Documenting the history of OpenStreetMap

Following my last diary entry, I would like to discuss how to deal with the wiki pages containing outdated content. There have been some deletion requests (including reverts, list) and some template applications so far.

  1. Which pages should be kept in the wiki for documentation purposes? (Background: some pages like “Kosmos” feature extensive subpages, some proposals were just edited once)

  2. Should they be changed to reflect their historical relevance? (How?) (Background: Are templates like ‘Historic artifact’ useful to avoid confusion with current content? Should we summarize long pages?)

Tigerfell

Edit: The post is available from https://www.openstreetmap.org/user/Tigerfell/diary/47605.

Generally proposals should be kept. Only if there is 0 usage of the proposed tagging they could be removed, ideally by their proponent, or if they are not active in OSM any more, the proposal could be put on a public deletions-list and automatically deleted after some time like 3 months or so. But usually tags from “inactive”
proposals are in use anyway, hence the proposal should be kept for documentation.

The case I referred to when commenting on https://wiki.openstreetmap.org/w/index.php?title=User_talk%3AMateusz_Konieczny&type=revision&diff=1784192&oldid=1708076 was page place=xian proposal. I requested to delete it, but I was reverted by Mateusz Konieczny. But I guess we agree in this case on deletion.

There are two lists already: Labelled for deletion for straight-forward and not disputed cases and Deletion Proposed for pages with ongoing discussion or proposals (in case the people were not sure and want someone else to review, before requesting deletion (first list)).

Going through the list of deletion requests, reverted by Mateusz Konieczny today https://wiki.openstreetmap.org/w/index.php?limit=50&title=Special%3AContributions&contribs=user&target=Mateusz+Konieczny&namespace=&tagfilter=&start=2019-02-01&end=2019-02-01, I request deletion in the following cases as well (pretty sure, I did not request it in the first place, because I just received 1 notification):

  1. https://wiki.openstreetmap.org/wiki/Proposed_features/employment_agency (same as number 1. proposed case, no history to keep: page contained information about the existence of the tag before the proposal already, nothing actually proposed)

  2. https://wiki.openstreetmap.org/wiki/Proposed_features/Tag:natural%3Dflower (author requested deletion of their own proposal, no occurrence in the database, apparently violates “Do not map temporary features”)

  3. https://wiki.openstreetmap.org/wiki/Proposed_features/Tag:natural%3Dfungus (similar case as above, original author and ‘deleter’ distinct)

  4. https://wiki.openstreetmap.org/w/index.php?title=Proposed_features/arena&oldid=436387 (original author figured out that other tags existed already and cancelled the proposal)

In the following cases, I would have proposed a deletion (meaning discussion):
  1. https://wiki.openstreetmap.org/w/index.php?title=Proposed_features/castle&oldid=1325615 (cancelled proposal, 2010, expanding the scope of historic=castle to apply it to areas as well as nodes (in use on nodes previously already), archived on a user page, 2 deletion requests by different users reverted by different users already)

  2. https://wiki.openstreetmap.org/wiki/Proposed_features/Bag_shop (unifying different spelling variants, proposed new tag already became common before voting already, so no voting took place, few additional information on the page and its history (how proposed, alternatives, definition of a bag shop)

I would not suggest deletion in the following cases:

I will not continue with the following seven proposals, but I guess you get the point…

Thanks for going through them Tigerfell. I agree with your assesement on all of them. Its a good starting point.

Thanks @Tigerfell for starting with the case studies, this is a good start. In the end I’d like to have a list of criteria that can be applied to these cases, to make decisions more objective. Naturally however, the criteria for keeping and those for deleting will always have some grey area between them, whereever the threshold is set.

In general, I see the threshold for deletion quite high. OSM lives from the community discussing its changes, and this discussion is expressed in the history of the communication channels. We do keep archives of mailing lists for example, and nobody sits there and weeds out mails that are not relevant to the topic. Equally, the history of wiki pages is an elementary element for transparency, and to understand reasons for decisions and developments. This particularly true for proposals.

So I’d like to start with an incomplete draft for both lists.

To delete

  • Typo pages, e.g. a page was started with a misspelling in the title, the page has been moved to the correct one. The remaining redirect stub can be deleted.
  • Spam unrelated to OSM
  • Violations of people’s privacy / data protection rights
  • Draft pages on request of the original author, that have not attracted discussion from other users
  • Pages that are contrary to basic principles in OSM (e.g. proposing to severely violating copyright, or Good practice). However, if this violation of principles is more subtle and therefore being discussed or disputed, that is a discussion to preserve.
  • Utter nonsense (needs further definition, maybe the bullet above is one)
  • Proposals that only mention a tag, that is not being used, without describing/explaining its meaning

To keep

  • Proposals with a vote
  • Proposals for tags that are being used, even without a voting process
  • Proposals that serve as a documentation for a tag, even if it is not widely used. This should prevent that a tag is being used for a different purpose later. Feel free to add a taginfo counter to such page.
  • Proposals that attracted other users who contributed to the page or its discussion. Feel free to add deep-red boxes about the proposal being abandoned or the tag being used inconsistently.
  • Proposals independent of their age, i.e. age should not be a criterion. OSM itself is just 15yo.
  • Pages linked from other pages

I don’t think “propsals that have attracted users” should be included in that. Most proposals have had minor “house keeping” edits done to them at some point by either one or multiple users. So it essentially disqualify 99% of the potential propsals right off.

The amount, quality, and usefulness of the information should be the criteria. Not that a few people over a ten year period changed “RFC” from upper to lower case, changed a comma to a period, or anything else along those lines.

Id also include “propsal pages with broad definitions of objects that aren’t osm specific and can be found somewhere else.” I’ve had a few deltion propsals reverted on pages that were something like boat=whatever, then a line half way down saying “A boat is a thing you put in the water to get across lakes in”, but nothing else. Then the person doing the revert said it was because that information was “relevent to the history of osm.” Which is ridiculous. So the content criteria for saving a page shouldn’t be “any comtent.”

As far as the age thing goes, I only use it as a measure for mostly empty pages as one veriable to help judge if the page might be worth saving in case someone might come along and edit it at some point. Like, if the page is blank, only has one edit, but is only a year old, who knows if the person who created it will come back or not. Whereas, with a ten year old page with no edits (beyond supperifical ones) or content since its original creation probably has zero chance of original creator of it ever wanting to come back and revise it. Plus, a close to zero chance of anyone else wanting to try and revive it. That’s the only situation where age plays a part for me and I never said otherwise. Obviously it would be stupid to delete all propsal pages after they have been around for a few years and I wouldn’t suggest it. Although for some reason I was treated as if that’s what I was doing.

Id also add propsals for already existing tags where the person that created the tag is “oh, I just realized there’s already a tag for that” and then abandons it. In those cases the content is often just a duplicate of the tags propsal page, or if there is something unique it would be more useful to transfer it to the propsal page of the tag that’s being used instead so people will actually read it.

We already have a list of general criteria on page https://wiki.openstreetmap.org/wiki/Delete which is based on someone’s experience, but it is more general. I hoped we could limit it to the proposals for now, because they seem to be very controversial.

The difference between mailing lists and the wiki is that ML are a historical source, but the wiki is a documentation based upon such sources (ironically the proposals serve as a source themselves). We seem to keep a distinction between the two kinds of pages with counters on feature (documentation) pages like https://wiki.openstreetmap.org/wiki/Tag:office%3Demployment_agency and proposal boxes on the proposal pages like https://wiki.openstreetmap.org/wiki/Proposed_features/Tree. This is done to avoid the people changing the proposal after the vote or getting irritated.
Thank you for the list, Polarbear. You find my opinion below:

To delete

  • I have the feeling that it rather kills the administrators’ time, so I would prefer redirects, but oh well…

  • +1

  • +1

  • +1 → I think that should be okay even if others request deletion after a discussion or a long time of no activity.

  • wording seems okay, but I think this is really tricky: in my list above, there were two proposals about mapping flowers, but we do map trees, right? Maybe the length of the discussion could be an indicator?

  • would limit this to hoaxes (I remember something like cuisine=high_quality, which was supposed to be a joke)

  • +1 (like in case a. https://wiki.openstreetmap.org/wiki/Proposed_features/employment_agency)

To keep

  • +1

  • Well, what if the proposal mentions the tags but they were used before the proposal already, so the proposal effectively did not influence the tagging?

  • This would relate to failed or cancelled proposals only. I am undecided in this case, because I also see the task of the wiki or encourage common tagging, so that we do not need to document tons of alternatives and discourage users to choose the less used one.

  • I guess you refer to significant content additions (excluding categorisations and spelling corrections). Adding this criterion would mean that proposals set up by multiple people would be less likely to be deleted than the others. I fear that people confuse proposals with the documentation and therefore change the proposals, not because they want to contribute.

  • Age should be a criterion (same opinion as Adamant1), because some users start with a proposal page that has a very low quality and then improve this over time while discussing the content on the mailing list. Since it is rather difficult to check every list (thinking of foreign languages), age is a good criterion to estimate if something is a draft or abandoned.

  • I prefer to value the quality of the links, because essentially every proposal is linked to a category like “Proposals with status x”. Apart from that, wiki users create links to make it easier to find a proposal relating to a specific topic. We once had a page version of Types of relation listing basically every proposal that could include a relation and even ideas of non-existing proposals. This would have basically blocked every relation proposal from deletion. In addition, “delinking” is part of a deletion process, you find better pages and finally figure out that the page itself is useless and request deletion.

Agree, current example: Bag_shop (this also compares different ones, so I would discuss first)

Go ahead and define objective criteria to measure them.

Indeed I mean users who contributed to discussion and content, not typography.

No idea what that means.

A criterion for being abandoned, yes, but being abandoned does not mean to be deleted.

Sure. What I mean is that content on other pages refers to this; not automatic categorisation.

Yes.

Quality and usefulness High quality and usefulness imply: not a duplication of other proposals, contains a definition + a comparison to existing or other proposed tags (like “a=b is better than b=a because …”, “you can not express … with the tags a, b, and c”)

I am talking about the conjunction of different factors here, for instance if we have a poor proposal which seems to be incomplete, and it was never voted upon, then its age is an additional factor for arguing in favour of deletion, because the chance that the original author comes back are pretty low.

minor edits, spell check…

I the problem are lists, updated by humans comparably to “manual categories” and poorly related “see also” links.

I forgot links originating from user pages like User:A/todo → Proposed features/B. In addition, User:A/todo was not modified for five years, but I guess this is obvious…

I thought the same about natural=fungus and the original proposal was indeed not so good.

Fungi however are the biggest and oldest living organisms on earth so I replaced that proposal with a new one.

+1

The cost of deciding what to delete is pretty high even in reasonably clear cases because it is an irreversible act purging all content and history. Once a delete request creates controversy this cost goes further up substantially.

What is the benefit of deleting a few ancient useless proposals? Does it ever justify the work involved?

Are there better alternatives to permanently purging useless old pages? Aren’t a few amboxes good enough? Moving to another wiki-namespace?

You avoid maintenance tasks such as categorisation (changes over time continuously), watching out for vandalism, and reverting destructive edits. In addition, you are able to focus the documentation of currently used content (not that relevant for proposals, but in case of software lists or recommendations/comparisons).

Lastly, I guess some users like to clean up and condensing content (a similar process happens in the brain with our memories).

Going through the list set up by Polarbear, the process is straightforward: You check if it qualifies for deletion and otherwise you leave everything as it is.

Honestly, you could think of a read-only system for old proposals. Employing such a system would save some maintenance, because no one needs to watch the changes and you can basically keep everything (like with the mailing list archives). There would be quite a lot of “low quality content” though (proposals we currently delete for instance, although I am not saying everything I would delete is of a “low quality”, some things are just duplicates or discussions that never took place because no one was interested, proposals based on misunderstandings…).

Such a system would not be like a wiki, but rather like a database storing a log of all actions with no option to change it later on. User Ezekiel “stored” some of the proposals on a user sub page, but I dislike that because still everyone could change it, but it is more tricky as it is in a user space and it is not really related to this user.

Still, you would need to decide when a proposal is ready for archiving and this is not trivial either…

Edit: grammar, wording

I don’t think whether to do basic up keep tasks on a website or not should be based on an “outrage meter.” Since it can easily lead to self-selection bias, where the outraged people think their opinions are in the majority and most important. Simply because they are in the outraged group themselves. There is also a strong chance of confirmation bias that way also. Since strong, negative opinions are over represented, because people who are fine with something don’t usually comment about it. Or if they do it’s usually after the fact to complain that things didn’t turn out how they wanted. I see it happen all the time, both here and on the main style’s Github page.

Therefore, it’s all around better to have actual quantitative guidelines to follow, that we can keep unbiased as possible and be able to point to later if someone complains or wants the page deleted again. Outrage tends to ebb and flow and can always be disputed. The fact that a proposal page shouldn’t be deleted because it has votes doesn’t though and it can’t be disputed as easily. Someone would have to revise the guidelines to get their way. Which is much harder in practice then finding a couple of other people willing to be outraged at the same thing as them.

In general, the more we can step out of our own ways (which also includes me) and “automate” this stuff the better IMHO. It reduces needless discussion and arguments if nothing else. I’m sure we all have better things to do with our time. It also partly deals with the possibility of someone using shill accounts or similar tactics, that could create a fake controversy and make it seem like there’s a majority opinion on either side when there actually isn’t.

I merged the suggestions into the draft and published it at https://wiki.openstreetmap.org/w/index.php?title=User:Tigerfell/Crafting&oldid=1785670. Note that I moved some general content to the top section.

I was not sure about point 3 in “To keep”:

I understand the intention, but what if the tag’s meaning changes, because someone else sets up a new proposal and ignores about the few previous occurrences?

as to point 4, I would remove the exceptions (except for relevance and deletion discussions). Obviously, if the only discussion is the autor (that’s what the paragraph is about) raising an issue for deletion and nobody is defending the page, that would mean deletion, but otherwise, if there were arguments in favor of the proposal, I would keep it.

No, the intention was to include an option, that *a second person *could step up (after one year without changes) and propose and discuss about a deletion of a draft meaning incomplete proposal(use cases: initial author lost interest, did not know how/if to delete, abandoned draft without the qualities defined in https://forum.openstreetmap.org/viewtopic.php?pid=736853#p736853). In this case, the deletion action should not be blocked by the fact that there was a discussion with the final result leading to a deletion.

The key point is this “second person”.

I was hoping that every discussion would finally lead to some result (one or the other way).

Edit: highlighted insertion

I think one reason why the discussion was so heated in the first place was because people were using {{Delete}} instead of {{delete proposal}}, perhaps not even knowing the better alternative???

Even if an old proposal is total garbage it is almost never very urgent to delete it. I am thinking that even if the case for deletion is pretty clear it might be much better to use {{delete proposal}} so that ideally a few users can leave their comments making the decision for the admin easier. Delete proposal also has the big advantage that in case the decision is to keep the page the discussion about deletion will be preserved and won’t repeat every few months.

The https://wiki.openstreetmap.org/wiki/Delete already says that {{Delete}} should be used only “If you are very sure that a deletion is the right way and you have some experience with this wiki” but apparently this is not enough. It should not happen that someone puts the {{Delete}} template on 13 pages out of which only 3 turn out to be uncontested.

I would suggest to very strongly discourage the use of the {{Delete}} method on anything except “own” content.

I would also suggest to prolong the deadlines to respond to a delete proposal by at least an order of a magnitude. Currently it says

This might be enough at wikipedia but here many people are lucky if they manage to login every month and discussions tend to be slow and long lasting. We have now some really old proposals laying around so I do not understand the hurry to get rid of them so quickly.