2009/3/11 Tal <tal....@gmail.com>: > On Wed, Mar 11, 2009 at 4:53 PM, Dave Stubbs <osm.l...@randomjunk.co.uk> > wrote: >> Don't encourage people to try to insert random crap in free form tags >> such as name. > > It is no more crap than your answer ;) It is an attempt to offer a > better solution to a serious problem in some parts of the map. > I agree that this might be considered as a temporary solution since, > as I see it, at the end there will be no "name" tag. It is simply not > flexible enough for a multi language project. For that, a lot of tools > will have to change, or a common preprocessor agreed upon. > Until then, this is the lesser evil. > >> Most people dealing with this will just print it and >> it'll look stupid. > > You may be right. That is a disadvantage. But as we don't tag for the > renderer, we definitely do not tag for "Most people dealing with > this".
Yes. Yes we do. What we don't do is mis-tag for the renderer. We don't tag construction sites as landuse=allotments because we like the colour mapnik uses to render allotments. We do however tag things with highway=motorway and not highway=omgponies because that's what the renderer is expecting and only osmarender will correctly handle omgponies as a motorway of arbitraryness. > I know it might look like this a classic tagging for the renderer (I > already said that my problem is only with the slippy map on the osm > site), > but this is not the case. It is purely to have better data quality. > The best - and in fact the only reasonable - way to store Hebrew names > in a multi lingual map of Israel is using the name:he tag. > Currently we depend upon the slippy map. We tag, we see it rendered, > we fix problems, add missing street names, fight about naming > conventions etc. > The map is steadily improving (with the natural focus upon hebrew > names, but not only). The problems is that we are basing our all our > work on the name tags, which DO contain random crap. Sometimes they > also contain English, sometime Hebrew and transliteration, sometimes > just English, and most of the time Hebrew. A computer script can't > automatically go back from the "name" tags to the "name:he" tags. So > as a result, we have lesser quality data - We have no idea what the > name:he tags contain, because we do not see them, and the improvement > to the name tags can't automatically go back to the name:he tags. > This is the disturbing situation which I want to make a little better. > > >> Find some other way of doing this. >> Several perfectly good ones were suggested on the recent thread. > > I've mainly seen my proposal, copy-paste in the editor and name:local=XX. > Maybe I've missed something? > Are you referring to the name:local=XX ? This is OK by me, but less > intuitive to the mapper, since one needs to remember a new tag. And, > old tools will have NO name tags with this solution, so it'll be just > a little better then using name tags with garbage. Think driving > instructions with no name tags... > and you think name=$(name:he) is intuitive to a non-programmer? For an intuitive end-user-experience you need to fix the editor itself. And when you start playing with the editor you can make the copy-paste automatic. Dave _______________________________________________ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
