|
From: Andrew T. <an...@tr...> - 2008-09-09 00:09:00
|
Tom: > On the names of the rules, I think we should think about those names just > a bit more; once we distribute, those names will be widely distributed. > Slightly longer names don't hurt anything, and the descriptiveness can > be beneficial. I don't think people will ever *type* rule names, just select > them from a drop down or read them from a file. Hmm, I'm in the habit of opening Set Rule and typing in rules, so shorter is better as far as I'm concerned. I sometimes find it quicker than using the named rule menu. But I guess I'm not a typical user. I can see the advantage of descriptive rule names, but having to type them in docs and email messages will get rather tiresome. No big deal though. > I've never been a fan of the "EVN" name; I presume E is for extended, but > of course there are a million different ways to extend. I'm tempted to > suggest "Nobili-JvN-32" as the rule name for these; at the very least this > gives the novice user a search term to use when looking for documentation. Speaking of documentation, I'm in the middle of making some changes to the Set Rule dialog. I've added an algo menu which can change as you type in a rule, so you get immediate feedback on whether or not a rule is valid. I've also added a help button which will display a html file based on the current algo. For example, if the algo menu shows QuickLife then the help will display Help/Algorithms/QuickLife.html. The idea is that each <algoname>.html file will contain a description of the algorithm and the rules it supports, along with some example rule strings that can be copied and pasted into the rule box. > Following this convention, I suspect yours might want to be named > Hutton-JvN-32 (or however many states there are). The fact that we drop > Nobili from that name doesn't worry me too much; you made the last > changes and the provenance would be clear from searching for your name > and your description of the rules. > > I'm just not completely happy with three-letter rule names, especially as > we move into a future with many rule sets extended in various ways. > > Consider the precedent set by Mirek's Celebration, where the rule set > names are user-oriented. We also allow users to assign names to rules so I'm not sure how that is relevant. > I don't think we need to preserve any sort of backwards compatibility; > those files we ship will be corrected, and that's all there is to it. > Any files created up to now I would consider testing/alpha level files > and converting them to use the new names would be trivial. Yep, I've no problem with that. Anyway, we need to make a decision on this. Tom, jvnalgo is your baby so you should have final say on the rule names. I have a slight preference for the short names Tim suggested, but I don't really mind if we go for descriptive names. So the canonical rules should be: JvN-29 Nobili-JvN-32 Hutton-JvN-32 Correct? Andrew |