Talk:OpenAddresses.io

Warning text

With respect to the Warning text, OA contains data that has been obtained by scraping sites that have been left open due to security holes and were/are clearly not to be publicly accessed. Further in many cases if licence/usage terms haven't been found or not in English, nonsense has been entered in the licence field. Whoever is maintaining the dataset does not seem to be prepared to proactively vet what is being added (as that would be work and would likely make the whole project untenable), it is, as said, also not possible to determine a legally responsible person/entity.

The warning was added as concret response to an OSM contributor questioning why data that he couldn't use because of restrictive lience/use terms, was being bandied arounf on OA. SimonPoole (talk) 17:17, 30 November 2017 (UTC)

Hi Simon. Instead of writing inflammatory, underhanded attacks on a wiki page warning box, how about you participate and report these deficiencies so they can be fixed? You've already done that and we've responded, so I know you're capable of doing it. If there are still problems, then let us know! Instead of participating in a revert war, why don't you propose updated text that still makes your point but in a more even-handed way? Or maybe let someone else do it, since you clearly have a problem with me or the project. We're chatting on this GitHub ticket if you'd like to make suggestions for improvement rather than just hitting the "undo" button. Yellowbkpk (talk)

As you know I have reported a few selected issues in the past (in an OSM context that would have been enough for the DWG to go back over all the persons contributions and check all of them). But essentially as you are suggesting again, your policy is to make QA somebody else problem. SimonPoole (talk) 18:13, 30 November 2017 (UTC)

So now we are down to Tom Lee rewriting history and deliberately twisting the truth and adding aggressive spam marketing to this page. I will suggest that the page is deleted to stop further misuse. SimonPoole (talk) 00:11, 1 December 2017 (UTC)

Spam marketing? With all due respect (not very much at this point), what are you talking about!? Yellowbkpk Yellowbkpk (talk) 00:39, 1 December 2017 (UTC)

To all users involved in the ongoing edit war: Please behave sensitive, don't do edit wars and discuss your changes beforehand.

I have the impression that this page is the battlefield of an ongoing (?) edit war. I would like to ask all how are involved not to edit the page directly but to copy the content to a page in your user namespace, do the modification their and suggest the version on this talk page or a suitable mailing list. I hope that this will cool down your heated hearts.

The page is currently in an ugly state. It contains statement by OpenAddresses.io which Simon claims to be advertisment. He wrote his opinion below these paragraphs. Please get to a consensus which is not considered as soaked with advertisements. A good page describes both views, e.g., "OpenAdresses.io claims X but it seems that …". In addition, please add references to your allegations in order that others can vet them.

If I see no progress, I will ask an admin to revert all edits by User:Tom Lee and all edits done after these edits, i.e. the page will look like this. In addition, I will ask to lock this page for some time.

Please play the game and don't violate the unwritten Code of Conduct which IMHO contains the rule "Don't do edit wars." --Nakaner (talk) 12:39, 1 December 2017 (UTC)

I'm disappointed to see my edits have been received so poorly and am eager to defuse this conflict. My goals were to answer some of the questions raised by SimonPoole (for example, his assertion that the people behind OpenAddresses.io weren't clear--a fair criticism, so I added contributors and relevant organizations); acknowledge and expand upon the central point of his edits, which seemed to be warning OSM users that OA data cannot be imported to OSM without additional scrutiny (a point I kept featured as the first section after the introductory text and elaborated upon at length, while retaining the warning box); and to make the page more useful as a wiki entry by adding context and links to additional resources (an introductory talk, the github page, some stats, etc -- nothing objectionable, as far as I can tell).

I was prompted to undertake the effort in the first place because of edits that included claims that were derogatory and false. I understand that Simon dislikes our project and disagrees with how we approach our work. That's unfortunate, but I know we are unlikely to resolve those differences here. I do think it's reasonable to expect acknowledgement that both sides are operating in good faith. Similarly, I hope we can all agree that claims that our efforts are "potentially criminal" are extreme and don't belong in this wiki. The claim that we "do not validate entries submitted to [the project] in any form" is also false.

I have no desire to fight someone of Simon's stature in this community, but I think it's reasonable to ask, prior to any consideration of reverting my edits, for specific details about which parts of my edits are false or in violation of OSM wiki policies. Deletion would be a regrettable outcome but would be preferable to reverting to the page's previous state, in which it included false and libelous information -- that is not an acceptable outcome. --Tom Lee (talk) 14:47, 1 December 2017 (UTC)

Just so that I don't get accused of libel: Question (from the OA repo); Are we comfortable redistributing the data when we're unsure of the licensing? Answer (Tom Lee): I think this has to be yes. And there are lots of similar statements and sources the prove the point in their repo (just in the small sample that I checked) that support everything I've pointed out. Essentially their position is that anything they can retrieve programmatically is fair game completely regardless of the local legal and regulatory situation, regardless of if the data was made available on purpose or just accidentally. That is not our business, the principles of OA have to live with their own decisions, using the OSM wiki as a marketing platform and claiming exactly the opposite -that- is not OK. SimonPoole (talk) 20:12, 1 December 2017 (UTC)

We didn't create this page. We did improve it by adding factual information and links to talks, though. There are plenty of other pages on the OSM wiki that have a similar form. Vespucci's, for example, has links to its home page, GitHub repository, and is marketing for itself with a link to the Google Play Store. As Tom said, I would rather see this page deleted from the OSM wiki than have it contain misleading, incorrect, or libelous information. --Yellowbkpk (talk) 20:32, 1 December 2017 (UTC)

Redistributing data for which the license cannot be fully determined is not a criminal act--a good thing, too, since not every edit in OSM can possibly be vetted in this way (although of course the contributor agreement does shift liability; and I hasten to acknowledge that the OSM community does strive for a high level of certainty around data source licenses). This is the nature of large-scale, collaborative geodata projects.

OA collects and indexes publicly available data. When we receive takedown requests, we comply with them. Although I realize you dislike it, this is not an uncommon strategy--see Google and Archive.org, to name just two precedents. Here is a link to the full comment you excerpted -- it contains these same caveats and limitations on the quoted portion, as well as a discussion of the practical difficulty we often have in determining a license, even after contacting the officials running the sites in question and affirming their intent to share the data.

But as I said: I know you will not be convinced in this venue. I consider our policies to be reasonable; you do not. I think the best outcome for OSM Wiki users is a clear explanation of what OpenAddresses' policies are and how they relate to OpenStreetMap. I think the page does that well right now, but additional constructive edits are welcome (at the risk of repeating myself: specific suggestions for improvement of the content would be very helpful).

Other than that I can only echo Yellowbkpk: we didn't create this page. Our interest is in improving it and correcting some inaccuracies it contained. I think these changes will benefit OSM wiki users, but if other admins feel deletion is preferable, I can accept that decision. Reversion to its earlier inaccurate state is not an outcome anyone should consider acceptable. Tom Lee (talk) 21:56, 1 December 2017 (UTC)

Lets get this straight:

• you claimed that you are only distributing meta-data, now you are now saying that you are actually distributing the data
• you claimed that you "make a good-faith effort to ensure that the data sources the project indexes are available under open licenses that fit within with broadly-defined "Freely Shareable" or "Share-Alike Required" categories" however admit that you don't care and will distribute it in any case (btw I did not claim that was criminal)
• you do retrieve data from obviously non-public arcgis servers (wasn't aware that google and archive.org do that), bad luck for the operators if they didn't secure them properly and have a business model which boils down to "har har we're in the USA, sue us if you want".
• somebody in country X using data that you nicked from a server in country X could very well run in to serious legal issues outside of copyright (how is the user supposed to determine if they are allowed to use the data if it isn't from a public source?), which my original warning from way back was designed to convey.

Further notes: Justin Meyers is now contributing to the OA repo and as always is squeaky clean about what he is doing, but he also shows the clear problem for OA, it is a lot of work and slow process.

Back to this page, it had been in a stable state since September 17th with essentially no content, the warning added by me (and a small factual correction). The warning itself was and is completely factual, that does not mean that is was something that OA wanted to read, but then this isn't OAs website, or is it? The comparison with the Vespucci page is farcical, Vespucci is an editor for OSM data, OA has nothing to do with OSM, just as we would not host a page for google extolling the virtues of google maps. The only purpose of the page was to give a short explanation what the OA is and clearly warn against using it. Not to provide a platform for a competing project to attract more contributors.

-- User: SimonPoole 23:32, 1 December 2017‎

Great, so we agree! Let's delete the page. Yellowbkpk (talk) 23:37, 1 December 2017 (UTC)

Feels worth noting that I don't agree with the above characterizations but yeah, deleting the page seems like the best path forward--I'd like to find a way to communicate what the OSM community needs to know about OA but this conversation's stuck in a loop. That's a shame, but if this wiki can avoid making pejorative statements about OA that's good enough I guess. Tom Lee (talk) 21:13, 4 December 2017 (UTC)

The page shouldn't be deleted. "find a way to communicate what the OSM community needs to know about OA" Yes please. That's what I intended to do when I created the page in the first place. Although not just what the OSM community needs to know, but also an outward expression of what the OSM community agrees is a fair description of the topic.

A little quote from the Wiki guidelines may be useful here: "On these pages there is no point duplicating lots of 'about' information found on the external site. The page should describe the site in an OSM context. Be more neutral and less promotional with your description, although do aim to use language which promotes OSM and uses of OSM.". Obviously this is a guideline we could expand into a page long wiki editing policy if we enjoyed writing policies like they do at wikipedia, but hopefully it's clear. Projects which are obviously aligned with OSM's aims would be promotional in their presentation and language here on this wiki. If that is something people are casting doubt on, then we'll have to try to be more neutral, describing the project factually, and expressing concerns people have.

Simon is obviously on a mission to be negative about it, but I don't think OA is on the same level as Wikimapia for example. Compare that wiki page. It's describing a website and project which we (OpenStreetMap) have every reason to bad-mouth and point out the negatives of. The wiki page is our chance to do so. To explain why wikimapia sucks and you should contribute to OpenStreetMap instead. But actually the language used on that page ends up being, not all that strong, partly because people have made small wiki edits in defence of wikimapia, and partly because the original author (who was that smart guy? ;-)) knew that such edits would happen, so he attempted to strike a neutral tone and stick to the facts when highlighting the negatives.

I would've thought OpenAddresses is nowhere near the same level of anti-OSM-evilness as Wikimapia, so I would expect we would arrive at a description which is mostly positive, but with some criticisms mentioned and a clear pragmatic description of the problems around importing from it.

...and maybe we have. The current page looks pretty good.

-- Harry Wood (talk) 02:01, 6 December 2017 (UTC)