2010/1/4 Martin Koppenhoefer <span dir="ltr"><<a href="mailto:dieterdreist@gmail.com" target="_blank">dieterdreist@gmail.com</a>></span><br><div class="gmail_quote"><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
what does this conclusion imply? Does it mean if you encounter a separately mapped cycleway and there is not (yet?) enough "significant" different tags for the cycleway and the street nearby (say name, ref, maxspeed, width, surface, lanes are the same, the same turning restrictions apply, ...), you will delete the way in favour of a simple cycleway=track? IMHO you will nearly always find differences in detail, and we might be discussing about which differences meet your significance criteria, but I'd prefer not to...<br>
</blockquote><div><br>Just rewind a bit; in the start of this thread, I cited a guideline developed on talk-dk to aid the choice between separate and tagged cycle-tracks/lanes on a street with the aim of getting the best map description in the end. In the following discussions, I have tried to outline why this guideline did not end up saying: "always micromap everything".<br>
As for all other mapping, some judgement still needs to be left to the actual mapper, such as what is "significant".<br><br>My conclusion simply implies that I do believe the community would benefit from such a guideline.<br>
<br>Such a guideline can not tell people when to delete/convert/alter existing data
- just describe a "best practice" on how to map for the best end result.<br>I would certainly think twice and carefully before I converted an existing separate cycleway to cycleway=track/lane, just as I trust that you would always think twice and carefully before you fragmented a properly tagged cycleway=track/lane into separate ways.<br>
<br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">I'm not saying that it is impossible to enter all the details using just one way (as long as they stay strictly parallel or in case you don't care for positional / form detail), but there's a long way to go (currently there are no solutions or </blockquote>
<div><br>Well, I already admitted that it is not possible to represent all of the properties of the real world in the OSM description. Neither with extreme micromapping, macro/metamapping or combinations. But I have tried to show that micromapping streets into fragments also leads to loss of information, and why I think that in some cases this loss is more severe than with metamapping.<br>
</div></div><br>Our map will always be "wrong" in some respects. Live with it, but be aware of it. Only then will you be able to evaluate implementations properly.<br><br>-- <br>-- <br>Civilingeniør ph.d. Claus Hindsgaul<br>
Edvard Thomsens Vej 19, 5. th<br>DK-2300 KBH S<br>