The lessons from eNATIS were not learnt

Nine years ago, an electronic system for license renewals was rolled out in South Africa. It was called eNATIS, and the rollout was a mess. After a lot of money and time was wasted fixing things, it was eventually a success. But numerous principles of proper Project Management weren’t followed.
eNATIS had not been properly tested, configuration tested nor performance tested, and had been rushed out in a big bang approach instead of being implemented on a few pilot sites to make sure everything was working before being generally rolled out. Tasima, who coded, implemented and maintained eNATIS, is currently embroiled in a legal fight over non-payment by the Government.
Last week, we had eNATIS Round Two with the rollout of an online registration system for pupils in the Gauteng Education Department (GDE). It is abundantly clear that the lessons that should have been learnt in the wake of the eNATIS fiasco weren’t.
The first time I learnt of the registration system was hearing about it on the radio while driving to work. A news report said that the system had crashed under the sheer load. It was immediately obvious that the system had not been properly load tested, otherwise this would not have happened. And that was just the start.
I have read reports that the site was buggy and error-ridden. In addition, if a person made an error, there was no way to go back and undo the error. The result is that there are duplicate registrations because one has wrong information. this is a violation of a central tenet of usability: enable users to detect and correct errors.
I test software for a living. Even though I don’t have a project management qualification, I’ve worked on enough projects, both properly and badly managed, to spot the signs of a badly managed project. This has a lot of the signs.
Firstly and most obviously, the project was rushed into live without being properly tested and was riddled with bugs. There should be testing at all stages and the code should not be allowed into the next phase until all the major bugs are found and fixed. This has the appearance of a vanity project, where a go-live date is given and the code is implemented, regardless of whether or not it is ready.
Secondly, the system was rolled out across Gauteng instead of for just a few schools. Unless a project is exceptionally well-managed (and sometimes even then), there are teething troubles on go-live. Deploying the system on a few pilot sites allows errors to be detected and fixed without too much trouble. It also helps users and the people maintaining and running the system in the back end become familiar with it. In fact, operating manuals often include things that were learnt in this phase.
The third sign is the bad usability of the system. A properly specified system should include a good user interface. User testing should have begun in the design phase. Had that been done, the system may have avoided a lot of the duplicate applications.
Fourthly and finally, non-functional testing was not properly done. Good performance testing (incuding load and soak testing) would likely have revealed the system needed recoding or more hardware. Usability testing would have exposed the problems with the interface and led to one that was less likely to cause user error. It also makes me wonder if security testing was done, and if so, how thoroughly.
I mentioned above that this has the air of a vanity project. Reinforcing that impression are news reports that some school governing bodies have said that the system is currently illegal (online registration is currently not an option in the law), and that they’re prepared to go to court if the GDE persists in going ahead with this.
One last thing has me concerned: the articles I’ve read seem to imply that the GDE intends to only use online registration.
I sincerely hope I’ve misunderstood. There are still a great many families in Gauteng who have no access to computers, never mind the internet. Mandating the online route would prejudice them, and any of their children.

Advertisements

About autismjungle

I am a Software Test Analyst. Shortly before I turned 21 I was officially diagnosed, although I had long suspected I was autistic. Welcome to my blog
This entry was posted in Life, Software, Work and tagged , , , , , . Bookmark the permalink.