Ask HN: What information around commits and code would be useful?

[jpatil]

At tenXer we are working on tools to help development managers and developers identify software development problems and improve skills. We have been focused on getting information from project management tools to date but are now ready to look at commit level data from git.

When I say commit level data, I am referring to things like identifying if someone is thrashing on a piece of code (writing a lot of lines and rewriting them right away), identifying commits that have touched really old code (probably want to check for regression defects), identifying huge pull requests (hard to review), etc. Here are a couple samples -> http://cl.ly/image/1e1d2b0a0f0m http://cl.ly/image/0F2X2u1A3R2w

I was wondering if people had other thoughts around what might be useful information to extract. What might you want to know about your team based on their git activity?

Thanks!

[mtmail]

You already split by programming language, that should be useful.

I wonder if one could identify code vs configuration, maybe by pre-categorizing directory names.

Plus 'secure' certain files. E.g. a positive list of developers that is allowed to touch conversion-rates.csv. A warning after edit is still better than nothing.

[jpatil]

The second one would be a bit tough unless we knew previously what directories are associated with what.

We could definitely do the last one. How about sending an alert if someone who has hardly any experience in a particular language touches a file of that language. For example say I have little experience in js and I touched a js file, tenXer would send an alert saying that the lead or someone else on the team needs to review that commit more carefully? We could do the same with lines of code that are really old that have been changed or code that has been written by someone who is no longer at the company/ org.