Talk:Quality assurance

Discuss Quality Assurance here:

---

2008: Undo (resolved)

I would like to add (or see :-) a section about "Undo changes / how to get information about "who did add/change/..." / vandalism and how to revert the data (if possible), but before I start investigating the topic I would like to get some opinions about it. So: opinions? Drbobo 10:39, 11 December 2008 (UTC)

On this page there is a program called osmdiff. You can use it to monitor an area of interest. It will tell you who did what. It cannot undo anything unfortunately. --Gary68 13:39, 11 December 2008 (UTC)

I wonder whether it is possible to revert the data, eg. after vandalism? Seems that currently there is no chance to revert it, http://wiki.openstreetmap.org/index.php/Talk:OSM_Protocol_Version_0.6. --Drbobo 16:00, 13 December 2008 (UTC)

See Change monitoring and Change rollback. For a run down of the (limited) techniques available at the moment. It's more technically tricky to revert changes on a map as compared to a wiki, but this is direction we developing in. -- Harry Wood 18:24, 15 August 2009 (UTC)

linked the rollback changes to the change monitoring. marked this section as resolved --Werner2101 (talk) 20:04, 5 May 2014 (UTC)

Remove unnecessary nodes

There are a lot of highways with nodes in a straight line. I think it would be a good idea to remove those unnecessary nodes with the utilsplugin of JOSM (simplify way). Less nodes = faster map drawing and less database size. With a simplify-way.max-error setting of 1 (m) there won't be any details lost. --GP4Flo 22:23, 12 August 2009 (UTC)

-- Lesson, don't throw away data you a don't understand. Unfortunately when GP4Flo edited/simplified a great amount of data he throw away a lot of very accurate information I had been amassing over a long period of time. Many of the paths that were accurate to less than a metre are suddenly missing nodes because the tool was ran with a large max-error setting. This type of behaviour should be considered VANDALISM and will drive away users with access to good data. Alpine walking tracks were effectively changed to straight lines! This tool should be used on individual traces and not on the complete dataset.

We have already talked on PM about the problem. The default max-error-setting was set to 2.5 (meters) which is a bit too high for some very small details. I'm sorry that this was the case with some of Evan's nodes. Changing the setting to 1 m should avoid such problems. In my opinion it's the question whether it does make sense to draw super-detailed paths out of GPS data which is only measured with an accuracy of (for the best case) about 2 m. I've used the simplify option on a lot of paths I've collected with my GPS and couldn't notice any loss of important detail. You won't be able to see a difference of 1 m in your GPS device or in the slippymap anyway. --GP4Flo 12:47, 31 August 2009 (UTC)

Although we can't see a error of 1 meter in the slippymap for one way (as in most rural locations with current data density), it becomes quite visible once there are other features that have been positioned relative to that way; say a hedge, building or a cliff that was seen to be parallel to the road at that point and 6 meters away from the centerline (lane 3.5m + sidewalk 2.5m): if the "extra" node is removed from the road, the feature next to the road would then be at a wrong angle and possibly even overlapping the road. Just something to be aware of when simplifying in areas with other features present. Alv 15:01, 31 August 2009 (UTC)

List of all QA tools and services

Hi, would it be better to collect all QA services here or there: List_of_OSM_based_Services ? --!i! 16:53, 8 August 2010 (BST)

We've had this page for a long time as a useful list of OSM data bug detecting/analysing/reporting tools. These tools are very much within the ecosystem provided entirely for mappers to improve the data.

The List of OSM based Services ...is a list of OSM based services. i.e. services for people outside of the OSM world, but using OSM data/maps.

-- Harry Wood 10:15, 9 August 2010 (BST)

So ok I will add the tools to the list and add a nice list of the related tools here, too --!i! 12:21, 9 August 2010 (BST)

2010: Cleanup Request (resolved)

Hi, would be great is someone can resort the items e.g. into:

• bug reporting
• monitoring
• error detection
• assistants

Linking to List of OSM based Services, ... would be cool, too --!i! 13:57, 4 November 2010 (UTC)

I made a proposal for the order based on your idea. The List of OSM baded Services is linked at the end of the site under "more" or do you mean a more prominent way? --DINENISO

sections are splitted into the suggested tool categories. marked section as resolved. --Werner2101 (talk) 20:10, 5 May 2014 (UTC)

2010: Tool for highlighting dead weblinks (resolved)

A Quality Assurance tool for highlighting dead web links where 404 errors occur for the website tag. Should be a useful way to help people update old web addresses when things change?--Hawkeyes

So you talk about website=* tags and their values?--!i! 22:29, 8 November 2010 (UTC)

Yes, website=* tags. --Hawkeyes

keepright detects broken and bad links. marked problem as resolved --Werner2101 (talk) 20:08, 5 May 2014 (UTC)

FLOSM

I added FLOSM to the QM site. The contribution was reverted because of copyright issues with this service. My questions to prevent potential future copyright problems:

• Why are the road lists, where each missing street has a link to google maps, OK where this site is not?
• For which task is FLOSM written, when we are not allowed to use it for QM?
• Is OSMcompare also illegal?

Thank you very much --DINENISO 21:20, 10 April 2011 (BST)

The road lists thing you mentioned sounds like it could bad idea too. I haven't looked into it.

It's clear that we shouldn't use FLOSM to copy roads from teleatlas, and using it to gain a rough idea of levels of completeness is also something OpenStreetMap contributors should not be doing. External people assessing OpenStreetMap's completeness will inevitably use tools like this to make a comparison. We should avoid feeding that directly into our processes for contributing or deciding where to go mapping.

OSMcompare is also problematic. I wouldn't say illegal. As I say, people will inevitably create and use comparison tools. But it is quite likely to be illegal for OpenStreetMap contributors to use comparison tools for their contribution, because basically you're copying.

There's some tricky grey areas here, but fundamentally we're creating an open licensed map by re-surveying the world from scratch. That's the idea of the project. This Quality Assurance page should avoid listing such tools altogether. On the wiki in general we should either avoid linking to comparison tools, or make it very clear that contributors should not be using these tools while mapping.

-- Harry Wood 11:42, 11 April 2011 (BST)

Thank you for clarification, then services like OSMcompare and FLOSM perhabs should be listed in the List of OSM based Services when they are intended for external use only. --DINENISO 12:55, 11 April 2011 (BST)

Perhaps. Although we should label them clearly. Maybe a compact way of saying this is "Not to be used as a source"

But I'd also recommend creating dedicated wiki pages. This is always a good way of giving us plenty of space to express ourselves more clearly. If we had a page on FLOSM and OSMcompare then we could write some clear guidance on there.

-- Harry Wood 22:33, 11 April 2011 (BST)

Harry, could you please clarify why you see the use of tools like these by mappers as a problem? As long as they are used only to find out areas where we need to collect more data (on the ground or from aerial imagery with permission, NOT from the maps we compared to), I can't see where copyright would be involved. But as I'm not familiar with common-law based legal systems I might be missing something here. --Lyx 22:46, 11 April 2011 (BST)

See Completeness#License issues (Note that this page was written by me, but please feel free to discuss and adjust the text to reflect consensus) "As long as they are used only to find out areas where we need to collect more data"... fine but now let's just zoom in a bit, and oh look, we've just exactly identified a missing street. Might as well just copy the map. If you list a tool like this on a page along with lots of other quality assurance tools for mappers, that's definitely going to give people the wrong idea. -- Harry Wood 23:12, 11 April 2011 (BST)

Well for this kind of mapping applications we maintaine a list at Maps#Compare, so wouldn't this be a better place? --!i! 11:47, 5 September 2011 (BST)

Public transport validator

The links on Public transport validator are dead. I couldnt find the correct links but maybe anyone else knows them. Ogmios 07:48, 27 February 2012 (UTC)

Hello, a few months ago the PT map openptmap.org has moved to a more powerful server. It seems I forgot to care about the validator, SORRY! However, I'm not sure if the validator project is active at all. I will try to contact the maintainer and keep you up-to-date. Meanwhile you can reach the Validator here. It is still running on the opengastromap.de's server, but I'm not sure if it's working correctly. --Marqqs 18:09, 27 February 2012 (UTC)

Request to rename or clean up this page

This page is more "QA Tools" than "QA," with a heavy emphasis on the "Tools" section. It's a great summary, but it doesn't help describe the general process of reviewing data quality, how to work together to get this done, what acceptable levels of QA might be, etc. As such, rather than inform or help the reader, it just continues their quest for understanding OSM QA by burying them in an avalanche of research. I suggest we split up the two topics and move this content to a QA Tools page. Jeffmeyer 03:42, 7 December 2012 (UTC)

You are right. Do you dare to move the page and fix the pages that link here? --!i! 08:51, 7 December 2012 (UTC)

Change to bullet points instead of headings (level 3)

Just as further explanation for my recent revert and possibility to discuss if Mcke likes to: revert to version of 17:20:42‎ (going back to headers). If you don't like the h3 style then you can use a different CSS. The headings provide more semantic value and offer the possibility to link to a section.

Furthermore stuff like

</p><!--

--><p>

is not how a wiki should be - easily editable (therefore the wiki syntax).

A clean-up is needed, yes. But the headings have advantages. Maybe a merge with the other list (there is much redundancy) (see #List of all QA tools and services) could be useful.

I am interested what others think. At least I think this change should be discussed. And I would like to add that I still appreciate the will and effort to improve that wiki page. --Aseerel4c26 (talk) 22:32, 19 May 2013 (UTC)

The first thing that you see when going to this page is the huge table of contents with all the tools. And most of the sections on this page only have 1 or 2 lines. A section of 2 lines is not really a section, but more like a list item. It's true that directly linking to each section on the page is lost by using lists. This is what I overlooked, but is this actually being used?

Using lists the way that I did is maybe not the best way, but it improves readability and simplifies the table of contents. I used the comment and paragraph html code to put longer sections in one list item. This is the only way that I found (on Wikipedia) how to add breaks in the syntax of a list item.

There is still something that needs to be done to improve readability of this page. Mcke (talk) 07:16, 20 May 2013 (UTC)

Yes, right, most "sections" are very short. I was looking more at the first part of the page - Keepright and Osmose. Those are longer. However, maybe they are too long since they have own articles and the info here is maybe even outdated.

In general we could use anchors to provide linkability{{User:Aseerel4c26/VisAnchor|anchors to provide linkability}}, however that's also not easy to use. I don't know if it is being used (I also see no possibility to search for such section links (inside the wiki or outside)).

For the list style: At least we could leave away the html comments -

... blablabla.</p>   <p>new paragraph blablabla ...

also works (update: but without providing a new paragraph indication (empty line) in edit mode). The code with headings provides much functionality with very simple code, I like that. However, if I look at the "list of osm based..." page with the tables - the table wiki code is more complicated than your bullet point list suggestion... so, well, I am okay with trashing the linkability and easy code for your suggested layout. --Aseerel4c26 (talk) 13:09, 20 May 2013 (UTC)

I have reverted to Mcke's version with some minor changes (real unordered lists instead of faked ones), link template. Will re-add the new changes now... --Aseerel4c26 (talk) 22:17, 28 July 2013 (UTC)

New content re-added. By the way: there should not be a empty line between list items - a new list is started otherwise. I have removed those new lines therefore. --Aseerel4c26 (talk) 22:52, 28 July 2013 (UTC)

Chapter about RSS-Feeds

I'm wondering if it would make sense to add a chapter about RSS-Feeds in this article. Currently, the Feeds page is only linked in the "See also" section. Instead we could add a short description about the variety of QA-Feeds and maybe add a link to this qa-feed-aggregator tool: http://tyrasd.github.io/osm-qa-feeds/ --Tyr (talk) 09:09, 19 September 2013 (UTC)

I am currently not quite sure how we can split the feeds to the QA and the Feeds page to avoid duplication/redundancy.
I have an old, but still current, writeup about QA tools which includes a RSS section. Since it is in German it may be useful for the German version of this page (which is quite out of date sadly ... maybe we should re-think the style of how translation is done here. ToDoToDoToDo ... ;-) --Aseerel4c26 (talk) 18:09, 10 November 2014 (UTC)

Add relation-specific tools

• The Relation_analyzer is missing
• Servers/analyser.openstreetmap.fr is missing

--Voschix, November 2014

Please feel free to add them, it is a wiki! --Aseerel4c26 (talk) 14:22, 10 November 2014 (UTC) I have done so now. Please help to improve the descriptions! --Aseerel4c26 (talk) 21:11, 12 November 2014 (UTC)

List/Category of common tagging mistakes?

There is Template:PossibleSynonym - is there any more? A comprehensive list or category could help people maintaining QA tools to pick common errors they want handle but could not find anything. RicoZ (talk) 16:08, 24 March 2017 (UTC)

404/dead services

Looks like ITO Maps isn't working anymore. Should I remove it? Thanks -Valerietheblonde (talk) 20:15, 11 September 2017 (UTC)

http://product.itoworld.com/map/ works for me. However, I cannot find a click-path to actually see a map, I just see categories of map thumbnails and when I click one I see more thumbnails on a new page. Maybe it only works after login?
In general, I think removing a permanently dead service/tool is the best way to go. --Aseerel4c26 (talk) 18:48, 12 September 2017 (UTC)

Who stole the treasure chest?

Let's say I remember years ago there was a treasure chest etc. Let's just say "something", there on the map (within a certain bounding box), but now it is gone. The only tools I found that can search for changes within a bounding box only go back a month, etc. How can I tell that I am not just imagining it? Jidanni (talk) 00:45, 11 June 2018 (UTC)

There's quite a few tools/approaches described on Find the id of a deleted node which may help you -- Harry Wood (talk) 00:12, 5 August 2019 (UTC)

Need to add Atlas Checks

QA tests based in the Atlas project. See https://github.com/osmlab/atlas-checks

Adavidson (talk) 01:21, 6 December 2019 (UTC)