2009/2/24 sly (sylvain letuffe) <[email protected]>: >> mess for someone who wants to use our data. Can't hear that argument >> 'let the application developers decide what they want' any longer! > > Looks like an holy war is gone start. Let's try however to keep it low. But > I'm a part of that war against duplicates tags. > > Quoting "Dave" : Handeling duplicates is not complicate, I don't agree. > Handeling one is easy, handeling hundreds become more and more complicate for > nothing.
a) that's not a quote, it's a paraphrase. b) you're taking it out of context. I'm not advocating hundreds of duplicates, or encouraging duplicates in any form. So here is a proper quote: "given that there /are/ two tags in use, why go to all the effort to change it?" See the difference? I also said a number of times that handling duplicates was relatively trivial -- it is. It's a pain in the arse, but a trivial pain the arse. So sure, duplicates add complication, and it would be better if they weren't there. That doesn't mean we run a bot. I don't care about the tags, I care about the bots. They're a blunt tool that more often than not will fly out of control if people start using them frequently. > > I can't rembember how many (oneway='yes' or oneway='true' or oneway='1') there > are in the mapnik style's sheets I use. > > not talking about tunnel='yes', bridge='yes', and of course the later > amenity='doctor(s)' > > Not that this is a enormous deal for developper, but I think it is one more > confusing thing about OSM for begginers. > > I'm in favor of disussion, describing only one, removing the others from any > documentation, removing slowly support in renderers and slowly, replace old > values by new values. That sounds like a more sensible way of handling it. I'd document both for a while, specifying a preference, introduce the changes to validators early and people can change as they find things, and you can remove the duplicate handling and extra docs when the number as reduced to a "low" level. Dave _______________________________________________ talk mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk
