On Thu, Feb 17, 2011 at 8:57 PM, John Henderson <[email protected]> wrote: > The downside I can see is the difficulty in rendering software being able to > make use of the information given in the exact text. The same information > may be expressed quite differently in different locations.
I think there's a kind of continuum of interpretation, starting from the hard, cold fact (the sign) through to what the human being at the other end will eventually do with the information as presented to him. Documenting the hard cold fact is certainly useful, and allows for very precise analysis of the data. But it also makes the data virtually unusable without further work: can you imagine writing code that has to deal with the raw text of signs. So we do some interpretation of the facts to make processing easier: assimilating a range of different signs that might indicate that caravans shouldn't go down a road, and in the process, losing some granularity. (Obviously I'm talking more generally than just caravans...) So...recording both bits of information seems like a good idea: caravan=no signposted:caravan="Road unsuitable for caravans" Or something. Btw, if we're distinguishing between "bad idea" and "prohibited", perhaps the tagging should be "no"/"prohibited" rather than "unsuitable"/"no". It's pretty unusual that you'd want to disregard an "unsuitable", but obey a "prohibited". Steve _______________________________________________ Talk-au mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-au
