Why you should not use commercial software in university courses

Sandra Porter has an interesting blog entry about using commercial software in university courses.
She states that she has big trouble with Microsoft Excel related to different platforms and software versions related to the application itself and the underlying operating system. She is surprised by all this trouble that is topped that users (and therefore students, too) are simply not able to read and to think while sitting in front of a computer. Well, guess what: This was to be expected. Everybody with a minimum experience regarding software architecture or software useability simply knows, that users are mostly not able to use the keyboard correctly, do not read error messages, do not read help files and are simply not able to follow step by step instructions.
This all was to be expected. The really interesting fact is that she got bitten by multiplatform and multiversion issues regarding Microsoft Excel. This was to be expected, too. Her requirement was to be able to do pivot tables.
Why Microsoft Excel is used for doing statistics is beyond me, there are open source packages for that kind of work (R comes to my mind, but there are surely others, too) and if it should really be an office application - why don’t use OpenOffice? There is even a nice tutorial for Pivot Tables. That would solve the portability problem, I guess.
She also states that being a Microsoft Office shooting star helps on finding a job. I don’t think that it is knowing the specific product that helps finding a job - it’s the attidue towards learning new software systems. It’s all about the priniciples regarding office suites. I know that I am not a Microsoft Office guru, but I find my way around it. But I know that I am able to efficently work with different programming paradigms (and programming languages, by that matter), a good hand full of object relational database management systems and relational database management systems and a whole range of operating systems. I use LaTeX, troff and DocBook where they solve the problem at hand. And I use other systems when they are more appropriate to the situation. Oh, and I did never have a problem finding a job. Microsoft Office knowledge was never even mentioned in any interview situation. It’s the attitude that counts.
So her final question remains: Should online courses require specific commercial software packages and specific versions? In my opinion they should of course not. University is not there to learn specific products. It’s there to teach students about the underlying principles. If a specific software is required to understand any course content then the course is not teaching principles.
And I am not even starting to ramble about the problems arising when students are enrolled in different programs and therefore need different versions of operating systems and applications. And I am not even thinking about the additional cost factor in purchasing all that stuff. Version chaos can be handled sort of by virtualizing installations - but how should students not even able to handle simple tasks like creating pivot tables be able to work with vmware? Perhaps a solution would be to hand out vmware images containing all required software - but then there would be all sort of legal issues. And what about the student with the dell laptop with this odd networking chipset? Not able to get on the internet with that vmware image. And the mac guys? No, specific software requirements are simple not worth the trouble. If that is a requirement for the course, the course should be changed, in my very humble opinion.
Always remember: Software is just a tool. And a fool with a tool is still a fool.

2 Responses to “Why you should not use commercial software in university courses”

  1. I disagree that spreadsheets are not good for stats. They are great for certain circumstances. In some cases, this is where students can actually calculate stuff “by hand” and get involved in the guts of a particular statistical procedure.

    Of course, using Microsoft Excel is a very very bad idea for all the reasons you cite and more.

    OpenOffice is good, but Gnumeric is even better. You should give it a try if you have not yet done so.

  2. I agree with you that spreadsheets are ok to do very basic statistics or to try to understand how a specific method works.
    I currently do only very limited statistics, but in my previous job I had to do a lot of it and used a mixture of sql and R.
    Of course I know about gnumerics, but at the end of the day I prefer R. One of the reasons is how easy it makes it to import data and process them by previously defined scripts - there are very few things I hate more than clicking the same buttons time after time, day after day :)

Leave a Reply