Mythic Logos

This is about CMSs and why HTML rocks while custom behavior is bad.

My first CMS usage was with Blackboard in 2004. That was horrible. I put some basic materials up and ignored it. I used my own blog software for my students (using perl) and generally it was not pleasant though it was fun.

Then I lived CMS free for awhile despite being an online instructor. That worked fairly well, in fact. After a time, I moved up into management and had to deal with CMSs for other courses. And we transitioned an old, homegrown system into Moodle 1.9. We had an individually paced model which no one uses and so it was a rather big nightmare to figure out how to work all of that. While I don’t have administrative access to Sakai or Blackboard, I have the feeling that Moodle was the most accommodating of  the three for us. It also seemed to offer the easiest and clearest way to have automatically graded tests in it. The transfer of materials involved a combination of PHP scripting and lots of button clicking. We used jsMath for displaying math and that worked well enough. Looking back on it, I wonder how stable it is. And you know what? It is great to be able to not care because it is someone else’s problem, a problem not of their own creating so solving it is a plus, not par. Win-win!

After striking out on my own, I ended up teaching a course covering all of math from algebra to statistics to calculus. I first taught it in the classroom and quickly evolved it into presenting slides and giving practice problems to students, thereby developing a variety of support PDFs, all of which were done with LaTeX, GeoGebra–>PSTricks, Beamer, and written with TextMate. Then I was asked to develop an online version.

So I spent a couple of months filming video lectures, i.e., far over the slides with my face in the corner. I recorded them with ScreenFlow and compressed them with ffmpeg into a very compact, but high quality recording which doesn’t play on iOS but works well otherwise.

So I had videos and I had PDFs. I was given Sakai to work with. I decided that each week should remain a single unit with the PDFs and videos being the delivery mechanism. So I wrote Python scripts  to generate HTML with embedded links for all of this material. The first two online sessions I hosted the files (PDFs and videos) on my own server and then I finally had my host institution host the videos; I uploaded the PDFs into Sakai directly–400 files loaded up with clicking! Argh.

When I first looked at Sakai, I thought about playing around with their auto-multiple choice, etc. test framework. But ultimately I decided that it was easier both for me and my students to go with PDF assignments and file uploads. It has worked well. And I am glad I went this route, as I will expand upon in a moment.

I also decided to work with the discussion boards, a wiki, and blogs all in Sakai. They worked reasonably well, but it was all rather disjointed. The wiki drove my students nuts and while I kept the material they had generated, I dropped the requirement to contribute in the third semester. The blogs and discussion boards worked well, but I feel that Sakai’s setup did not call attention to new stuff thereby dampening the interactions.

In summary, my use of Sakai was to host an organization of files, manage a gradebook, and have a managed place for students to generate and share text. They sometimes tried to share images but that was mostly a failure due to permission issues.

I have now been asked to convert into Blackboard. Which I did in less than two days, probably 10-15 hours of investigation and uploading, etc. Now done. And this is why I am writing this. Because I realized that I would have had issues had I used the system made tests. Extracting that kind of information out of a CMS is painful. The uploading files was actually trivial thanks to CyberDuck and WebDAV on the Blackboard site. Attaching them to the lessons was a little painful–a lot of clicking. But still, not as bad as uploading each file in a browser, one at a time.

So I modified my Python scripts to generate HTML for the new environment and then copied and pasted my way to completing the Lesson content. I turned off their rich text editor and it defaulted to having HTML-aware plain text. This was great because pasting in copied material was largely a mess due to styling coming along for the ride. Raw HTML works much better.

Setting up the blogs was trivial. Though when I was setting it up, I had a piece of helper text that I copied over and over again. And then, I said to myself, “Well, I am glad that I am clever enough to not make a mistake with that.” Oops, I had. So I had to do an extra 30 clicks to clean that bad paste job. Never think that you are clever. Ever. Try to avoid doing the same thing over and over again.

The wiki was somewhat easy. I could not copy the source directly because that was in special wiki language. In Blackboard, they use their standard HTML editor. So I had to cross my finger and paste the displayed HTML into the editor. It looks okay. Has some gunk, but not too bad.

The discussion boards and assignments. Well, I did not generate that stuff outside the system. So I had to go into Sakai, edit each blurb, click source, copy, and paste into the appropriate Blackboard bit. Again, the common language of HTML was great (copying the displayed output led to a very bad choice of fonts and gunk in the code). But it took a lot of effort because the primary source was in the CMS.

While I have not had to use it with students, so far, the setting up side has been nice. I was able to put all of a week’s tasks into a single place that they go to. I hope this helps them out. It was much easier to simplify in Blackboard than in Sakai. Also, all the inputs were consistent. I could formulate the start and end dates separately and then just paste them into various boxes without issue. Sakai had a different date format for every different type content–a product of open source plugin stuff.

The lessons I have learned about CMS is that for content, it is best to treat it as a HTML-front-ended file manager. Do not use anything that is custom to that environment. And produce everything outside the system as HTML. It is okay to generate it from something else, but having that HTML on the outside ready to copy is crucial to CMS agnosticism. And we all must be agnostic even if we stick to the same system–upgrades cannot be trusted.

The interactive stuff such as blogs, wiki, discussion boards, they all work well. Expect that stuff that goes in does not come out easily. But a big item I care about is whether new stuff gets flagged. I can’t tell that yet, but Sakai’s help manual on private messaging does not feel me with confidence: “Students are not notified if they receive a new message, so routine checks must be made for new messages.”  This is not good. The most important part of a communication system is that you know about an attempted communication. Moodle allowed a lot of notification to email options. Sakai grew less and less over new versions. What will Blackboard hold?

For interactive tests and the like, I think it is largely hopeless. The best that might be hoped for is support for HTML-CSS-JavaScript rather like how WordPress works with themes and stuff. JavaScript would need to be a secure subset. And one would need to talk with the backend. Accessibility becomes an issue. All in all, truly interactive content seems very distant. At the very least, good tools to generate the various possible formats need to be created. But it seems like from a content-testing perspective, the PDF route is the surest bet. It works, it is minimal hassle, and it is transferrable. But it breaks my heart.