On 31/07/2010, at 11:07 AM, John Smith wrote: > Maybe we should stop using words for key pair values and just come up > with a database that issues ID numbers, half the problems with the > current scheme is due to people treating enumerated key pairs in the > same way they are used to using english, the other half is english > isn't used consistently.
Which won't solve the actual underlying problem, that there are a lot of similar but not identical things in the world. How will people know what numerical tag id to use? It will be by reading a description of it, which will have exactly the same language/cultural problems as textual tags do now. Lets say you used tag 192854 to represent a place which sells coffee and some food (i.e. what in some places gets called a café). I want to tag something that sells coffee but not food. Should I reuse 192864, even though it is not identical? If I do, then how do you distinguish between ones that sell food and ones that don't? If I make up my own tag, how do mappers choose between the hundreds of very closely related tag choices, and how do things interpreting the data know that they are all places which sell coffee? The only advantage I can really see is that you could have a database holding metadata about tags, so your 192854 and my tag could both be tagged with "items_sold=coffee". That doesn't really need magic database id to do that, someone could set up a db doing that with textual tag. For example "junction=roundabout" could be tagged with implies="oneway=yes". -- James _______________________________________________ Tagging mailing list Tagging@openstreetmap.org http://lists.openstreetmap.org/listinfo/tagging
