David Earl wrote: >So what you're saying is that > >- each editor and data consumer has to have its own set of national >rules and defaults rather than defining them centrally (so inevitably >they'll end up different);
The editors must have some way to set defaults, the consumers will get a full dataset. So they must know the defaults plus the interpretation of the tagger(!) *now* but not later. >- we have to massively increase the amount of data we store by saying >for every road that it is open 24 hours a day (because some aren't) and >has a 44 tonne weight limit (or whatever it is by default in your >country) except for the few cases where it isn't; all cycleways don't >permit llama pack animals (because some in Peru do) and all motorways >explicitly do or don't permit horse drawn vehicles. The most common values (by highest count) can be left out from the *db* and only be stored once. So yes, there must db-wide-defaults. >- we can't type a simple tag any more, we have to go via a menu or a >form because there are so many of them. Every highway would have to >carry maybe thirty or forty tags giving use cases, Shure you can tag "cycleway" and nothing else, but you'll have tell the editor once, what a cycleway means to you. >and every time we >realise we are missing a use case (say we discover motorways in Ecuador >permit learner drivers to use them [please don't tell me this isn't the >case - it's only an example]) we have to add tags to every other highway >in the world to say that there learner drivers can't, otherwise we're >assuming a default. If you'll need to update any record in the table for this is a question of design. >- and that we have to update almost every way in the system already why? >and change every bit of software we already have why? Norbert "playing advocatus diaboli" _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk