Le 21/09/2016 à 17:35, Vard Antinyan a écrit :

We have undertaken a task to assess code complexity triggers and generate recommendations for developing simple and understandable code. Our intension is to share the results with you, developers, so everyone can learn the triggers behind complex software.

Thanks, it's quite interesting.

Well, I have some remarks.

I don't understand the lack of examples with the questions. Just four lines of code each time, then say whether you are actually talking about 40 lines or 15, so that we can have a visual representation of what you want us to think about, and can see which ‘lengths’ you imagine. It would give the questions much funnier while it's quite hard to remember of an experience of ugly code (well, I just clean it once, then it's become beautiful and I can work), or even imagine such an experience in my poor mind, for a newbie of collaborative projects like me.

For example, see this question: ‘Generally very_long_names_of_functions_or_variables in a unit of code make that code ...’ Reading this, I imagined a code with six lines having method calls (so very elementary results) with method names just as long as `very_long_names_of_functions_or_variables`, which quickly makes the code hard to reuse. A small example would help having a reference.

Other remark: I don't have anything of a diploma in philosophy or even a simple interest, but the meaning of the word ‘complex’ seems to be very central for completing the survey. But well, my first definition of a /complex/ code didn't sound interesting to me. To me, ‘complex’ qualifies abstract things. An algorithm or an architecture can be complex, just because it takes many things from the base language, has many components. For code used by developers, it's more… complex, you see! The interesting thing with this survey is the actual result, the influence of bad code on the productivity of developers; so how large is the need to explore code and the influence on motivation, when trying to add a feature, to reuse existing code. Then I had to formalize this definition to complete the survey. Anybody might be able to do it, however I think being a bit more concrete to participants, would be far more productive. Furthermore, maybe my definition is not strictly yours; maybe you're only focusing on readability and understanding (just as if you'd check that a source code does not include backdoors — do you?), rather than real hacking. Maybe you don’t mind? It would be weird…

So, I hope this makes sense. I got average results everywhere, except for comments that I absolutely don't mind about, from my personal experience.

Regards, -- Sylvain