To paraphrase a well-known saying: Quality is in the eyes of the consumer.
How long do you think we can survive with this policy of refusing to acknowledge that there is such a thing as good data and bad data? Interpretation of the definition of the name tag (and many others) is incredibly subjective. Abbreviations? Ampersands? Including or excluding suffix type information such as "and Sons Ltd"? Capitalisation? Spacing? The list is endless.... As an architect I can only applaud Matthijs' initiative towards some level of normalisation. If there are objections to this particular method, let's have some alternatives out on the table. The fact that this discussion is taking so much time and energy would seem to confirm that we have so far failed to address the data quality issue, which starts with a definition of a norm - even if it is a woolly one if that's what is called for. Once we have a norm, we measure our data against that norm, and take appropriate steps to improve it. If the norm is "anything goes", then there should be no discussion, right? Colin On 2014-11-01 09:15, Lester Caine wrote: > On 01/11/14 01:45, Matthijs Melissen wrote: > On the whole the changes seem logical, but I'd rather they were applied > manually than setting a precedent to allow changes that may not ACTUALLY > apply as some small outlet is being a little economic with the name they are > using ... > I would guess that's a rather hypothetical situation. Do you have any example? More a case that any attempt to do the job properly' and create proper relational data by using any table of 'valid' data and only allowing a reference to that table to be used has always been frowned upon. If we are going to 'dictate' what is acceptable content it should be managed properly rather than blindly re-writing tags. I'm not saying that the changes are wrong ... just that this is not the right method to get things done.
_______________________________________________ Talk-GB mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-gb
