what to look at
Fri, Dec 28, 2007
Just a few things I should prolly look at next year.
- OpenID. “Poor man’s Shibboleth”. 2.0 is out, with new authentication and attribute exchange mechanisms. I should have a look at how Guanxi can support OpenID OP and RP functionality. I particularly like the combination of Information Cards and OpenID. Looks interesting. Adding OpenID to Guanxi would automatically turn Sakai into an OpenID OP and RP via the GSK.
- CardSpace. I quite like the idea of selecting identities graphically as Information Cards but the major problem for me is the Mac. It doesn’t support CardSpace, although the Bandit Project claims to fix this. Parallels seems a good way to develop CardSpace functionality if it wasn’t so unstable. Crashes on me all the time on the laptop and it won’t run on the G5 as it’s too old. There are also plugins for major blog and wiki products to allow you to login using cards, from PamelaWare. I particularly like to CardSpace enable the Guanxi SP.
- Guanxi 2.0. I’m a fair way through the Spring-ifying of the Guanxi SP. Once that’s done I can start to add different SAML profile support as well as the OpenID and CardSpace goodies.
- SAML 2. Must get back up to speed on SAML. Guanxi already talks SAML2.0 metadata but I need to reacquaint myself with the identity landscape.
- Blackboard. I’ve got a brief to investigate integrating our SITS student record system with our Blackboard VLE, to extend the current real-time account creation to the VLE. Since the demise of CLAN, VLE accounts are no longer in sync with enrollments and there’s a 48 hour lag between being able to login to your eDirectory driven portfolio of accounts and Blackboard. It seem possible to do this via a Building Block that speaks some form of web service lingo.
- C#. Should I have a look at C#? There’s a requirement for a major web service framework to allow long term storage of video recordings of lectures but the munger that allows you to format the recordings (quality/placement in stream etc) is a windows commandline application. I could just wrap it in a Java service but new languages are always good. Just have to ignore the hacking element of the dept and do what I think is best. Thing is, these days you can hack just about anything. You could wrap it in a Ruby web thingy, or a Java servlet or even a C application listening on a port. For just stuffing params into a commandline app, perhaps C# is overkill. We’ll see.
- Ruby. I should perhaps get up to speed with Ruby. I’ve looked into it a bit as Zoto were offering a free lifetime account for linking it to Facebook but they backed off when I explained how much work was involved and they would have to host the Facebook application. So my interest in Ruby waned after that. Have to have a reason for learning something.
- RDF querying. As part of CTREP I need to look into resource index queries for RDF. Fedora 3.0 doesn’t support RDQL any more so I need to look at either ITQL and SPO.