Re: [OSM-legal-talk] Questions about CTs 1.2.4
On 14/04/2011, at 8:06 AM, Francis Davey wrote: > On 13 April 2011 22:24, James Livingston wrote: >> * If so, how do we know what data must be removed in a switch to ODbL? > > That clause doesn't appear to put any obligation on you to remove > data. All it requires of you is that _when you contribute_ you have a > right to give that authorisation. Okay. So those maps that people are producing showing how much data would need to be removed if we changed to ODbL are a work of fiction then, since that are based off who has agreed to the CTs, which don't guarantee the data is re-licensable. Hmm. -- James ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] Questions about CTs 1.2.4
On 14/04/2011, at 6:57 PM, Robert Whittaker (OSM) wrote: > This method seems a much more satisfactory way of doing things to me > -- assuming it could work legally (IANAL). We would still have the > flexibility to re-license if we needed to without individual mappers > being able to hold their data hostage, and we have much clearer rules > for contributors to follow about what sources it is acceptable to use. I think the big problem with that would be people editing, and combining licences. Say for example there is some CC-BY derived coastline, and a road traced from Bing. I know that the area in between is a sandy beach, so add that to OSM. What licence is my beach under, and what are the restrictions for changing it in the future? Right now this works because everything is CC-BY-SA or compatible with it, so it's not really an issue. I honestly don't know how it'd be figured out for a change to ODbL, probably just ignoring the issue, but if we're tracking what licence all the data is, then it would seem like something people would want us to figure out. ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] A plea for meaning ful changeset comments
On 30/07/2010, at 9:52 PM, Steve Bennett wrote: > For me, very frequently, the changeset just represents a random bunch > of edits I happened to be doing at one time, with not much cohesion. > There are different suburbs all in the same changeset as I flitted > about. My editing falls into two categories, casual editing and big tasks. I think I put in reasonable comments for big tasks (my current one is uploading National Parks data). For casual editing, I'm not sure what I could put in that would be useful. Often I start off adding some street numbers I've collected, and then trace those houses from nearmap, and then start tracing a creek, and then start doing something when that ends. When I set the changeset comment, I don't know exactly what I'll be fixing up - I know the location, but you can get that from the changeset anyway without any comment. For any kind of semi-automatic or large scale things, I agree that good changeset comments shouldn't be difficult to write and would be very useful, but I'm not sure about small-scale editing when you go along with things. -- James ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Defining critical mass...
On 14/07/2010, at 9:52 PM, John Smith wrote: > On 14 July 2010 20:59, Richard Weait wrote: >> What do you suggest would be acceptable / unacceptable? > > I would consider things to fail if more than 5-10% of data disappears > in any region. At the very least it would be demoralising for anyone > that spent even a few hours working to make OSM data better. Every keep talking about 5% of the data disappearing, but being kept as-is and being remove aren't the only two possibilities. Being removed is only necessary is the person who first created it refuses. If the object has say 6 version and mapper 4 refuses, it can be reverted to version 3. If mapper 5 says yes and just added the street name, you should in theory be allowed to re-add that to the version 3 data. How all that will work in practice, I don't know. However part of it will still need to be dealt with, if nodes get removed but a way they are in doesn't, or things that are part of a relation. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] What could we do to make this licences discussion more inclusive?
On 14/07/2010, at 10:28 PM, Ævar Arnfjörð Bjarmason wrote: > I'm no expert on this sort of thing, but there are probably a lot of > well known pitfalls to avoid when trying to run an inclusive > international project in many languages. I'd think having English-only > discussion at a set time convenient for Europeans would be pretty high > on that list. I don't know if you'll get out of being English-only, since like it or not it is the main working language of OSM (as with many open projects on the Internet). Using any other language is probably going to exclude even more people. One thing that I've seen done in other projects is rotate between three meeting times eight hours apart. So for example one meeting would be 1800 UTC, the next 0200 UTC and the next 1000 UTC. >> Further to what Frederik has said, there's a couple more points that >> are important. The OSMF receives legal advice on matters relating to >> the license change, and as far as I'm aware they are forbidden from >> making the legal advice public. I can't speak for them, but I would guess it's more inadvisable than forbidden (with respect to licensing anyway). If you get advice saying "we believe that sections A, B and C will hold up in court, section D probably would, E should unless XYZ happens and we don't know about F", then telling everyone that means anyone trying to get around it knows about the potential holes you found. Of course, people using the license will want to know about that kind of thing, so it's a trade-off. > I.e. can the legal advice only be shared among people actually on the > LWG conference call, and not all OSMF members? Who can be on the call - LWG members, any OSMF member, or anyone involved in the project? Actually, I can't even find how you get on the LWG in the first place. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-legal-talk] public transport routing and OSM-ODbL
On 09/07/2010, at 12:24 AM, Matt Amos wrote: > I agree with Andy. This is what I understand the ODbL to be saying. > Unfortunately, as with any legal text, its difficult to read and this is an > unavoidable consequence of the legal system. If you need interpretation of > the license, new or old, the best route may be to consult a lawyer. Consult a lawyer with the caveat that what they tell you may only apply to the one jurisdiction. While they can tell you whether you're likely to be able to do or not do[0] something, it's harder for them to tell you whether other people can/can't do something in a different legal environment. [0] as far as lawyers can advise about things, without being tested in court it's just an (obviously educated) estimate of your chances. ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] new license use case questions
On 09/07/2010, at 1:07 AM, David Carmean wrote: > They use a shapefile generated from > a filtered snapshot of OSM data--leaving only roads--as a base layer. > > If they do nothing else but serve this one-time snapshot as a base > layer, what are their obligations? My opinion is that since they've changed the data (stripping out the non-road bits) that creates a derived database, which strictly speaking has to be provided. However like with similar situations with the GPL, this may not be as onerous as it sounds - many people package binaries of GPL'd software without putting the source along side it. Per ODBl 4.6b, you would just need to tell anyone asking the filtering algorithm you used - for example "remove all ways that don't have a highway=* tag" (in machine readable form). > Secondly, what if one of their staff, being unfamiliar with OSM but a > GIS expert, sees a problem with one or more roads about which they > have personal knowledge, and "fixes" those problems in the GIS data > only--then publishing the result as a mapserver layer only. What are > their obligations in this case? Making either the derived database or the changes available. Since it's not algorithmic, you have to provide a diff (or the whole derived db) if someone asked for it. If you are doing a one-off import, you'd just need to either have a copy of the original to produce a diff upon request, or be willing to produce a dump of the OSM-based part of the database. If they're not doing a one-off import of OSM data, you'd presumably have some way of merging you local changes with the upstream changes, so a diff should be easy. > Third: there is the usual problem/condition (depending on your > political leanings) of divergence in the tag values. For example, > "surface=Dirt" vs. "surface=dirt", "surface=cement" vs. "surface=concrete", > etc. I'd certainly want to "fix" that, but if I were that agency, I woulnd't > have the time/skills to make a 'bot fix back in the OSM database. Strictly speaking you'd have to provide those changes. But if it's a rendering rule then it's not a DB change, and if it's automated you can just provide the algorithm "change surface=Dirt to surface=dirt". The two things to note are, 1) like with the GPL, you have to offer it to the recipients, however if it's not "interesting" (e.g. just filtering), no-one will probably ask, and 2) if it's an algorithmic change, you just have to provide the algorithm. -- James ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] Calling all bulk importers
On 17/06/2010, at 2:21 AM, Mike Collinson wrote: > If you have been involved in bulk import of data from third-parties, may I > ask you to check that this is on > http://wiki.openstreetmap.org/wiki/Import/Catalogue . > > Why? Now we have final versions of everything, the License Working Group is > checking compatibility with the proposed change to the Open Database License. > We are aware that in some cases the donor's permission will need to asked. We > like to leave you as much time as possible to do that and to be prepared to > assist you if needed. There is a new support page here. I'm going to update the page with some of the data imports that have been done in Australia, most of the ones from government sources being CC-BY (not -SA). As far as I'm aware the conditions of CC-BY should be met by being ODbL, so the licensing wouldn't be a problem, however I'm wondering about the contributor-terms bit. If I recall correctly, there was discussion that bulk imports could be exempted from the contributor terms - for example AND isn't going to let us arbitrarily re-license their data, so we'd have to exempt that if we wanted to keep it. We're trading off the usefulness of the data for being beholden to the company if we want to re-license in the future. How does a decision about whether the tradeoff is worth it, and hence gets exempted from the contributor terms, get made? Presumably AND data would get one, but my personal edits wouldn't be worth the tradeoff. We're getting more Australian government data coming along, so what should we be doing to either know that it can be exempted or that we shouldn't import it? -- James "Doc" Livingston ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Changed highway=*_link meaning?!
On 23/06/2010, at 8:56 PM, Andy Allan wrote: > Don't worry, it hasn't actually changed the meaning of anything - it's > just that the wiki is now wrong. The easy way to fix the situation is > to correct the wiki - it's as straightforward as that. You could argue the wiki is now wrong, but you could argue the wiki is now right too. Just because the wiki previously said "X or lower" previously doesn't mean it was correct. I, and from what I see in use where I live quite a few other too, have always used xxx_link tags to join a highway=xxx with a higher one, because we think what was documented on the wiki (xxx_link joins highway=xxx with a lower one) is silly. > Many of us refer to this kind of activity as "wikifiddling", or the > counter-productive deliberate insertion of false statements onto the > wiki in an attempt to influence the real world. You could argue that it's wikifiddling in an attempt to influence how people map, or that it's documenting how a lot of people already map. It's all a matter of perspective. Short of a tagging dictator, how do we decide which camp wins the argument? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Bug reporting problems (was: Big sponsors)
On 18/06/2010, at 4:56 PM, Ben Last wrote: > The existing editors (and I include Potlatch, JOSM and MapZen in that) are > powerful and well suited to users who understand mapping and OSM and are > motived to deal with the UI complexity, but are not at all well suited to > generalist users who want only to make small, simple edits. Providing > simple, targeted UI support for making such edits I'd been thinking about this for a while, and the talk-au discussion had reminded be about it, but I don't think I replied there. What we really need is a task-based editor, where you tell it what you want to do, and then it help do you do just that one thing. Click "contribute to OSM", then get presented with a few choices: 1) Name streets 2) Add house number 3) Draw streets from imagery 4) Add traffic sign stuff (turn restrictions, give way, etc) 5) ... When you click Name Streets, you see the map and can click on a street then enter it's name - and nothing else. Want to add house numbers too? Go change modes. For number (3) it could offer to switch to (1) when you'd finished, in case you know what they're called too. As well as being simpler, it will stop people (newbies and experienced mappers alike) from accidentally breaking other things. I'd use it when uploading house numbers I've collected, since there would be no chance of me accidentally pressing a key and wondering what I just did. -- James ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [talk-au] Fwd: Re: Why PD is not better for business
On 11/12/2009, at 8:02 PM, Elizabeth Dodd wrote: > so we don't need imported data? In most cases we don't need imported data, but it can be useful. For example rather than painstakingly crafting the entire coastline of Australia from a few GPS traces and a lot of imagery (much is relatively inaccessible), we can import it from someone's dataset and spend that large amount of time doing other things to improve OSM that we can't import data for. There are also some things where importing external datasets is the *only* way to get it into OSM. For example boundaries of areas that have no physical edge, just a (not necessarily straight) line on someone decided on at some point in time. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Question that will not be asked
On 09/12/2009, at 7:29 PM, Pieren wrote: > The poll contains the three original questions which will be asked to > all contributors if the new license is approved by the OSMF: > - yes > - yes and consider all my data PD > - no > Then I added two more questions for those who would say 'no', for this > part of the public who feels that they have a 'gun on the head' > because they don't agree with the Odbl. Then I added two options to > give them a chance to explain why they would say 'no'. Going back to the original point of this thread, that is why it really needs to be a simple Yes or No question. Once we start having extra options, we need to have a debate about what all the options are and exactly what they mean. For example, that poll has nothing for "I want attribution (which PD/CC0 doesn't cover) but not necessarily share-alike" (like CC-BY) or "ODbL is fine, but I don't like the contributor terms", and I'm sure people can think of other things which fit their exact preference. You also need to sort out what voting system you want to use, and deal with people trying to vote and not being sure what the difference between option 7 and option 13 is. A simple Yes/No vote is the best option, because people either like what is being present or they don't. It's not the best for getting a feeling of what the community wants, but it's much better for an actual vote on changing. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Announce] OSMF license change vote has started
On 06/12/2009, at 10:05 AM, Frederik Ramm wrote: > James Livingston wrote: >> For example, I have inferred road positions from the CC-BY-licensed >> Queensland DCDB-lite dataset, and have uploaded national park and >> world-heritage areas from the CC-BY dataset on data.australia.gov.au. >> As I'm not the copyright holder of those base datasets, I don't see >> how I could agree to the relicensing, or contributor terms which >> allow for future relicensing. Does that mean everything I've ever >> contributed (even my own work) has to be deleted? Probably. > > Are we sure that CC-BY is any more incompatible with ODbL that what we're > doing now? I mean, nominally we have CC-BY-SA but data.australia.gov.au is > not listed on the maps anywhere... data.australia.gov.au is just a repository for data, the actual copyright holders are various departments in various levels of government. For example various ways have source=au.qld.dcdb_lite attribution=Based on data from State of Queensland (Department of Environment and Resource Management) 2009 or source=ABS_2006 attribution=Based on Australian Bureau of Statistics data and the ABS is mentioned on the Attribution wiki page. Aside from any incompatibility between CC-BY and ODbL, the contributor terms would prevent us using CC-BY data in OSM unless the copyright holder agrees to the terms. I think that by satisfying the ODbL you also satisfy all the conditions of CC-BY, so it would be distributable under the terms of ODbL - but strictly speaking I don't think you could say it's all ODbL-licensed. > If we're talking about a lot of data, and if you have put proper "source" > tags in or tagged your changesets in a way that makes them discernible, then > we can find a way to open a new account and transfer this "tainted" data to > the new account and you then accept the relicensing with your old account. As above, I think everyone has been putting those tags (and similar ones for other datasets) on the ways, so hopefully they'd be extractable. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [Announce] OSMF license change vote has started
On 06/12/2009, at 8:44 AM, Ulf Lamping wrote: > Tom Hughes schrieb: >> Polling the OSMF members is just the first stage - there will another >> vote later when all contributors will be asked whether they want to >> relicense. > > With a gun at their head: "Refuse: After the migration (currently 26th > February 2010), your contributions will not be included in ODbL licensed > downloads and you will not be able to continue contributing.". > > If you call this a vote, then we have pretty different understanding > about voting. I'd say it isn't a vote, it's asking whether you agree to relicense your contributions under the ODbL subject to the Contributor Terms. I would imagine there are quite a few people who couldn't legitimately agree to that, even if they want the ODbL. Have you ever important any CC-BY(-SA) data using your account? If so, they I would think that you can't legally agree because you are not the copyright holder of all your contributions. For example, I have inferred road positions from the CC-BY-licensed Queensland DCDB-lite dataset, and have uploaded national park and world-heritage areas from the CC-BY dataset on data.australia.gov.au. As I'm not the copyright holder of those base datasets, I don't see how I could agree to the relicensing, or contributor terms which allow for future relicensing. Does that mean everything I've ever contributed (even my own work) has to be deleted? Probably. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-legal-talk] [Osmf-talk] New license proposal status II
On 03/12/2009, at 10:19 PM, Ed Avis wrote: > That was my interpretation too. It appears to me that if some well-meaning > body released a set of data under the ODbL (which presumably we recommend as > an appropriate licence for geodata) then the OSM project would not be able to > use it. In other words, under the proposed way of using it (with these > contributor terms), the ODbL is not compatible with itself. A somewhat similar situation happens with open-source software that is licensed "GPL 2 or later" - people can license their changes under "GPL 2 only", meaning they can't practically be used upstream. The difference is that doing so is an active thing, you are not likely to release your changes like that without knowing that you're doing it. With the contributor terms however, you have to actively choose to make your data importable, rather than it being importable by default. > The 'without giving up the ability for easy re-licensing' part is not a > disadvantage of the ODbL or the proposed contributor terms; it applies to any > licence. (Currently, data released under CC-BY-SA can be imported into OSM, > but the project doesn't have the right to relicense it without separate > permission.) I probably could have left out the references to the ODbL in my mail. > However, if the policy is that no data (ODbL or otherwise) can be imported > without agreement to contributor terms that allow broad relicensing, then in > practice data derived from the OSM data cannot be merged back in without > special permission. This does seem to defeat most of the point of share- > alike licensing. (The data set may be available, but without permission to > reincorporate it into OSM, it becomes much less useful.) This is the main point of what I was getting at. We'll have to see what the LWG thinks, but as I read it the proposed contributor terms defeat the main point of choosing a share-alike license: that we can benefit from when a derived database contains some useful information. ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-legal-talk] [Osmf-talk] New license proposal status II
On 03/12/2009, at 10:19 PM, Mike Collinson wrote: > - Whether friendly or unfriendly, they never have any obligation to merge in > their data improvements into our database. > - However, you or I can. > > Does that make sense? I completely agree that they don't have to do anything towards merging any data, I meant "unfriendly" as not actively working towards improving the upstream data in the main OSM DB (where sensible), not in any way hostile. > The key will be clause 1 of the Contributor Terms[1] allowing you or I to do. > Is the ODbL license itself explicit permission from them? I'll bring it up > for review at an License Working Group in the light of your question. I can't see how the ODbL would give a third party (e.g. you or I) permission to agree to that, as it would mean we could relicense data without the copyright holders approval, but I'd be interested to know what they think. ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] Instead of voting
On 11/10/2009, at 12:08 AM, Martin Koppenhoefer wrote: > This proposal includes the deletion of all voting-related stuff > including the casted votes of the past. I'd say that this helps prove the point that different people reading different things into what pages on the wiki say. The proposal wasn't really serious, but I didn't intend it to mean that we would delete all voting-related content from the wiki, only ban new votes. > the thing is that not everybody will write a documentation for every > key he uses, and in the end (we're already in some tags at this > stage), there will be many same tags with different intended meanings. > By deleting the voting-process things will get worse. I'm not entirely sure how not having the voting process will make things worse. Instead of having a tag with several different meanings, one of which is "approved" but may not even be the most common meaning, we'll just have a tag with several different meanings. The voting process doesn't mean that the "approved" version is what people actually use. If there is a wiki page which describes a tag in a limited way, and I want to document how I've used it, what should I be doing? * edit the main page, which could annoy the people who created the page * add a note to the discussion page, which someone searching the wiki for how to tag things won't read * create a new page describing my version, which leads to conflicting information ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-legal-talk] ODbL "virality" questions
On 06/10/2009, at 11:30 PM, Matt Amos wrote: > so far, all the responses seem to indicate that everyone thinks > linking to OSM data by ID is OK. what about Andy's idea, though? is it > OK to take a location, name and possibly an ID as well to perform > "fuzzy" linking? > > my view is that all the linked-to OSM information would have to be > released; the list of (location, name, ID) tuples. but that it would > still be OK to not release the linked-by proprietary information. That sounds good in theory, but I think at some point getting out the locations and names of things could be "Extraction and Re-utilisation of the a Substantial part of the Contents". Am I allowed to mine the database for the name and location of all the pubs and restaurants in the world, without having the data fall under the ODbL? If not, how could it become okay if I claim to just be using them as lookup keys? I guess you could have a database with all of your proprietary data, and second one which acts as a link between the fuzzy-OSM data and IDs in your database, and only release the second. ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] Tagging schema
On 06/10/2009, at 10:58 PM, David Earl wrote: > On 06/10/2009 13:35, James Livingston wrote: >> I can see things getting ickier than they are now if you can just >> go around adding new shop= values, without having some prior >> discussion to what it means. If I saw a suggested option in an >> editor, I would generally assume that there is some agreement as >> to what it is supposed to mean. > > You can already add new shop values willy nilly with no discussion, > and lots of people do (and value this capability and would be loathe > to give it up). Sure, and I uses that all the time. I was just trying to say that I think a lot of people (myself included) would tend to assume that suggestions being offered by an editor had at least some vaguely consistent meaning, which a lot of the shop tags don't. > The action to add a new tag/value ought still to be simple in an > editor, so you're not held up for lack of anyone adding shop=joke > previously, but it should at least ask you to describe what you mean > so you can spread the word and at least minimally document your tag. > (Of course, you could code an editor to bypass all of this, but that > would be rather unhelpful to everyone else). I think that being able to document your tag would be useful, even (or especially) when someone else has already done so. It would allow us to collect more data about how people are actually using the tags, and might help to find things that need to be ironed out. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Landuse areas etc. abutting highways
On 05/10/2009, at 8:18 PM, Marc Schütz wrote: > IMO (a) is the correct way to do this. ... > For a road, we can either choose to map it as a linear object (this > is the common case), or we can map its geometry more exactly by > using an area. In both cases, however, the object in our database > represents the entire road (i.e. not only the middle line). Because > in reality, there is no gap between the road and the areas next to > it, there shouldn't be one in the database either. I agree with this, for things like landuse (which is what is mentioned in the topic) where the road is represented by a way. A residential area or farming area abuts the road reserve, so the polygon should abut the road's area or way. If you're actually mapping the road and not the road reserve, so putting things like footpaths as separate ways, you obviously wouldn't want to have the landuse cover those, but for just a single way it makes sense. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Tagging schema
On 05/10/2009, at 7:54 PM, David Earl wrote:
> * Three new primitives, tagkey for describing the k part of tags,
> tagvalue for the v part of tags and tagdescription separated off to
> allow for multiple descriptions in multiple languages without having
> to
> download all the data for languages you're not interested in.
> ("tagkey"
> etc can be anything we want, don't get too hung up on the
> terminology, I
> just use it for didactic purposes).
I'd been thinking something along these lines for a while, due to
having used a similar system at an old job. There everything in the
database was described my metadata in The Dictionary, and The
Dictionary lived in the database too. Essentially, we'd just allow the
tagging of tags (keys and values) in the same way that we tag nodes,
ways and relations.
I think being able to add arbitrary metadata to tags would be handy,
because you can come up with cool new things. For example, you could
tag the proposed incline=up tag with "edit:reverse_way=value_flip" or
similar, which says the value needs to be flipped if the way is
reversed. If an editor knew what to do for that tag it could do it
automatically, and if it didn't it could present a warning to the user.
Control would be an issue though, as someone accidentally or
maliciously breaking tag metadata could really screw things up, if
editor and renders went straight off it rather than "verified" copied.
> (d) the meaning of newly introduced or changed tags goes along with
> them, so that the intention is described to others.
I can see things getting ickier than they are now if you can just go
around adding new shop= values, without having some prior discussion
to what it means. If I saw a suggested option in an editor, I would
generally assume that there is some agreement as to what it is
supposed to mean.
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 4:29 PM, Gervase Markham wrote: > Because sometimes, occasionally, a benevolent dictator (a phrase > used by > lots of open source projects) has to break deadlock and dictate. > Things > are working well when that power is used very, very rarely, but it > needs > to exist. Mozilla has two - one code, one non-code, and I can't > remember > the last time they had to break a deadlock in this way. But it's vital > that they _could_. As well as a benevolent dictator, they also have a power structure underneath them, which gives the community guidance on how to work together to solve issues. The guiding force of the power structure is why the benevolent dictator doesn't have to use their power very often. As far as I'm aware, OSM doesn't have any kind of power structure - there is no equivalent to the contributor->committer->core developer->maintainer chain. As I mentioned in an earlier post, I think we need a Tagging Committee. Not to develop tags themselves, but to oversee working groups who develop tags. If we had one, they could create a "footpaths, cycleways, and tracks" working group, who would then sort out the highway=footway/cycleway/path mess. The WG would work together to sort it out, and the TC would then say "we think your WG has achieved a consensus, your tags are now approved". ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [tagging] Feature Proposal - Voting - boolean values
On 03/10/2009, at 5:54 PM, Konrad Skeri wrote: > Consensus will never happen and we don't have a dictator, which makes > voting the option left. I actually agree that we just need to pick one, and since "yes" seems to be the most commonly used one, that should be it. However, I just don't see how voting is going to help. > There are a lot more mappers, especially newbees, that care a lot more > what's on the wiki guidelines than what is being said in a > 50-messages-a-day-thead on a mailing list only saying "i said" "no you > said", and frankly isn't leading us anywhere. I'm not saying that looking things up on the wiki is stupid, only the voting. We obviously need a way of documenting tags, and the wiki can do a fine job of that. But how is having the exact same arguments on a wiki discussion page (or spread across several) any better than the mailing list? The problem isn't tied to a particular mechanism, it's a social problem where we currently don't have any form if power structure, and the one mechanism we have for choosing stuff (voting on the wiki) isn't accepted by a whole bunch of people, and doesn't get the buy-in it needs. I just don't think winning a wiki-vote 10-to-8 (or similar) is enough to convince people that the community actually approves a change. > http://wiki.openstreetmap.org/wiki/Proposed_features > But then again, you could argue that since it's on the wiki it's > stupid and noone cares about that either. People have argued before that require a majority of 15 votes is in no way enough for changing existing tags that have any significant use. Although the majority might be "yes", I would presume that "true" and "1" also have significant use, or we wouldn't be having the debate. > Or are you just opposing any idea that doesn't instantly answer the > question of life, > universe and everything. (Which is 42 by the way) I'm opposing the wiki-voting process, because I think that it's broken, and a sizable chunk of the community don't believe in it. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [tagging] Feature Proposal - Voting - boolean values
On 03/10/2009, at 5:02 PM, Konrad Skeri wrote: > Time to end this debate > > http://wiki.openstreetmap.org/wiki/Proposed_features/boolean_values Oh, and this: http://wiki.openstreetmap.org/wiki/Proposed_features/VotingOnTheWikiIsStupid ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [tagging] Feature Proposal - Voting - boolean values
On 03/10/2009, at 5:02 PM, Konrad Skeri wrote: > Time to end this debate > > http://wiki.openstreetmap.org/wiki/Proposed_features/boolean_values How precisely is that going to end the debate? a) Voting isn't the way to do this. It either needs consensus or a dictator. b) Lots of people don't care about some stupid vote on the wiki c) Who decided the rules for voting? d) How are you going to enforce the result, when lots of people will simply ignore it? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 5:02 PM, John Smith wrote: > This was not only highly frustrating but demoralising and as a result > I've not been bothered tagging any more school zones because I don't > see a point until there is a "One True Way" to tag school zones. Just do what I and a lot of other people have done - give up on the wiki being useful, and just go ahead and tag it however you like, checking tagwatch and similar to see what other people are actually using. If you do see an article on the wiki, you have no idea whether anyone actually uses it. If it's "approved" it was probably because three people voted on it and everyone else got sick of arguing and didn't vote. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 4:25 PM, Gervase Markham wrote: > Wikipedia has much less need for consistency than we do (e.g. it > doesn't > matter if one article is in American English and another in Australian > English; articles are not machine-parsed) and yet they have all > sorts of > mechanisms for ensuring it. I believe those mechanisms are something along the lines of "if it's related to a particular country, use their spelling, if not whoever creates the article first". Which of course leads to articles like Orange_(colour) and Grey using the British spellings, and most other colour articles using the American spelling. As long as it's consistent within the article, they don't seem to mind. I'm all for working groups, who have their own mailing list (reporting to the rest of us occasionally), and having regular meetings. Pick an arbitrary time in UTC and rotate between that, that+8 and that+16 for the meetings. I think having their own mailing lis is important - because anyone who just likes to argue, and doesn't actually care, hopefully won't bother subscribing. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 1:24 PM, Russ Nelson wrote: > I suggest instead that in cases such as these, SteveC should bless one > of them with his Holy Water of Antioch (and the number of the tags > shall be 3, no more and no less). His blessing will tip the stable > disconvergance in one direction. For cases where it's something like picking a value from "yes/true/1" or key from that list, sure. They all mean the same thing, and for the most part it's easy to change. I'd be less inclined to have one person, or the Ministry of Tagging, to use argument-by-authority is there was actual contention about what the tag actually means, as in some other discussions. Those ones really need consensus, or you'll get groups of people (or whole countries) ignoring the authoritative decision, and actually changing existing use is much more complicated. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 1:16 PM, Andrew Errington wrote: > If I am a map maker then I know whether or not the street has a name, > because I've been there and seen it. I can look at the map and see > that > this street has no name, but I know that it does. So I will edit > the data > to make it right. This also covers the case where the name is > wrong, or > misspelled. Provided you like somewhere that has all of the streets names, yes. If you live somewhere that has large swathes of unnamed (or completely unmapped) streets, then it certainly makes a difference. I could go drive somewhere to get that street name, but then find out it doesn't have one. Next week, someone else who lives nearby can go drive there to check the street name, and find there isn't one. If we can record the fact it hasn't got one, we could all spend more time doing areas that aren't mapped properly, rather than re-checking the same places again. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 12:53 PM, Jeremy Adams wrote: > If different regions want to use the map for different purposes, > display different tags, etc then they can apply their localization > when they create their map. It's not so much that there are different uses, but a lot of the assumptions and hidden implications in tags aren't the same. A lot of the tags are a bit Europe-centric because that's where OSM started, and where a lot of the mappers are. I'm not blaming anyone in Europe for this, it's just how it's developed. As such, people in other regions often take liberties with what the descriptions on the wiki, and historical consensus, to fit them to the local area. If we want to have globally consistent tags across the world, we're probably going to have to go and modify a bunch of core tags that are extensively used everywhere. Like whether highway=* is a physical or importance thing, and what it actually implies. As I understand it, it's used consistently within some countries as the former, and used consistently within other countries as the latter. Either we have it not being globally consistent (see the International Equivalence table), or half the world will need to change. > Otherwise there's no way for applications (routing and otherwise) to > know how to work across the whole globe. They're always going to have to know local quirks. From the discussion a while ago about highway=residential, I got the impression that in Europe is has a semi-implied access=destination for the purposes of routing - that is, you shouldn't drive down a residential one unless you have to, because they're very thin. In Australia, we have a lot of residential roads that are wide enough for four cars (one parked and one lane, either way, so driving down random residential streets isn't uncommon. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 02/10/2009, at 7:12 PM, Nigel Magnay wrote: > That's fine, so long as the tags themselves are namespaced. Otherwise, > just as now, the semantics get confused. > > I.E, It should be the case that if I tag as > > FredericRamm:interesting=true Going this route is really just reinventing XML, without the advantages of actually being XML (e.g. tooling, schemas). For example: versus: http://osm.org/users/FredericRamm";> true Not that I'm saying we should go that route, but if we want to start namespacing everything, we might as well do it properly. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] SteveC should decide
On 03/10/2009, at 7:02 AM, Gervase Markham wrote: > More examples from the Mozilla project: if one vocal group want > something one way, and another vocal group want something the other > way > in Firefox, the _worst_ thing you can do is make it a preference so > that > both sides can have what they want. That just makes everyone's life > more > difficult, because there are now two code paths to test and maintain. > Multiply this up by a number of decisions and you get complexity > explosion. That mostly works because you're talking about code, not paragraphs of description of what a tag means. If they're knowledgeable enough to figure it out, two people reading a chunk of code should come up the same idea of what it does, which doesn't happen with tag descriptions. If a tagging overlord who happens to be English writes a description of a tag, I can pretty much guarantee that some native English speakers from another country (e.g. Australia or the US) will read it a different way, or people who have English as a second language will rad it a different way. If we wanted to go the tagging-committee route, I think that voting people on to it is the wrong things to do. Geographical and cultural diversity is much more important than how many votes you get, otherwise you'll end up with a group that doesn't include large potions of the world. > If I were considering using OSM data in my business, I would > consider it > laughable that after 5 years there had not yet been a decision on what > value or small set of values I needed to look for on boolean > attributes > to see whether they were true or false. Laughable. As people pointed out the true/yes/1 things isn't really what we're arguing about, it's the principle of how we decide things. What does "forest" mean, or "residential", in a global sense that can be explained to everyone? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] address interpolation
On 28/09/2009, at 2:22 PM, Marcus Wolschon wrote: > 25A-25C should work with addr:interpolation=alphabetic . > However not all software that supports interpolation at all, > supports this interpolation-mode yet. > > 25-25A would not. I'm not sure you how you can interpolate things like this correctly if you're just using a single interpolation way. For example I've seen both 23-25-25a-27 and 23-25a-25-27, with the 'a' house being whichever one was built second or isn't the primary residence. Whenever I've encountered something like this, I've just broken the way up, so that there is one for ...-23, two plain nodes to 25 and 25a, and another way for 27-. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-legal-talk] New license status
On 28/09/2009, at 11:16 PM, Gustav Foseid wrote: > Well... There is no copyright that expires after 15 years. Sui > generis database rights expire after 15 years, but copyright is > hardly very relevant for an OpenStreetMap database dump. In Europe maybe - however there are countries where database do have inherent copyright separate from the copyright over their contents, for example in Australia. I think the copyright wouldn't expire for 70 years here, which is definitely more than the 15 for European sui generis database rights. I see the qualification that "substantial" is in terms of quality, quantity or a combination of both - but out of interest, is it supposed to mean basically what it means in terms of the underlying copyright/database rights? ___ legal-talk mailing list legal-t...@openstreetmap.org http://lists.openstreetmap.org/listinfo/legal-talk
Re: [OSM-talk] Should Bridges be independent of their ways?
On 22/09/2009, at 10:57 AM, Martin Koppenhoefer wrote: > 2009/9/22 Anthony : >> It is possible to represent different surfaces and different >> maxspeeds >> without using more than one way. "maxspeed:lane=130;110"; >> "surface:lane=asphalt;concrete". That's not necessarily the best >> solution, > > indeed, it won't be understood by none of the apps that are using our > data and it doesn't say, which lane has which value... The first bit is always going to be a problem, however we decide to map lanes the tools will need changing. Special tags or relations? They'll need to parse those to be useful. A new database construct? They'll need to understand that. There is no way around needing to modify the tools if we want more than a pile of tags that somehow relate to ways. The second bit we can easily get around by picking either counting from the left or right. Maybe something like lane[n]:key=value where there is data, and on the node in a way where the lanes change or merge we have lane-mapping data that says lane A back along the way changes to lane B after the way. > but that's exactly what I propose (map lanes explicitly) and it's > against the separate-ways-only-when-physically-divided-paradigm > (because an ambulance could change from one way to another)... There's a fun question, what counts as "physically divided"? I'm seen many emergency vehicles cross "physically divided" roads. If we exclude those, then what counts as a vehicle for the purposes of dividing a way? Are bicycles includes? Four-wheel drives? > Usually there will be ~100mtres for this where you can > change at any time, but in OSM you have to decide on one merging > point. I haven't seen a proposed scheme yet that really deals with this properly, they pretty much all pretend that a lane starts at a point. Personally I don't think that figuring this out is necessary for having multiple lanes, as we could continue the fiction that they start at a point. It'd be nice, but not necessary. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] how to map this? cycleway or footpath?
On 10/09/2009, at 9:01 PM, Sybren A. Stüvel wrote: > Perhaps my perspective is very Dutch, as here in NL you are always > allowed to walk on a cycleway. When a pedestrian sees a cycleway on > the map she'll know that she can walk on it and use it as a footpath. > However, when a cyclist sees a footpath on the map, he'll assume he > won't be allowed to cycle there. > The standard mapping in NL boils down to: > >- Only pedestrians allowed => highway=pedestrian >- Also cyclists allowed => highway=cycleway This is where it starts to run into the "what does designated mean?" argument :-\ I would agree that if there is a sign/symbol indicating you can ride there (bicycle=designated) and not a sign for pedestrian traffic, then it's a cycleway. Are the ones you're allowed to cycle on signed as such? The problem is when there is no signage - for example in some parts of Australia (ACT), you can ride on footpaths unless it's signed otherwise [0]. Simply following the above rule would mean that 99% of the footpaths are marked highway=cycleway, which I think would be silly since although you can ride on them, they are primarily meant for foot traffic. I also seem to recall somewhere (Germany?) where you can't walk on designated cycleways without a sign indicating you can. I don't really want to get into this argument again, but I believe that either we're going to end up with local rules for the access mappings, or some regions are going to have to tag every single cycleway/footway with overrides. Personally, I think the former is better because it's a lot less work and there are going to be other things that need local interpretations - such as whether highway=residential should be practically treated as access=destination for the purposes of vehicle routing. [0] I didn't know that earlier, but someone mentioned it in the last debate and it did explain why cyclists kept almost running into me when I lived in Canberra. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] how to map this? cycleway or footpath?
On 10/09/2009, at 8:32 PM, David Earl wrote: > Therein lies the problem with each of these debates that comes up > every couple of months - while everyone would agree* that cycleways > accommodate cyclists, the rules vary around the world about what > else is allowed by default. > > I don't see why we can't have a convention/system that copes with > this, but people have tied themselves in huge knots over this in the > last year. Most of the arguments I saw in the previous debate stemmed from two questions: * If a path can be used by both cyclists and pedestrians but has no signage (or has signage for both), should it be footway or cycleway? * What does *=designated actually mean? I think the first question mostly was the cyclists wanting cycleway and the non-cyclists wanting footway. Both ways are perfectly valid, and I can't see either being picked without flipping a coin. I won't talk about my opinion on the second, in the hope that we won't have another argument, look at the archives if you care. The big problem here is that the different groups have actively been using the alternative meanings, so we'd need to go back and edit a lot of data to make it consistent with whatever eventually gets chosen. Of course, not having a consistent meaning is worse than having to edit the data. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] how to map this? cycleway or footpath?
On 10/09/2009, at 8:05 PM, John Smith wrote: > 2009/9/10 James Livingston : >> Because of the presence of the bicycle symbol on the ground, I'd say >> highway=cycleway;bicycle=designated;foot=yes. If that wasn't there, >> I'd say footway=yes;bicycle=yes > > Isn't that redundent? > > I assume highway=cycleway to imply bicycle=designated I do too. However after the last discussion a lot of people seemed to think it only implied bicycle=yes not bicycle=designated, so I'd add it explicitly because of the designation marking on the ground. Of course, I don't think anyone agreed on what designated meant either... ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] how to map this? cycleway or footpath?
On 10/09/2009, at 7:01 PM, Valent Turkovic wrote: > how should I map this - > http://www.flickr.com/photos/valent_turkovic/3900795904/ > > highway=cycleway + pedestrian=yes > > OR > > highway=footway + bicycle=yes > > Are these two the same? What is the difference? Be prepared for a long drawn out argument on this, like the one we had last month on this topic. Because of the presence of the bicycle symbol on the ground, I'd say highway=cycleway;bicycle=designated;foot=yes. If that wasn't there, I'd say footway=yes;bicycle=yes ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] New dimension of vandalism
On 28/08/2009, at 9:23 PM, Alex Mauer wrote: > On 08/28/2009 03:46 AM, Gervase Markham wrote: >> If dieterdriest has found a number of people who've been ignoring the >> definition, > > Nobody (that I know of) has been ignoring the definition. It's just > that the definitions didn't match the top-leveldescription. *None* of > the definitions of the highway values has ever described the physical > characteristics of the road, apart from motorway in a very limited > sense. Indeed. I'm wondering how things like "Administrative classification in the UK, generally linking larger towns" (primary, from the highway=* page) could possibly be taken as describing the physical structure of the road. Personally, I think that the road hierarchy from trunk down to tertiary doesn't really have a strong definition, and that region- specific mappings (the International Equivalence table) is what people generally go off[0]. The distinction between motorway and trunk seems to be somewhat consistent globally, and the sub-tertiary values (unclassified, residential, et al) have globally useful definitions, even if people don't always agree on what they are. [0] Although I just had to fix the Australia section to match the Australian Tagging Guidelines and what people do. The ABC classification mappings were off by one level ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] New proposal: Bad data
On 27/08/2009, at 9:09 AM, Martin Koppenhoefer wrote: > I'd say the only thing you know for sure is that the source is unknown > unless it is explicitly tagged. I wouldn't assume anything besides > that. There are people who don't upload their traces (i personally > always do) and who have all rights to not do it. Not uploading traces is perfectly fine. If someone doesn't upload their trace _and_ doesn't add a source tag, then I'll assume that my GPS trace is more accurate than whatever they used. I think that's fair, because all someone has to do if they have an accurate way is let other mappers know that by tagging it. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] New proposal: Bad data
On 26/08/2009, at 7:31 PM, Liz wrote: > we've had a lot of trouble in Au because group X decided that > unmarked was > landsat and they would mark survey, and group Y decided that > unmarked was > survey and they would mark landsat I take the approach that unmarked is landsat, yahoo, or something else that deserves to be checked/improved unless there is a public GPS trace. > could we simply extend source=survey with a year > and source=landsat similarly? > > source=survey09 > source=landsat_trace09 > source=yahoo_trace08 That sounds like a good plan. I was going to suggest source:date=*, but it seems that things that way around are already used by source:name to say where the name came from. It sounds backwards to me, but date:source? ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [RFC] Deprecating the use of Tag:highway=stop in favour of Key:stop
On 26/08/2009, at 1:10 AM, Lester Caine wrote: > I think the point here is that of being able to see easily what has > been > applied to the data. Nodes and ways are easy to see, but this extra > data > is probably not so obvious as you would not know that a node ON the > way > actually has extra data, or perhaps that some other relation is > involved? This is starting to head in the territory where it depends on what editor you're using, or what tool you use to visualise the data. I'd argue that is your chosen tool doesn't tell you when something is a member of a relation, it should be fixed :) Using Potlatch, a node in a way belonging to a relation is just as easy to notice (blue outline) as a node in a way which is marked with tags (black fill). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [RFC] Deprecating the use of Tag:highway=stop in favour of Key:stop
On 26/08/2009, at 1:38 PM, John Smith wrote: > I agree, we need more tags to describe the railway crossing's > feature set, boom_gate=no, lights=no etc, however this is a special > case for stop signs because they will exist either side of the > junction and never applies to the railway line. Unlike junctions of > road traffic which needs to be differentiated from the way. This brings up an interesting question, when you're "finding the nearest junction" to use for stop key on a node, what counts as a junction? It's going to be a node which belongs to the current way and at least one other way satisfying certain conditions, but what are those conditions? If we are to use the stop key, I think those conditions will need to be explicitly spelt out, so that you can process the data. One obvious possibility would be ways that have highway=* tags - should footway/cycleway/path crossing count as junctions? If we're going to automagically determine which junction the Stop applies to, why do we even need a new key with yes/both/-1 values? Surely we could just say that if the existing highway=stop tag is applied to a node belonging to a single way (and not an intersection, which has the current meaning), then the Stop applies to traffic on the current way approaching the closest junction. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] talk Digest, Vol 60, Issue 157
On 25/08/2009, at 9:37 AM, Roy Wallace wrote: > What is everyone's preference? I quite like the relation described at: > http://wiki.openstreetmap.org/wiki/Proposed_features/Relation:type%3Dstop > > In fact, that relation avoids the need to split the way at the > junction if the stop sign applies in both directions along the way > through the junction. I like the idea of that, but think it might be better if the member ways had roles like 'stop' or 'give_way'. I know several intersections which have both Stop and Give Way signs, so being able to use a single flow-control relation to represent both of them would be nice. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [RFC] Deprecating the use of Tag:highway=stop in favour of Key:stop
On 25/08/2009, at 10:22 PM, John Smith wrote: > --- On Tue, 25/8/09, James Livingston wrote: >> Or we could just always use a relation, so that mapping and >> software >> don't have to check for two different things, when editing >> and >> processing data respectively. > > Or in other words, tagging for the routing software, this > information can be dealt with in pre-processing. Sorry, I had a typo in that sentence - it should read "so that mappers and software ...". As well as software, it makes it easier for mappers who wouldn't have to check arbitrary nodes around a junction. I completely agree that tagging for renderers/routers/whatever is a bad idea, but I'm not certain this counts. You shouldn't tag things incorrectly to make it render properly, or make a tagging scheme worse to that some piece of software can avoid doing some work, but what is wrong with the tagging scheme making it easier to process if the scheme is otherwise just as good? Of course, whether it is just as good is pretty much what this argument is about. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [RFC] Deprecating the use of Tag:highway=stop in favour of Key:stop
On 25/08/2009, at 7:39 PM, Pieren wrote: > That's exactly what I say. Stick the node where the real life stop > sign is - nothing arbitrary here If you're arguing that the tagged node should be where the stop sign is and not on the intersection, surely it would be beside the way not one of the nodes forming it? Most of the stop signs I've seen aren't stuck in the middle of the road. > then let software calculate the > distances and find the nearest intersection. If and only if the > nearest intersection is not easy to find (e.g. because the distance > from the two intersections is the same or nearly the same), then add > the relation. Or we could just always use a relation, so that mapping and software don't have to check for two different things, when editing and processing data respectively. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [RFC] Deprecating the use of Tag:highway=stop in favour of Key:stop
On 24/08/2009, at 8:53 AM, Roy Wallace wrote: > I don't like this, because "before" is arbitrary. If the stop > requirement applies to the intersection, I think it should be applied > to the intersection itself (either directly or as a member of a > relation). I agree that these kind of things should be related to the intersection or way, rather than an arbitrary node before the intersection. What happens when someone wants to reverse the direction of the way? Currently you need to check the tags on the way, in case one of them is direction-dependent - I don't want to have to start checking all the 'nearby' nodes in case one of them has a tag which is dependent on the direction of the way. Personally, I think that using a relation (and splitting the way if necessary) would be nicer than having to check a bunch of nodes in case one of them has an way-direction-sensitive tag on it. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Business Building Conventions
On 20/08/2009, at 10:29 AM, Andrew Ayre wrote: > If I draw the outline of a strip mall (a connected string of shops) > this > represents several businesses together. If I then put nodes on them > and > give the nodes names Mapnik won't render the names unless they are > amenities. But not all businesses are amenities. I'd say that you need to tag them as what they are. If they're shops, then use shop=*, and if it's a company's office then we need to have a tag for that and add it to the renderers. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [english 95%] A process for rethinking map features
On 16/08/2009, at 2:20 AM, Tom Chance wrote: > Probably sensible to start with something more manageable than path/ > highway. > Maybe the forest/wood debate. Sounds good to me. The important thing is that the group has their goals set out explicitly, so they know exactly what they should be doing, and they know when they're finished. To me, this means that we need to collect a complete list of all the tree/wood/forest-related things that people may want to add to OSM (even if they already have tagging solutions), with good descriptions, if possible photos and what implications people think they have. The WG could then sit down and figure out which of them are actually the same, and then find a good tagging scheme. I think the "complete list of what we want to tag" is something we're missing in the current arguments. How are we supposed to know if what people are talking about are actually the same thing? Especially since language is an issue, either not having English as a first language, or not having the same English (e.g. British vs Australia vs American). > The one missing part to work out is how we respond to the proposal. > The best > thing I can imagine is if we could set-up a poll that uses our > OSM.org logins > and we notify as many users as possible through every channel > available. We > could set the bar at something like >1000 votes and a 66% majority > needed. I think that if the WG comes up with a solution after taking into account, then it would likely be acceptable to most people. If not, then it probably didn't represent a crosssection of the community, or people didn't add their items to the list of things to tag. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Proliferation of path vs. footway
On 12/08/2009, at 10:38 PM, Greg Troxel wrote: > But if there is "no default" for foot, then what is > routing software to do? If it uses the way, the default is yes, and > if > doesn't, it's no. So the notion of no default does not make at lot of > sense to me. ... > With highway=path, the wiki page does not give the semantics when > there > are no tags. For highway=path and no tags, is that horse=yes or > horse=no? Is it paved or not if there is no tag? What I'm trying to say is that not having the tag would mean "currently unknown" rather than "depends on local defaults", and so someone should find out and add the missing tag. The same way that not having a maxspeed tag indicated that we don't know what the maximum speed limit is, rather than there not being a limit. Obviously software processing the data will need to pick a default, but while editing it would mean that someone should improve the tags. > The biggest problem is that there needs to be an unambiguous mapping > From these highway=foo tags to the implied value of the access > subtags. > The next biggest is non-operational semi-circular definitions like > 'highway=cycleway' being for 'designated cycleways' which talk about > 'intent', although in practice one would ask (in en_US) "do most > people > think this is a bike path". I think it's mostly around the use of the word "designated". Some people (including me) take that to mean "there is a sign, or other signal present saying that it's for bicycles", but other people obviously disagree. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Proliferation of path vs. footway
On 12/08/2009, at 8:14 PM, Pieren wrote: > Note that in France, pedestrians are not allowed on cycleways. I don't > see why we should add "foot=no" now in all cycleways in France. I read > somewhere that some motorways in US gives access to bicycles. Does it > mean that we have to add "bicycle=no" to all other motorways in the > world ? Either you have to do this, or software (in this case for routing) needs to know about all the quirks of different jurisdictions, which is a point I tried to make in the Australian-residential-tagging thread. In practice you probably wouldn't do it in the routing software itself, but when transforming the data from OSM format to whatever the router uses. Having tags whose implications change from region to region is asking to confuse people. Back when I lived in the ACT (Australia), I wouldn't expect to be required to tag things differently if I'm a few kilometres away from home, across the border in NSW. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Proliferation of path vs. footway
On 12/08/2009, at 3:51 PM, Nop wrote: > There is no consent on which way to go to express the strict use case. I think the only two solutions are to either have this be country- specific (at which point routers/renderers have to start knowing these kinds of things), or we have highway=cycleway not imply any value for foot at all. Going the other way and not having highway=footway imply any value for bicycle would mean that people like me could tag something as a footway and say that I don't know whether it's suitable for cycling on by leaving the bicycle= out. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Do we care if its forest or wood? Natural worldmapping ...
On 09/08/2009, at 8:17 AM, Jason Cunningham wrote: > Wood and Forest have not had clear definitions for centuries in the > UK, and as Mike Harris states the trees within Forests were > incidental (the famous Sherwood Forest was mostly heathland). Just because it's called a "Forest" doesn't mean that it should be tagged as landuse=forest. I know several a few Lakes without water, and Beaches that no longer have a beach. > Looking at the discussion Mike Harris has already suggested the tags > I would suggest, but I may as well repeat them > natural=woodland land covered with trees (Minimum Crown Cover = 20%) > landuse=forestry Does landuse=forestry mean that it is a managed forest (like landuse=forest was supposed to), or that it is an area used by the timber industry? For the latter you could make the distinction between natural=wood;landuse=forestry (old-growth) and landuse=farm;produce=tree (plantation). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A process for rethinking map features
On 11/08/2009, at 11:27 PM, Tom Chance wrote: > The principal reason for suggesting SOTM is that - in my many years of > experience with these matters - it's incredibly hard to sensibly > discuss > complex matters online. With a good facilitator and a well defined > process > of preparation, you can often solve these matters in a 100th of the > time it > would take over IRC or through a mailing list + wiki. I completely agree, and this works really well for open-source projects. It does mean that the discussion is biased towards those from the region where SotM is being held. That isn't really a problem for most open-source projects, but probably would be for OSM due to the fact that a lot of arguments fall along country/region lines. Take us Australians for example: if some proposal was discussed at SotM this year in Amsterdam, you are unlikely to have many of us present to say how things work in this part of the world. Then when it came time to present the results of the discussion to the general community, you'd get us complaining that we hadn't been involved and that you'd not taken into account how things work here. It's nothing to do with us specifically, but you'd get the same thing from any group who is under-represented at a given SotM. I imagine that SotM is like the other (generally Linux) conferences I've been to, where there is a lot of discussion outside a formal event, so that allowing remote people to participate via IRC and video wouldn't really help. I don't have any solutions to this, but think that doing it at a conference won't really solve a lot of issues. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] A process for rethinking map features
On 11/08/2009, at 12:49 AM, Tom Chance wrote: > - If the proposal throws into question existing, accepted tags, > defer the > proposal to small working groups This sounds good, but deciding the makeup of a working group is likely to be an issue. If you allow anyone to join, then your working group won't be small. If you limit the membership, then you're going to have a fun time making up the rules for such decisions and how to deal with the (possibly large) group of annoyed people who don't get to be in the working group. A working group is pretty much going to need a very diverse set of members otherwise you'll get into situations where how things work in country X isn't taken into account, and people in country Y mis- interpret what the resulting tag means. There is also likely to be some discussion in the community before a vote, which I think would end up just being the same as the current arguments here. > - At SOTM present and discuss their proposals and vote As others have mentioned this is bad because it penalises those who can't go to SotM. IRC meetings could work, but as soon as you get more than a certain number of people involved they need to be moderated, and then tend to go on for a *long* time. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] residential and unclassified in Australia WAS definition of the main highway-tag
On 06/08/2009, at 12:58 AM, Renaud Martinet wrote: > There has been a lot of discussion on the talk-fr list but once we > came to a consensus, > it was easy to put in place because we have our own MapFeatures > page. Probably > you should have one also... That might work for countries where English isn't the main language, but I doubt it would work for Australia and similar countries. I would guess that many people would just read the general MapFeatures pages and follow their interpretation of that, rather than notice the small links to a country-specific page. It'd probably be impossible to find out, but it would be interesting to know how many Australian mappers (not just those here or on talk- au) actually know about the Australian Tagging Guidelines page. I had been doing some mapping in OSM for about two months before I happened to stumble upon it, but in that time I misinterpreted several things that were written using UK/European terms. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] [tagging] Feature Proposal - Voting - 4wd_only
On 05/08/2009, at 5:54 PM, Roy Wallace wrote: > The sign says "4WD ONLY" - I therefore suggest that 4wd_only is indeed > the correct terminology, at least in regions (e.g. Australia) where > the sign appears as such and the phrase is in common use. While true, it would also be useful to know whether you can't drive an average sedan up the road, or if you need to bring your recovery equipment (after checking it still works). ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] residential and unclassified in Australia WAS definition of the main highway-tag
On 03/08/2009, at 11:23 PM, Martin Koppenhoefer wrote: > 2009/8/3 James Livingston : >> In any case, if you have a router that does this kind of thing, >> wouldn't it be better to base it off landuse=residential/industrial? > > the problem is, that it is far more timeconsuming to check this for > all roads instead of having the information already avaible as such. It'd probably take a bit longer to convert from the OSM data to whatever format your router actually uses, but it also means you could treat roads in other landuse areas differently too. > well, tag whatever you like, I just can tell you, that the definiton > in the wiki says for residential, that there must be at least at one > side residences. The highway=residential wiki page doesn't directly say that, but may imply it. The problem is that a lot of the words used seem to be based on the British way of defining roads and that doesn't necessarily translate into non-British English very well, let alone into other languages (as seen in some of the other discussions). Most of the Highway page talks about British road classifications, and things like "(tertiary) In the UK, they tend to have dashed lines down the middle, whereas unclassified roads don't", which doesn't really help people figure out how it is supposed to apply to other countries. What exactly does "This tag is used for roads accessing or around residential areas but which are not a classified or unclassified highway" mean? If you take 'highway' to be a synonym for 'road' then suburban residential streets shouldn't be tagged like that because they are unclassified. If it's not a synonym, then how do industrial streets get tagged, because they're not highways. In addition the "Australian Tagging Guidelines" (which Liz mentioned were written a year before the residential page) explicitly disagree with the residential page. Which brings us around to one of the major questions in this argument. If the consensus (which may exist in Europe, but I'm far from certain is global) is to use one definition, but within a region there is a consensus to use a different definition, what do people want to happen? > If you don't care about this definition, do as you > like. You'll IMHO loose a datum and gain nothing. There are other ways of storing that data (e.g. landuse) and roads in Australia aren't tagged according to the highway=residential wiki page at the present time, so what exactly do we lose? We might not be able to use exactly the same routing settings as in Europe, but I'm pretty certain they are never going to work as-is anyway, simply because things are different over here. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] residential and unclassified in Australia WAS definition of the main highway-tag
On 02/08/2009, at 9:56 PM, Martin Koppenhoefer wrote: > yes. A residential road should be avoided if possible (slow, dangerous > and noisy for residents / playing kids), while I don't see this in > industrial or commercial context. Not having been to Europe I can't say for sure, I wouldn't say that in Australia. I'd generally prefer residential over industrial roads, because the latter have more trucks, more variability in road condition (due to heavy vehicle damage), and the like. In any case, if you have a router that does this kind of thing, wouldn't it be better to base it off landuse=residential/industrial? On 03/08/2009, at 7:50 AM, Martin Koppenhoefer wrote: > sorry, but I can't believe that. All roads in your country have the > same width? The same minimum radius for curves? They don't, but that's more to do with tertiary <-> residential/ unclassified than it's not really on an industrial/residential basis - what we tag as tertiary is different to what we tag as residential in both areas. > (IMHO) residential: > http://maps.google.it/maps?hl=de&ie=UTF8&ll=-37.675859,145.165879&spn=0.000252,0.000597&t=h&z=21 > (IMHO) unclassified (~25% wider in the aerial): > http://maps.google.it/maps?hl=de&ie=UTF8&ll=-37.769521,145.02807&spn=0.000504,0.001195&t=h&z=21 Sure, and I can find a heap of examples where they're the same. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] definition of the main highway-tag
On 01/08/2009, at 7:38 AM, Martin Koppenhoefer wrote: > Well, you can do this, but most routers will try not to use > residential roads if there is another way. Maybe things are different over in Europe than here in Australia. My Garmin when using commercial maps and a friend's NavMan are both more than happy to take you down what we have been taggin as "residential" streets, if it will save you a few seconds time. The exception is roads marked as "local traffic only", which I'd tag as "access=destination". > both of these are IMHO not valid for industrial > areas, that's why your aussie-way might produce slightly worse routing > results (don't know, just an idea). If we ignore the Australian tagging guidelines, what should we use for roads that are the same as residential ones but in an industrial area? According to the wiki, unclassified roads are wider than residential ones, and while roads in industrial areas are often wider than residential ones, that's not always the case. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] question about untagged green points
On 01/08/2009, at 1:36 AM, SLXViper wrote: > Those untagged nodes sometimes show up quite a lot, depending on the > area. In one case there were a lot of untagged nodes along a way, each > very close to another node belonging to the way. All untagged ones > were > created by a potlatch user, I think someone tried to draw another > way on > top of the first one, but failed doing so. I've seen this once before, about three or four months ago, and I also probably performed the action (in Potlatch) that lead to it occurring. I made a note to myself to report it when I'd finished the task I was currently doing, but seem to have forgotten to pay attention to the note. Oops :( What happened in my case was that there was an existing way (riverbank) which was not particularly accurate, and we had gotten some much more accurate data. So I copied the tags from the existing way to my new one, editing all the places where it joined to anything, and then deleted the old way. In Potlatch it looked fine. When I came back to the area about a week later, I saw that all the nodes that were part of the old (now deleted) way were still on the map. It looks like the way had been deleted but the nodes, unused by anything else and with no tags, had not been deleted. As SLXViper noted the nodes appeared very close to the current way, which was because they were just from a less accurate version of the same thing. Sorry for not reporting this earlier, it had slipped by mind due to being a bit busy at the time. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Business listings
On 29/07/2009, at 5:45 AM, Jack Stringer wrote: > Should we be charging to upgrade businesses details on OSM? > > I think it should be free. You could pay OSM to have a OSM member put > all the details onto the map for them, saving them signing up etc. But > I would not like to see charging being the norm. Only because OSM > exists as a free map service, the same I believe should go for the > Business data on it. I'd say it depends on what you are charging them for. I can't imagine we (as a community) would be happy with someone using OSM to scam money out of people, but there are ways to get money from businesses that I think would be fine. I know a lot of restaurants (and other businesses) that have *really* bad maps on their websites. What if you charged them a fair amount of money to put their business' location in OSM, give them the HTML needed to put an OSM-based map on their site, and went and checked the streetnames and landmarks in their area? You get paid, they get a map that people can actually find their business off, and OSM gets better data. ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
