Sakai Status Update

Well it is a Sunday morning and the “Night before Alt-I-Lab” and it seems like a good time for a status update. I have been the Executive Director for two weeks and things are already changing – increasingly more of my time is spent on RFQ responses, talking about directions for the web site, server move logistics, etc etc. It feels good as these have been areas that I have kept on the back burner up until now but now am moving them to the front burner.


I think that the biggest communications activity for the next few months is the building of the www.sakaifoundation.org server – right now this is a redirect to www.sakaiproject.org. We really need to build a nice, ground-up web site here that is the foundation message without any “echoes” of the project which currently can be found all over the www.sakaiproject.org – by building the foundation website from nothing on up (we will keep all of the articles, press releases, etc), we can figure out what information folks really need and craft a site together that reflects the foundation and not the project.
I am looking at things like search logs for the site to try to get a sense of what people are wondering when they arrive on the site. This whole effort will take some time and we will get broad involvement in the design of the site once things start to get tangible.
One thing that I want to do with the foundation site is to use it to develop applications (say like the Sakai map on dr-chuck.com) – right now we often use confluence for things that are somewhat challenging in confluence (like maintaining the partner’s production status). Just as an example, we need to get this partner information into the foundation site and write some PHP/Python or whatever to keep the partner and production site information in a database and then reuse that information for things like the map or partner list – we need to get to the point where our important data and documents are not scattered all over the place in flat files.
So the foundation site will become an active development area beyond its current uses for conference planning and web presence.
I also want to switch to Joomla as well as tightly integrate our Joomla, collab, JIRA and confluence so we don’t have to keep four or five bookmarks for Sakai’s seemingly disparate servers. This integration will be done using web services and also depend on Aaron/Virginia Tech’s new Sakai LDAP server.
All this will happen after the server transition is complete. We are moving servers from UM to Virginia Tech and Indiana. This is a great contribution from both sites. This is a great indication of the increasing diversity of the institutions with deep commitments to Sakai.
— End of marketing bits – on to technical bits —
On the technical side of things, of course we slipped version 2.2 until July 15. We had shipped 2.0 and 2.1 right on time, so this is the first slip in a year. We need to focus on the 2.2 release to make sure it does not slip even more. Unfortunately, some of the code was not really as strong as it should have been at the code freeze. We had some lost time for QA because of missing / broken functionality in the 001- 003 releases of 2.2 – so we really did not fire up full QA until the 004 release. The good news is that 004 was pretty good and 005 is even better.
Probably OSP suffered the greatest loss of QA time due to the non-readiness of the 001-003 releases – hopefully OSP is now recovering and get sufficient QA done in time for us to make the July 15 release.
It is my feeling that any further slip would be pretty bad for Fall – we would likely end up with very very few sites running 2.2 in the fall – and with cool new stuff like Melete 3.0 being based on 2.2 – we need a representative group running 2.2 for Fall.
Also remember that we want folks to use Java 1.5 JVM’s in production this fall if possible and report back on results – we won’t allow requiring Java 1.5 for the 2.3 release until we know Sakai works well at scale with Java 1.5 on the major important operating systems. Developers are chafing at the bit to use the new 1.5 semantics as it leads to safer and cleaner code in many cases.
You my have heard that Samigo had good live scaling test where a number of people in the community used a MySql of Samigo at the same time and experienced no deadlocks and performance was satisfactory. There is still a ways to go and we still need to do a review of Samigo toward the end of summer – but this is promising and yet another reason to get sites running the 2.2 release – and another reason *not* to slip 2.2. If there is some interest in back porting the Samigo fixes to 2.1, this may be done.
There *will* be a number of sites using 2.1.x this fall – Rutgers and Yale are already committed to 2.1 for the fall. Those sites running 2.1 should get together and work on finding a branch manager to get important fixes into the 2.1.x branch throughout the fall – the Samigo improvements are a great candidate for the 2.1.x branch.
In terms of Samigo alternatives for Fall – this is a status update (sorry if folks wanted to keep secrets :) ). IU will give users the choice of Samigo or OnCourse Classic Quizzing engine. The proxy code to put OnCourse’s quiz engine into Sakai is already done (integration via single sign on and a simple proxy iframe) – they are working on getting grades back *into* the gradebook. Foothill has (for the moment) decided not to add any testing engine to Sakai other than Samigo for Fall – too much to do and too little time. Etudes users who need high-stakes or high performance testing will be encouraged to use/stay on Etudes classic for a while – Vivie will reevaluate in the Fall. Rutgers is working on the generic proxy tool and has a version that is substantially complete and Chuck Hedrick is proceeding on integrating his first external application and then will move on to integrating CAPA testing.
Next week is Alt-I-Lab in Indianapolis. The Common Cartridge demo will happen on Tuesday. In my last status updated I lamented that we might not make it. But thanks to Zach, we have a basic import capability that should suffice for the purposes of the demo.
I guess that is all for now.