On Fri, Apr 27, 2012 at 2:40 PM, Martin Koppenhoefer <dieterdre...@gmail.com> wrote: > Am 27. April 2012 20:14 schrieb Anthony <o...@inbox.org>: >> A default set to the value which is correct 99.999999999999999999999% >> of the time is not arbitrary. > > how would you distinguish between default values and incomplete > data/missing information?
You can't distinguish between them. That's why defaults should be set to the value that's right most of the time. At least, they usually should (sometimes other considerations come into play, for example if the cost of guessing wrong on one side is much higher than the cost of guessing wrong on the other side). Note that I'm talking here about situations where presenting the end-user with "unknown" is not reasonable. I wouldn't suggest for a map to show default speed limits when nothing was provided. (On the other hand, a routing program probably is going to have to provide default speed limits when calculating the fastest path, and this is a situation where "the value that's right most of the time" might not be the best one to use.) > We could have a tag > > defaults_checked=area;surface;lanes;oneway;lit;width;... > > to show which defaults have been checked, and maybe have different > default-schemes maintained by different people or for different > topics. We could define default sets in the wiki, > default_set_garry:version2=yes, with Versioning, so we can later amend > the defaults_sets without bothering... This doesn't sound reasonable. > But for simplicity when in doubt I'd set the tag explicitly. I think that's probably a good idea for most tags, where the split is 80/20, or 60/40, or 90/10, or where the cost of getting it wrong is very high. For something like railway=platform I wouldn't bother. _______________________________________________ Tagging mailing list Tagging@openstreetmap.org http://lists.openstreetmap.org/listinfo/tagging
