I failed to convince my students about the benefits of code-generation

Add a Comment

Comment by Jordi Cabot on February 23, 2015 at 9:17

Lothar, I agree this could be a way to do it. Other people have also suggested other kinds of last minute changes (e.g. on the requirements) so that they can compare the cost of modifying the code vs modifying the model

Comment by Lothar Behrens on February 23, 2015 at 7:07

I think, one point should be shown when trying to convince students or professionals:

Measure their time to get used with a UI framework, the time to write a prototype manually as the source for a template (like XSLT) and the time to actually templatize the prototype. Just to get these times into the brain.

I had these times: Installation and trials = 3, creating a prototype = 3, writing the template = 1.

So to convince, each new employee to a team may have to learn the UI framework. (That was the case for me). Then tell them that a decision has been made to change the UI framework. Throwing away the code! (Happened in my professional work).

Also demonstrate repetition with several forms and one new feature to be added into a form template once compared to do this manually.

At the end compare the manually written code for each student or professional. Each will introduce different solutions and bugs. Measure the bugfixing time.

I think, only showing a tool to generate code that may has problems and bugs, will not convince. Indeed each tool will have their bugs and time needed to get used. I think, more aspects should be shown by comparing manual steps, repetition, bugfixing sessions and diversity of these bugs to help to convince (including a learning curve for an UI framework).

What do you think?

Lothar

Comment by Jordi Cabot on February 18, 2015 at 21:36

I don't want to signal out any specific tool. I´ve tried with two and both failed so I think it's a kind of general problem

Comment by Mark Dalgarno on February 18, 2015 at 21:11

What tool did you use?