Re: [OSM-dev] Reverts from the woodpeck_repair account
On 3 January 2013 14:43, Pieren pier...@gmail.com wrote: On Thu, Jan 3, 2013 at 2:59 PM, Andy Allan gravityst...@gmail.com wrote: Don't deliberately enter data incorrectly for the renderer We should add : Don't accept tag changes until the renderer supports it. Pieren The reason that OSM doesn't do depreciation is that there is no such thing as the renderer. You don't know if I have a private(or lesser known) render that relies on a used tag. And it's not limited to renderers, but editors, data sources, and non-visual uses/applications (such as driving directors). Introducing new tags should also take this into account. For example we had railway=rail and railway=tram type tags. I might make a map that displays railway=* as a grey line (because I'm more interested in some other detail but I want tracks for context). For a short time railway=historic got into popular use. My map would then start showing railways where they weren't! A better choice for the new tag is historic=railway because nobody is going to render that as an existing railway by mistake. Slowly people can start recognising the historic tag if they wish to. I can't remember the exact tags in the true version of this story, I think it was actually to do with stations. -- Gregory o...@livingwithdragons.com http://www.livingwithdragons.com ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On Wed, Jan 2, 2013 at 6:12 PM, Frederik Ramm frede...@remote.org wrote: Pieren said that the specific highway=ford edit was discussed before but I think it has already been pointed out that this is wrong; discussing a new tag is not the same as discussing a mass edit to convert old tags. But the message om 2010 said.: Any suggestions on what to do with the 4800 nodes also tagged with highway=ford? Change them to ford=yes all in one go as well? According to taginfo, we have currently (2013-01-02 00:58 UTC): 5981 highway=ford on nodes 282 highway=ford on ways 11215 ford=yes on nodes 5911 ford=yes on ways Perhaps because ford=yes is also in JOSM presets... But until mapnik OSM stylesheet supports the tag, we will have people reluctant about the change. The same who says that we don't tag for renderers... Pieren ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On Thu, Jan 3, 2013 at 2:59 PM, Andy Allan gravityst...@gmail.com wrote: Don't deliberately enter data incorrectly for the renderer We should add : Don't accept tag changes until the renderer supports it. Pieren ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On Thu, Jan 3, 2013 at 9:43 AM, Pieren pier...@gmail.com wrote: On Thu, Jan 3, 2013 at 2:59 PM, Andy Allan gravityst...@gmail.com wrote: Don't deliberately enter data incorrectly for the renderer We should add : Don't accept tag changes until the renderer supports it. I don't think Frederik was reverting because of questions about whether a tag is supported by a renderer or even if the tag is deprecated or supported. I think he was doing it because mechanically changing tags world-wide should be discussed. It wasn't discussed, so the change should be undone. ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
Hi, On 01/02/2013 02:31 PM, Paweł Paprota wrote: For me such reverts just introduce unneeded clutter in the history database - thousands of new versions in the history. Yes, that's sad, and I'd surely prefer not having to do such reverts in the first place. But what do you suggest? If someone makes a mass edit without discussing it with anybody, in blatant violation of our guidelines, simply shrug and ignore it? In OSM, a lot happens by example. If you want to do something, you look at how someone else did it and you do the same. If we were to simply ignore mass edits like this then everyone would assume they're ok. There's a huge warning on top of this page: https://wiki.openstreetmap.org/wiki/Deprecated_features And it isn't there for fun. And the end result will be that this revert will then again be slowly reverted by individual mappers replacing deprecated tags... The concept of deprecated tags is not widely used in OSM and I'm not a friend of the word because it suggests too much authority. The problem is that it takes just 15 or 20 people to deprecate a tag. Many tags in the aforementioned list of Deprecated features are actively used by mappers, and often the deprecated tag is more widely supported than the new tag that 15 or 20 people would like to see introduced. For a mass change of tags in the database, it is not enough to have 15 or 20 people on a wiki page somewhere who think that it is a good idea; you can work towards such a change but it would require a much broader consensus than the few people who are interested in tagging discussions on the wiki. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On 2 January 2013 13:31, Paweł Paprota ppa...@fastmail.fm wrote: For me such reverts just introduce unneeded clutter in the history database - thousands of new versions in the history. And the end result will be that this revert will then again be slowly reverted by individual mappers replacing deprecated tags... That's only one possibility. There are deprecated tags that are more popular than their voted on replacement, that have better support in editors, are used by more renderings and are still preferred by mappers. The assumption that deprecated tags will be replaced by other ones is a false assumption, and hence why we don't automatically change tags. Cheers, Andy ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On 01/02/2013 02:44 PM, Frederik Ramm wrote: Hi, On 01/02/2013 02:31 PM, Paweł Paprota wrote: For me such reverts just introduce unneeded clutter in the history database - thousands of new versions in the history. Yes, that's sad, and I'd surely prefer not having to do such reverts in the first place. But what do you suggest? If someone makes a mass edit without discussing it with anybody, in blatant violation of our guidelines, simply shrug and ignore it? Not ignore it - for sure there needs to be a conversation with the author of the original changeset so that it is clear that such large changes should not happen without consultation. What I meant by my message is that we seem to be shooting ourselves in the foot in pursuit of free tagging, no rules etc. Note that I'm not willing to discuss this at length as every discussion about this topic seems to lead to nowhere. I would like to keep this thread to one (or couple of) specific example(s) that I mentioned. In this specific case, what is the value of reverting other than making a point that people should not be doing such edits? Is the value greater than the cost paid in the long term by polluting the database and potentially discouraging people from editing like this in the future? As for the tagging related points you mentioned - I unsubscribed from tagging@ about a week after subscribing to it... Ultimately there should be no cases like this where there are huge changesets which basically bring no value at all to the project and just change syntax sugar (tags). How to do it? Well, that's a challenge, I have some ideas and am willing to do development but for now I want to finish with OWL and better history tab which is equally important topic... Paweł ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On Wed, Jan 2, 2013 at 2:55 PM, Paweł Paprota ppa...@fastmail.fm wrote: Not ignore it - for sure there needs to be a conversation with the author of the original changeset so that it is clear that such large changes should not happen without consultation. This change has been widely and publicly discussed on the 2 mailing lists: http://lists.openstreetmap.org/pipermail/talk/2010-October/054554.html http://lists.openstreetmap.org/pipermail/tagging/2012-August/011230.html What I meant by my message is that we seem to be shooting ourselves in the foot in pursuit of free tagging, no rules etc. +1 You cannot complain in one side about OSM tagging complexity and block all attempts to simplify it on the other side... Pieren ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
On Wed, Jan 2, 2013 at 9:13 AM, Pieren pier...@gmail.com wrote: On Wed, Jan 2, 2013 at 2:55 PM, Paweł Paprota ppa...@fastmail.fm wrote: What I meant by my message is that we seem to be shooting ourselves in the foot in pursuit of free tagging, no rules etc. +1 You cannot complain in one side about OSM tagging complexity and block all attempts to simplify it on the other side... I think we're trying to prevent artificial attempts to simplify. For example, I don't like the amenity tag so I'm going to download all amenity=* tagged items in the database and replace amenity with elephant and upload it. ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
Hi, Not ignore it - for sure there needs to be a conversation with the author of the original changeset so that it is clear that such large changes should not happen without consultation. As I said - this would ideally make the author of the changeset understand that he made a mistake, and ideally he wouldn't do it again, but the next day you'd have someone else believing that they do something good by changing 100,000 objects from one set of tags to another. In this specific case, what is the value of reverting other than making a point that people should not be doing such edits? I am not sure if there is a value other than making this point (only a proper discussion would have been able to establish that), but making this point is reason enough. This specific case consisted of a number of changesets where the author seemingly went through the list of deprecated tags and did *exactly* what the big banner at the top of the list told him not to do. It says that if you make a mass-edit without prior discussion then it will be reverted, and that's what we did. If we *really* wanted to mass-change a deprecated feature into something else, we could do that very effectively on the database servers themselves, but we don't. Is the value greater than the cost paid in the long term by polluting the database and potentially discouraging people from editing like this in the future? Yes, definitely, because for every mass edit where you say I don't think this is too bad there will be five others where a baby has been thrown out with the bathwater - and allowing outright policy violations for those that make sense means we'll have even more of those that don't. Changing thousands of objects around the world with a script is simply not something that you can decide for yourself and execute without talking to anybody first because you are very likely to make mistakes. The concept of deprecated tags is problematic. I have already asked the maintainer of keepright.at to stop marking them as errors and make them warnings instead, which he has thankfully agreed to do. There might indeed be situations where a mass edit makes sense but such edits may have wide-ranging consequences and they absolutely must be widely discussed before, no matter how well-intended they are. Pieren said that the specific highway=ford edit was discussed before but I think it has already been pointed out that this is wrong; discussing a new tag is not the same as discussing a mass edit to convert old tags. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
Hi, You make some good points but I still can't help but think that there has to be a more efficient way of preventing such things. It seems to me that changing and then reverting stuff back in this manner is a lose-lose situation for everyone. I see two problems here - one is the inability to agree what tags should be used, attempts to optimize tags etc. This problem is out of scope of this thread but I think a lot can be done to improve the situation, e.g. improving tools for documentation and discussion of tags. The second problem is the automated editing. Perhaps now as OSM becomes more and more popular it is time to start looking at some more general solutions to these kind of problems with data and bots. Just a quick glance at what Wikipedia is doing[1] shows that there is potential room for improvement in OSM in this area like introducing some workflow around automated editing. I believe there is a middle ground between OSM's free for all, do what you want, we will worry afterwards approach and Wikipedia's you can do X only after requesting permission Y from working group Z. I guess this is something that could be discussed with DWG? [1] http://en.wikipedia.org/wiki/Wikipedia:Bot_policy Paweł ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
Hi, On 02.01.2013 18:28, Paweł Paprota wrote: I believe there is a middle ground between OSM's free for all, do what you want, we will worry afterwards approach and Wikipedia's you can do X only after requesting permission Y from working group Z. There was the idea of introducting a bot flag for an account, and then require that you absolutely must have the bot flag for any sort of mass edit (something that might even be technically enforced). Requiring a different type of account would at last make sure that people understand that mass edits are an activity different from normal edits, and it would automatically put these edits under more scrutiny. How you *get* the bot flag would have to be subject of a different discussion; it could be that you can simply give yourself this flag (possibly going through a web page where you have to click on yes, I confirm that I have read and understood the guidelines...). It could also be that you get the flag only after enough people have agreed, or after some working group has rubber-stamped your plans, or whatever. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09 E008°23'33 ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
Re: [OSM-dev] Reverts from the woodpeck_repair account
Hi, The second problem is the automated editing. Perhaps now as OSM becomes more and more popular it is time to start looking at some more general solutions to these kind of problems with data and bots. The solution is simple and straightforward: A database design must be able to cope with the the edits that are uploaded. If you don't consider all edits valuable, you are free to drop data in the target database. By contrast, any additional decision logic in the main API does really clutter OSM because the main database as sigle point of failure gets more prone to errors. More important, it may shy away mappers if they found that their particular situation on the ground cannot be mapped properly. SomeoneElse got his work damaged by other mappers, for no important reason. A less self-confident mapper may have been lost at that point. That's the reason why we have freedom of expression in tagging and mostly in producing changesets. Data consumers still have all freedoms to process the data to any rules they may find suitable, but no harm is done to the community or the core database. Cheers, Roland ___ dev mailing list dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/dev
