Help talk:Contribute

From Rosetta Code

On testing task examples

What should be the minimum amount of testing done on posted code?

I would like to think that all code was at least run, and produced some answers that are present in other solutions or mentioned in the task description where applicable. I have twice, recently seen code posted and "untested" posted against it. I would hope that the poster was aiming to test their code shortly, and post the result, but it would be good if any code, especially for more than a couple of lines/operators, was tested to at least fulfil the task description, as the poster understands it. What do others think? --Paddy3118 07:35, 1 December 2009 (UTC)

Certainly that's the ideal, and it's not a very high ideal either. I prefer to put some output from the task's solution with the code, and it's easier to generate that by solving the task in the first place. (I also hate to put up wrong solutions anyway...) Maybe we should cook up a template ({{untested|lang}} perhaps) to mark the untested solutions so they stand out as needing attention (and they can also be marked for work by the maintenance bot)? –Donal Fellows 09:05, 1 December 2009 (UTC)
Done. —Underscore (Talk) 12:35, 1 December 2009 (UTC)
Thanks. I've even been inspired by that to go through and fix a bunch of pages that weren't using templates right so that now everything links together. Of course, it's now clearer that a number of things are not of good quality, but that just means that it is time for someone to get busy and fix them. –Donal Fellows 16:38, 1 December 2009 (UTC)
+1 on the thanks. --Paddy3118 19:20, 1 December 2009 (UTC)