacklight over the top. The Solr index which drives Blacklight derives its content from the FOXML (mainly the descMetadata MODS datastream; the multiple download links for a complex object are derived from the METS in the parent's contentMetadata datastream). The index updates automatically to reflect any change in the object. A customised Blacklight gives essentially the same splash page for a simple or complex object (in the latter case we have set the children not to be independently discoverable - it's a switch in the RELS-EXT); a thumbnail is displayed on an image splash page. The prototype supports OAI-PMH harvesting. Administrative functions and authorisation-based filtering are not yet implemented.
It is anticipated that a production version of this software stack, containing Hull's publicly accessible repository content, will be available to view by the end of September.
Friday last week saw a demonstration by Hydra's Stanford contributors of their ETD solution (submission workflow and display). This is now virtually finished and will go into production at the beginning of September. Whilst this is not Hydra it is a solution built on the same technologies and from it the Hydra tools will be generalised out.
Hull and Virginia are working with the repository side of things. Hull hopes to have a live system for displaying their ETDs by the end of September. The Solr indexing side of things seems to be in place (based mainly on a Hydra MODS datastream) and it remains to put Blacklight over the top and customise it. This implementation process has raised a number of issues about the design of Hydra objects which will be discussed in September (see below). (You may have noticed that our Stanford colleagues have just launched a new, Blacklight-driven version of their Libraries search system SearchWorks.)
The Hydra team is meeting at UVa at the beginning of September for an intensive face-to-face session.
2009-05-26: OR09, Atlanta GA, 18-21 May
Well, Hydra seems to have created a bit of a stir at OR09 if Twitter tweetings were anything to go by. Thanks to all the people who came up to us during the conference to express interest and wish us well.
Now that the Hydra idea has been officially launched, this will be the one-stop shop for information about the project and I will do my very best to keep it up to date for you. Health warning, though: please note the disclaimer on our home page. Just because we say here "we are thinking about" does not mean that an idea will make it to any final product. This wiki space is a snapshot of ideas in time - not a contract!
To answer here a question that I was asked several times in Atlanta: thanks very much for the thought, but no we do not need extra collaborators at this time. The team is nicely manageable as it is and we already have several groups lined up to do extended testing when the time comes. The code will be made available to the community soon (watch this space) at which time you will be welcome to try it and feed back to us if you will, however you'll understand that this process will not be supported by the project team. - RG