Zacker.org

My university decided to run with Sakai -- despite the comp sci people having a successful Moodle running on site for years.
I'm just going to summarise what I've learned from managing the Sakai pages for courses with enrollments of over 500 -- including four continuous assessment MCQ quizzes on each course.
1) Textbook publishers (who control the testbanks) have either never heard of Sakai, or don't want to hear about it.
2) I have yet to find any way of successfully converting WebCT XML to something that Sakai can import. In fact a technical rep for Thomson basically told me they supplied their testbank in "XML" (unspecified) and making it work with Sakai was my problem.
3) I wrote my own macros to import questions from .rtf testbanks, removing tables and unwanted formatting, spitting out a plain text file, uploading same into Claremount's Sakai Quiz converter and (finally) uploading the Claremount XML file into Sakai. Currency symbols needed manual re-coding along the way.
4) Attaching individual graphs/figures to quiz questions was ridiculously labour intensive and unreliable. Figures must be hosted remotely (not uploaded to Sakai) causing problems for naive end-users who get scared by IE warnings of secure/insecure page content.
5) In my experience, about 2% of quizzes mess up. Either the timed-test shuts down without warning, or the student's grade never makes it to the gradebook.
6) Sakai doesn't recored how long a student took to complete a test; leaving no way to verify if a student claims their test crashed after 20 seconds.
7) I'd like to see a way to code equations in LaTeX and include these in testbanks.

The URL to Mellon's page on Sakai is now http://rit.mellon.org/projects/sakai-sakai-foundation/.

I don't know how reliable most of your stats are, but the Alexa numbers are completely bogus. Alexa only counts hits by people who have installed Alexa's stupid toolbar. All that graph says to me is that Moodle users are almost 8 times more likely to install crap on their systems than Sakai users (if anything, that's a win for Sakai). Since you started your post with the Alexa stats, it makes the rest of your stats seem suspect even though the rest of them may actually be accurate.

Full disclosure: I work at a small university (<2500 students) that's currently running Blackboard Basic Edition. We're looking for a free replacement and are evaluating Moodle & Sakai. I don't really care who wins, but I do get annoyed when people try to cite Alexa stats as if they have any value whatsoever.

Thanks

The statistics/graphs presented here don't seem to be particularly relevant to the comparison of the two tools. You mention Moodle has been around a lot longer than Sakai, so how is a chart of install bases or community traffic related to the quality of the tool itself? Of course a tool that's been around twice as long as the other is going to have more community traffic.

I found this site in search of some answers to some security issues on Moodle. All the comments are very interesting.

I set up a Moodle site one month ago after trying to figure out how to establish an online quizzing system for a programming class I teach. I chose Moodle over other options because of its buit-in, ready-to-go, latex-friendly quiz generation capabilities. I wanted to be able to create hundreds/thousands of randomized questions that require special import. At University of Alabama, we have a large WebCT installation and its quiz generation capabilites are quite restrictive (note: UA will change to different system next year that will cost millions-not sure what it is yet). In theory, Sakai can do this as well. After reading through documentation for a day, I still wasn't clear about how to get it all set up (ditto for LON-CAPA).

Although it took me a while to get all of the Linux settings correct on my DIY 64-bit system before I got Moodle to run, since going online it has been doing great so far. Is it better than alternatives? I don't know. It does what the documentation says it will do, and it is set up in a way that allows for straightforward modifications to do additional things.

If Sakai (or LON-CAPA even) is superior, the people creating these systems up need to start creating documentation that clearly spells out what the systems can do and how to do it in language that idiots like me can fathom. (Note to the Sakai institutions: Although all your faculty and students may have genious mentalities that require minimal instruction to do amazing things, the vast majority of teachers and students in the world need a little detailed guidance.)

The fact that Moodle has many more installations than Sakai may have little to do with technical superiorities or user features. Documentation and installation for Moodle are vastly superior to alternatives, and Moodle works on virtually any system.

Octave (free) is a fantastic clone of MATLAB (~$2000 commercial), but few people use it because the interface stinks and the documentation (relative to MATLAB) is quite poor. If somebody took about 8 months to create a smooth interface that would work well under Windows (yes it sucks but it is the standard), Octave use would take off.

It may very well be that Moodle will scale to handle 100K users, but frankly I do not care. With respect to a centralized system, I appreciate the advantages of the central data coordination where I do not need to input class participants, or create/manage user accounts. However, I would much prefer a model where a central facility coordinates BIG picture details (like enrollments and course/database backups), and offloads course offerings to specialized (smaller) servers. My one small server easily handles the complex quizzing requirements for 100+ students at a time. Furthermore, I can set up my system to use many back-end features that might bog down or be unacceptable for a centralized system. Every class does not need every feature, so why create a burden on central server resources by loading processes accessed by a small fraction of users.

I am currently enrolled in a Sakai enabled course at IU. I also happen to have downloaded Moodle and have a test install running on my personal web site (as a class project in online course facilitation). From my perspective as a Sakai user, I am really disappointed with what I have seen so far in the IU Sakai enabled system. I certainly cannot comment on the backroom functionality or integration with other university systems, etc. However, I find working with the "basic" LMS features within Sakai to be extremely frustrating. Given that it is a tool designed to support online and distance education, it is far from "user friendly" and has done little to create a sense of community with class members. As a prime example, the discussion feature is a disaster. There is no way to track previously read posts from one session to the next. I literally have to "expand all" then squint to read the date posted to guess which I have seen - a real treat now that we are in our 5th week of class with hundreds of posts to scroll through. In one class I am taking, the students actually set up a Yahoo! Group to facilitate our communication needs. Obviously, I am not alone in my feelings as IU has (again) delayed pulling the plug on their old system. The old system is not without flaws, but it is still used by many professors who have switched back after trying the new system. Just $.02 from a lowly user.

This is one of those Godzilla vs. Mothra issues: Spectacular to behold, but do you want to take either one home as a pet? K-12 schools around our area are using Moodle since it is such an easy download and very quick to get up and running. My issue is not based on a popularity contest but rather functionality demands. As an observer from 30k FTE university, I appreciate the suggestions that Moodle could be run from several installation areas (i.e. individual colleges or depts) but what about the demands to integrate them all with Student Information Systems, LDAP authentication, Institutional Research data culling for student tracking for assessment, etc? In other words would the multiple integrations be an ITR nightmare? I mean this sincerely because nothing would warm my heart than to see open source step up to the plate and hit a homerun, but it has to be major league not the minors...

Hi, great piece and mostly I think very accurate. One small point (which I don't think invalidates your larger point at all) - the 8,900 'installed base' number for Moodle is, I believe, based on a built-in tracker that counts every single running instance of the software. The vast majority of these are not production installations. Guaranteed. I have 2 on my desktop and another on a server that are part of this total that have never delivered a single course.

Still, even if we were to just look at actual production installations delivering online education, I expect the number was easily 3000, quite likely more. Clearly trumping Sakai almost a hundred-fold. Not wanting to be pedantic, but I have seen that 8000+ number thrown around quite a bit, and its inflation actually debases what is in fact a valid argument.

Hi, great piece and mostly I think very accurate. One small point (which I don't think invalidates your larger point at all) - the 8,900 'installed base' number for Moodle is, I believe, based on a built-in tracker that counts every single running instance of the software. The vast majority of these are not production installations. Guaranteed. I have 2 on my desktop and another on a server that are part of this total that have never delivered a single course.

Still, even if we were to just look at actual production installations delivering online education, I expect the number was easily 3000, quite likely more. Clearly trumping Sakai almost a hundred-fold. Not wanting to be pedantic, but I have seen that 8000+ number thrown around quite a bit, and its inflation actually debases what is in fact a valid argument.

This is very, very helpful. Happens that I'm in the middle of evaluating Sakai and Moodle for a new online education project in West Africa. So thanks for making my work easier. What does this mean? -- Moodle: 150 per million
Sakai: 20 per million

Enterprise scalability.

Your analysis, whilst interesting, fails to take into account critical factors for large scale deployment. It's not *how many* sites are using Moodle, but *how large* they are. Large schools need to run multi-server resilient architectures that can suffer partial failure and continue to deliver service. This is the architecture that Sakai is working towards. Yes, it is more complex, but it is also more powerful.

Moodle is simpler to install and maintain, but is still working towards enterprise scalability. Moodle is good for small schools without the need for enterprise-class scalability and resilience.

You have also failed to take into account Sakai's use in other contexts outside learning and teaching, which is growing rapidly. Sakai is often not badged as 'Sakai', so web statistics may be somewhat indicative, but are not an effective comparison.