RfC: deprecate hashtag-only changeset comments

The kind of thing you mentioned actually happens every month in China. Most of it is because many changeset just describe an area that no one has ever described. It covers everything from improving streets to improving roads. It is difficult to summarize. We are commonly known as “pioneer on wild land”
Editors may not necessarily have the patience to fill in every changeset as carefully as they would like to fix a place, but out of the principle of not dampening the enthusiasm of editors, most mapper active in community is usually more tolerant of such bad comments, and in rare cases they can’t help it, will be reported to DWG
for example:

Does this comment make sense? Yes, but the 200+ consecutive ones are all the same “drawings on villages and other situations in the Zhangjiakou area based on satellite images such as MPI and Esri.”
But I decided not to report him a second time, because he is the only mapper in the whole city. I have always believed that countries or regions that allow everyone to write changeset comments must be regions with basically complete data.
As the saying goes, “Only when you are full will you consider more issues.”

For a prefecture-level city that requires 1000+ changesets to complete, it is pale and powerless to require any detailed changeset comments. I think this is not good, but forcing it may not achieve the purpose of improving the drawing quality by improving changeset comments.

Original reply in Chinese (Yes, above is machine translated, I'm lazy) 你说的这种事情其实每个月都在中国发生,大部分是因为很多changeset只是描绘一片从没有人描绘的地区,从完善街道到完善道路无所不包,很难去做总结,我们俗称“开荒” 编辑者不一定有耐心像修修补补一处地方一样仔细的填写每一处changeset,但出于不要挫伤编辑者积极性的原则,社区通常对这类bad comment更容忍,很少情况下忍不住了才会向DWG报告 举个例子: ![Zhen SHi|611x500](upload://vdLtSaqqTwpJqswHwGGfQWxDRdX.jpeg) 这comment有意义吗?有,但连续200+个都是同一份“根据MPI及Esri等卫星图像,绘制的有关张家口区域的村庄等情况。 ” 但我决定不会举报他第二次,因为整个城市只有他一个mapper,我一直认为能让所有人都好好写changeset comment的国家或地区,一定是已经数据基本完善的地区。 正所谓“仓廪足而知廉耻”。

对于一个需要1000+ changeset才能完善的地级市去要求任何detailed changeset comment都是苍白而无力的,我觉得这不好,但强制不一定能达到通过改善changeset comment来改善绘图质量的目的


What if it also resulted in 10 mappers “writing” (letting generate) adequate changeset comments (instead of useless ones)?


Until we have OSM software capable of mind control, any auto-generated comment is going to be no better than e.g. “added X buildings in Y place” (with another tag containing source or imagery). It’s literally no more information than is in the .osc file already.

1 Like

This is true. For example, I recently wanted to split an overly large industrial area into several small pieces (c143777767+c143778354). I might write a list of reasons to explain why I want to do this (because it can be expressed more precisely, because here It’s not all industrial areas and there are greenfields, and attach those reason in comment)
But if editors are asked to better describe their intentions, and their intentions may be simply to “pioneer on wild land”, how should we guide them in writing comments? For example, the case I took the screenshot above (first of all, I think that even blocking his account again will not solve all problem), how to describe those edits?

Original reply in Chinese (Yes, above is machine translated, I'm lazy) 这确实,比如我最近想要拆分一个过于庞大的工业区为若干小块(c143777767+c143778354),我可能会写一串理由来解释我为什么要这么做(因为能表示的更精细、因为这里并不全是工业区还有greenfield) 但如果要编辑者更好的描述他们的意图,而他们的意图可能仅仅是简单的“开荒”,我们应该如何指导他们写评论呢? 比如更上面我截图的这个例子(首先我认为即使再次封号也解决不了问题)

But it does not need to be (much) better than “added X buildings in Y place”. As a fellow mapper, I just need a clue what was changed to even decide whether I need to review the changes (for whatever reason). I don’t go policing after fellow mappers, but when I inspect an area using OSM history or Neis OSM Tiles the rough descripton of a changeset works just fine. I need to dig deeper using OSMCha or similar tools only if the comment indicated something of particular interest for me. And the bloke who only ever “changes some things” also has a mixed record of good and bad edits, so I’m seldom reassured when I find his name in the area history.


Well, sure, but what I’m saying is that literally 9 out of 11 examples of good changeset comments follow roughly (or exactly) that form, possibly with the source added. So when both a large amount (majority?) of existing useful comments as well as the documentation follow that form, and that form could (presumably) be generated, why shouldn’t we provide that generation to mappers?

Sure, no tool will be able to provide the why as in “added buildings I found during a bike ride through the town”, but it should be achievable to generate “added buildings based on a survey” (based on the contents of the changeset and the source= tag). We can still encourage people to write more (and maybe some even will who don’t today, having been freed from specifying “what”/“where”/source in their comment), but I actually think this might improve the overall quality of the comments.


9 posts were split to a new topic: Autogenerated Changeset Comments

The discussion turned away from the original question (deprecate hashtag only comments) to which comments do we want to have and how are they created.

Until now nobody said it was bad to have additional comments to changesets, many said it was good.

Therefore I support the change @Nakaner proposed in RfC: deprecate hashtag-only changeset comments - #21 by Nakaner as first step.

1 Like

In addition, the JOSM Remote Control API also supports adding hashtags when you “open in JOSM”.

Since it’s easy for data contributors to use the hashtags key, I support this “stop hashtag-only comment values”

1 Like

#I #would #like #to #ask #about #the #alignment #of #this proposal #with #the openstreetmap #values.

For instance, will the ban on hashtag-only comments be consistent with the OpenStreetMap Diversity Statement?

In other words, would it be discriminatory to target an age group (like Generation Z ) or a subculture that primarily uses hashtags,
or are these not protected by the OSM Diversity Statement?

Furthermore, how should this proposal be phrased so that it does not appear to be against Generation Z and avoids unintended consequences?

A post was merged into an existing topic: Auto-generated Changeset Comments

I’d say, stating that banning hashtag-only comments implies discriminating against Generation Z is an oversimplification of the facts. To begin with, not every Gen Z person actively uses the hashtag symbol in the sense relevant to the proposal, but also, I see little to no reason why Gen Z people couldn’t write more meaningful changeset comments, whatever those are (I think a discussion of that is outside the scope of my reply here, if not of the entire thread).

In the end I am but a minor contributor in this vast ecosystem (a Gen Z one!), and while diversity is an OSM value, and one I very much support, it’s not the only core value of the OSM project, and in my personal opinion, the improvement of the accuracy of OSM data is more important than alleged discrimination based on a vague relation between the use of a single typographical symbol and an age group.


Yes, equally bad because it doesn’t contain a description of the action contained in the changeset. Reviewing the changeset becomes at least partially guesswork.

I concur, and in my opinion we should sanction the organiser, not the occasional mapper who are just following bad instructions.

Nope. I for one choose which changeset to review based on the changeset description. Or I need to open each badly described changeset to check if it touches elements that I can review.

1 Like

Most people who use hashtag-only changeset comments do not do so because they think it is cool but because someone else - or a piece of software - told them that’s the way to go. I don’t think we’d be curtailing anyone’s free expression if we ask them to write human-readable comments; even if hashtags were somehow the lingua franca of a specific generation, a sentiment I don’t agree with, that would not be a valid reason to use that single-generation language in a multi-generation project like OSM.

I’d also politely request that this discussion not be derailed by whataboutism - “if you are against X then you must also be against Y and HERE’S MY ARGUMENT WHY BEING AGAINST Y IS BAD”. I have not suggested to reject contributions from users of software that forces them to use auto-generated tags and therefore I am loathe to discuss how useful (or not) they are in this context.


You have a point there, but a little redundance is not always bad! If someone writes “traced buildings in South Warokee” and instead the changeset has moved a POI to another part of town then it is likely that the POI move was a mistake, whereas if the comment reads “moved restaurant to new location” then we know that it was intentional.

Redundance is always part of human communication and language, and it makes a lot of sense especially as the group that is communicating becomes more diverse (different backgrounds, languages, cultures etc).


Then the core should be how to help mappers accurately describe their drawing motivations(such as “moved restaurant to new location”). But I think many times the mapper may not know exactly what it is doing. It just copies the elements that appear on the satellite map. It is as simple as “refinement”. How should they write comment at this time?

Anyway, at least I think the editor should give users a chance to choose

1 Like

You are again not discussing the original topic. The original topic is not about deprecating auto generated human readable changeset comments (like Streetcomplete or Every Door do), but hashtag-only (and therefore most of the time not human readable).

Your example of a bad changeset comment has one hashtag #added and then some words, so it is no hashtag only comment. Actually I don’t know why you put the “#” there, it is quite uncommon I think. The changesets of Streetcomplete and Every Door don’t use hashtags.

Even if it is off topic: I like the changeset comments of Streetcomplete and Every Door, because they describe what was done in a human-readable way.


“Mapping of xyz from aerial”?

Here source=foobar imagery and even auto-generated list of what was modified is good enough.

Though if user deletes data based on aerial then more involved description would be useful.


Yes, we should be more cautious about modifications and deletions, so the requirements on changeset comments are stricter (especially deletions)

A small tweak to the Organized Editing Guidelines which could be useful in this context is to change “Changeset comments should include” to “Changesets should include”. As written, the OEG currently do not allow the hashtags to be put into the changeset’s hashtags=* key.