From: V. D. <mic...@gm...> - 2006-05-24 14:33:14
|
I would like two point to Don's list: On Wed, 2006-05-24 at 06:09 -0600, Don Allingham wrote: > Language has little to do with anything. The problem with options boils > down on several issues: > > 1) Requests for features are many. The number of developers is few. We > have to prioritize what we do. Unlike some OSS programs, we have no > corporate sponsorship - no one gets paid. All work is volunteer, and > the developers have jobs and families. So the amount of time that > developers can contribute is limited. > 2) The number of options increases the number of cases that have to be > tested. Assuming that all options have only two possibilities, you > find that the number of test cases is 2^n, where n is the number of > options. 1 option = 2 test cases. 4 options = 16 test cases. 8 > options = 256 cases. And suddenly the possibilities for errors > explodes and the code starts getting difficult due to the many > possible branches. 3) There's no point in adding features and say: "if it's not what the user expected, he can always go back, tinker with the options, until it's ok". Each option that goes into Gramps should be the best solution for an existing problem. Let me clarify, because I understand this point sounds as if I'm not worried about your concerns. The problem is, it is véry easy to design features in the export & filters area that don't make sense for *anyone*. (Especially if you start thinking about the sources that are linked to a person, and point to other people as well, etc... Or this one: filter for all males, and then provide an option to include their wives and families -on export, not in the filter itself-... I guess there's a point, but it's not very compelling.) Features that don't make sense for anyone are a burden on the program and its usuability. 4) Also, even if options are included, Aunt Martha will thank you if the defaults are right for her. Therefore, options are not the end of the debate. In any case, nothing is lost by thinking about this issue... and I don't think we're there yet. Also, my views are not very relevant, since I'm quite sure I'm never going to use anything like this. This was my last e-mail about this issue written without understanding what it's about. If there will ever be a next e-mail, I promise it will bring more to the debate! :-) Michaël |