Joy of SCM.
Aug. 10th, 2012 10:14 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
izardThere is only one aspect of software engineers day to day job that I genuinely hate. It is not writing documentation, keeping bug tracker up to date, or finding sporadic multithreading bugs. Those all are OK to fun on my scale.
Software configuration management tasks are something I consider the most boring type of work. That thing when you have to link to a library, which was last compiled on SunOs using imake, and depends on couple of others that are as dated.
So you have to dig in output of nm, objdump, ar, autotools, make, imake, shell, ldd, ant, etc.
So to make this mundane work more fun here is what I do. When it still does not compile and I have to add simple fixes to the source, I assign each change a risk on a scale [0 - no risk, to 4 - may crash everything. (I don't do "5" grade changes)]. After all changes are done and compilation is successful, I try using the software and see what is a correlation between software quality and risk estimation I made for my changes. No correlation so far, perhaps my estimations are flawed.
Software configuration management tasks are something I consider the most boring type of work. That thing when you have to link to a library, which was last compiled on SunOs using imake, and depends on couple of others that are as dated.
So you have to dig in output of nm, objdump, ar, autotools, make, imake, shell, ldd, ant, etc.
So to make this mundane work more fun here is what I do. When it still does not compile and I have to add simple fixes to the source, I assign each change a risk on a scale [0 - no risk, to 4 - may crash everything. (I don't do "5" grade changes)]. After all changes are done and compilation is successful, I try using the software and see what is a correlation between software quality and risk estimation I made for my changes. No correlation so far, perhaps my estimations are flawed.
