On Mon, 06 Jun 2011 09:16:25 +0200, Stefan Götz <stefan.goetz@xxxxxxxxxxxxxxxxx> wrote: > Hm. I'd prefer a consistent rule for the complete code base because > code tends to move between daemons and the library. Personally, I > don't think that name clashes between HIPL code and other code will > ever be a big problem which is why I'm not a big fan of hip_ (which > should be hipl_, anyway) but then that's just my gut feeling. But this > is nothing too important - I'm fine with any rule, including the one > you suggest above. I ran into this issue when working on midauth in the firewall: There's code speicific to hipd, specific to hipfw and also shared between those (in modules/). Applying the least amount of creativity possible, this resulted in names like (don't remember exactly) hipfw_check_challenge_response(...) and hip_check_challenge_response(...), for example, the former calling the latter. Without prefixes, I could have come up with "process_challenge_response(...)" or similar purely to avoid a name clash of course, but I'd have to remember which verb belongs to hipfw or hipd then every time. Another example: main(argc,argv) parsing parameters and delegating control to hipd_main() (or hipfw_main()), rather than "run_daemon()" or something. That's just an annoyance of course (and you could argue that adding a prefix every time is even more annoying). And this applies to cross-module functions only... prefixes pose no benefit to static functions IMHO. -- https://code.launchpad.net/~stefan.goetz/hipl/hidb-lsidb/+merge/61833 Your team HIPL core team is subscribed to branch lp:hipl.