|
In Serbia, there are still many regions with very little OSM data. In 2012, the Slovenian OSM community obtained USGS map data in Public Domain and put it on WMS servers for mapping purposes. Some users in the Serbian OSM community disapprove of its usage though, citing that it is better not to have maps than maps that were not checked with GPS and highway classificiations that were not reviewed on the ground, as users might get into trouble following these not fully reliable roads and paths. Generally, IMO it is better to have a not 100% reliable map than no map at all; my solution would be to reliably mark the source of the entered ways and nodes as "USGS maps" and thus to leave them as they are until they are revised with GPS nodes from local mappers. Would you agree? |
|
In my opinion it all depends on the quality of the data. If the roads have an offset of a few meters or don't reflect the changes from the past year then this is acceptable for the start I would say. However if the classification is plain wrong (i.e. highway=residential instead of highway=track) or not present at all or if the data is several years old or the offset very large or the data contains road that aren't present at all then by all means throw it away and go surveying yourself. Of course it takes some time and effort to review the data and get an idea about it's quality. And more importantly: You should discuss import plans with the local community, help.openstreetmap.org is the wrong place. 3
Well I'm not discussing peculiar import plans here, though bringing up an example, but posing a question that could be of general relevance for mappers. The discussion you're advising has already been led, this is just an effort to see what the opinions are in the general community. The answer might prove to be useful for subsequent newbies. But yes, "it depends" will as always be the answer to go with (like in the first answer), and where to draw the line will always remain the point of discussion, and the discussion should always primarily be led by those most affected by it, true.
(16 Sep '15, 14:45)
cirko
|
|
"Would you agree?" No. It is not that we don't have experience with importing junk, we actually have too much experience with it. Simply filling the map with evertyhing you can lay hands on makes it difficult to find out what is correct, what not, if further work is necessary and so on... besides the issue you mentioned of including unreviewed data from a quality pov. And in general correcting mistakes is more work than doing it right first time. Now the maps in question may or may not be of suitable quality and if available on compatible licence terms may might make complete sense as a refeference and mapping support. It should however be noted that statements by US Military organisations on the copyright status of their material should be viewed critically, very often the original source material has been obtained in military conflict and may still be protected in the country of origin. Your pov is clear and stands for quality, and let's leave aside the question whether this particular example is licensed adequately or not, but how can someone know when he sees a "blank map" whether there will ever be a mapper having more knowledge than him about this certain piece of land? See armchair mapping: I am perfectly aware that entering imperfect maps means more work than doing it well from scratch, but how can you know whethere there 'is' anybody who can and will do it from scratch? Isn't the point of OSM "contribute now on the maximum available level with all you have"? I am absolutely not talking about mapping areas where "real" OSM editing based on on-the-ground knowledge clearly is already happening.
(16 Sep '15, 12:21)
cirko
2
I don't believe we have a principle of "contribute now on the maximum available level with all you have", if at all it might be "contribute what interests you" (within reason), and as a corollary if nobody is interested well then the map is blank. Now naturally the whole issue is not black or white. IMHO having the major roads (motorways, perhaps down to secondary level or equivalent) in OSM is so useful that arm chairing those at quite large distances is quite OK, however with increasing detail the nearer you should be and the more important actual on the ground surveying becomes. I'm not even touching on the community building aspect of actually having room to add simple things here.
(16 Sep '15, 12:45)
SimonPoole ♦
Well, "it depends" is an answer that is in almost 100% of cases correct, and so it is here. :) Of course everyone should just do what interests him, OSM is voluntary, but when you're saying "within reason", that's the point where opinions of different people as always diverge. Me personally, I'm having trouble looking at blank maps from regions I know but cannot travel to right now, and looking at these very nice 1:50000 PD maps I took a shot at mapping rivers and roads down to unclassifieds in one particular region I originate from, unfortunately at almost the same time a more professional local user started mapping and thus we got tangled up a bit. Not mapping there anymore, of course, but I still just wanted to know what the opinions are on the topic, though it's clear that there will never be one solution for all.
(16 Sep '15, 13:02)
cirko
1
@cirko: this is a Q&A site not the place for long discussions from your POV.
(16 Sep '15, 15:07)
SK53 ♦
I'm sorry if you got me wrong, never meant to go into discussions about my peculiar example. The comments just reflect that the line between black and white is hard to draw.
(16 Sep '15, 15:16)
cirko
3
We're not talking about just importing junk here I think, but about using an extra data source on top of aerial imagery and public gps tracks. I don't see the harm in that.
(16 Sep '15, 15:31)
joost schouppe
1
@joost the question was far more general than just the specific example, thats why he got an answer to both.
(16 Sep '15, 15:52)
SimonPoole ♦
showing 5 of 7
show 2 more comments
|
|
Go for it, copy as much as you can from maps that are out of copyright (OOC) or where the licence is compatible We make great use of old maps for various features in Ireland, see here Very shortly we'll be making use of maps from the very same source for #MapLesotho About the only caveat would be to not use them for classifying roads 1
The OP was not referring to "old maps" in his example, but to a map from a source that is known to not respect non-US IPR.
(16 Sep '15, 22:46)
SimonPoole ♦
Sorry, said old in reference thinking of the maps we use in Ireland. Edited to clarify what I should have said
(16 Sep '15, 23:06)
DaCor
1
While it makes sense in some cases, I certainly wouldn't "copy as much as you can from maps that are out of copyright (OOC)". In the case of features that haven't changed (much) since the last century, such as Irish townlands, it does make sense. However in many cases there are problems. One is that the accuracy of old maps often isn't very good. In GB we we've had a few people drawing old railway lines and footpaths from OS "New Popular Edition" and other OOC sources. Near where I live the former path of railways is often visible, and the equivalent features on the local NPE map are sometimes only within 50m or so (different features have different offsets so I suspect some were added laterm perhaps by someone on a Friday before they finished work for the weekend). Another problem is that some things just don't exist in the same form any more. Railways are an obvious example again here, but another one is footpaths. As an example, at the time of writing, there are a few at http://www.openstreetmap.org/#map=14/53.2628/-0.8339 (I was there yesterday). The footpaths that mappers have copied from NPE sometimes (but not always) correspond to a current footpath, but often not very closely. In the specific case of the footpaths that I surveyed yesterday, it was more difficult to do so with invalid OOC data present than it would have been without it. In that particular case, "not mapping" would have been better than copying something that's old and invalid, but as has been said already "it depends".
(21 Sep '15, 14:12)
SomeoneElse ♦
|
