On 2006 Jul 1, at 8:33 PM, Graeme Gill wrote:
> Ben Goren wrote: > >> The OpenBSD approach, which I rather like, is to use a config >> file for things that get that complex. I think this'd rather >> make a lot of sense for Argyll, especially seeing how so much >> of the time you'll be using the exact same options...and how >> the chances of you remembering exactly what ``-c 4'' is >> supposed to mean three weeks after you read the manual for the >> first time are virtually nil.... > > This approach has its merits, but it also has it's > problems. Too often I've struggled with systems that have > hidden configuration, and you're wondering "why is it doing > *that*!". Installation is made more complicated, and only by > reading the manuals can you figure out where the config is.
A valid complaint, but it's pretty easy to deal with: have the first line the tools print in everything but ultra-silent modes be ``read config from /foo/bar/baz.config''
> For some stuff it's great, for other things, not so great.
Always about balance....
> For projects that have a complicated/hard to remember setup, > I've simply creates shell scripts to encapsulate the knowledge.
I'm /almost/ at that point with Argyll. Right now, I've just got a file with some notes that I copy / paste as needed.
But, considering just how simple it all really is, I might never get that round tuit I need to put a shbang at the top of the file....
Cheers,
b&