I’m convinced that using something else than Bluesky’s ATProto or Mastodon’s Fediverse for place reviews (and images, part of reviews in my point of view) is the equivalent of creating a new geo database (instead of using OpenStreetMap) when faced with the challenge of bringing enriched geo attributes to posts of a social network.
Edit : that’s the path I’m taking for https://cartes.app
Haven’t thought about that yet. It would be a nice workaround, avoids developing another platform.
Just some unorganized thoughts … Allowing only objective, verifiable content about places (e.g., photos, restaurant menus) and not reviews or ratings might provide much of the value while reducing the moderation required. Honest ratings usually require patronizing or spending significant time at a place whereas a dedicated contributor could quickly submit objective content for many places. The amount of content, rate of content submission, etc. can provide signals about popularity, just as the number of reviews on Google Maps does.
Hi folks!
Fully agreed! For people’s convenience, it may be helpful to reiterate that there are only two solutions today (1) mangrove.reviews and (2) lib.reviews and the latter is invite-only—i.e. you can read but cannot start writing reviews without being invited first—so that leaves us with mangrove.reviews as the only “contender”.
This isn’t a competition of course! mangrove.reviews use CC BY 4.0 which is a very permissive license. I’d have preferred a stronger copyleft license like OSM does, but I’m happy to use the same license (CC BY 4.0) as them so that we can pool our data together (if they wish to synchronise with us of course).
I’ve considered it but then decentralisation makes everything—from aggregated statistics to moderation (and especially moderation, from spam to vote manipulation)—100x harder. As a solo dev, I don’t have much resources to bootstrap this proposal ActivityPub/ATProto today.
If there are any details you can share with us that’d be great! I’d love to learn about it more.
That’s the beauty of ATProto. Decentralised data storage, but global record stream and cache.
I just created a new Lexicon app.cartes.comment and my little server listens to Jetstream and stores an aggregated osmCode → [comment] disk database.
The source of truth is user’s PDS (mainly on Bluesky for now). My local DB can be recreated in case I lose it (it’s not simple, but it works : you ask the network which user repos have the `app.cartes.comment` lexicon, then loop to retrieve records).
In consequence, lots of different ATProto browsers can display the comments as-is. Exemple.
Also read this blog post Using AT Proto for more than just Bluesky posts
It’s live in v1 since yesterday 
Lots of things remain to be done : license, moderation (Bluesky has done some work to help moderators), etc.
When you say “I’m happy … to pool our data together” are you talking about your own review service, or OSM?
If OSM, we will need a waiver from them to use their CC BY 4.0 data, & they would need to agree to release OSM data under the ODbL.
can you explain why someone should create a third solution instead of joining one of these existing?
I’ve juste added the ODbL license to all reviews published via https://cartes.app, thanks for talking about this important matter.
Also, this page lets you see all published comments, freely available. Will be downloadable and an API will be documented.
(Cartes.app is still French-only, sorry we have to finish the v1 before i18ning)
Wikimapia already solved the “review” problem, as any user can add comments to any POI object. It has been up and running since 2006.
No offence but they arguably haven’t “solved” the problem if very few people know about them, and very few people contribute.
The main problem to solve is in my view a social one, not a technical one: get enough people to contribute for the data to be useful. This might need more active reviewers than there are active mappers: in many places around the world, we can barely keep up with which POIs exist and which ones have closed down, nevermind how happy people are with them.
Don’t get me wrong, I applaud these efforts and I don’t want to discourage them, but I’d love to see concrete ideas for how to solve that social problem of motivating people to sign up and review, including people who don’t even know what OSM is.
is there way to find say highly rated eateries in a given area?
if yes how it can be done?
Zoom in to the target area, click on “Categories,” then “Restaurant,” and manually click each entry shown on the map one by one.
There’s no explicit five-star rating system here; all you can do is submit text and photos.
In some places, its user base is still active and strong, so it still generates some organic mapping and review data. I occasionally come here to get a “second opinion” against reviews on Google Maps.
I didn’t down vote this, but I would point out that you are promoting a long time and dubious competitor to OSM on an OSM platform.
It is quite possible to construct permanent links to OSM objects, as in links to specific versions. It needs some hackery to make this work for ways and relations so that you can detect all location changes, but it is completely possible. IMHO so generated links are not really fit for the general public, but for a service that is closely coupled to OSM, why not. It has the added advantage that it is trivial to detect changes and potentially semi-automatically update such references.
That’s putting it a bit strong; on the site currently I see “This page cannot load this map properly” and “for development purposes only”. The only POIs that I see are a school and someone’s house that they have added themselves…
Not at all, OK it would have been more correct to say “a now defunct competitor to OSM” :-).
All I see on my end is haphazardly (!!!) traced municipality boundaries and only a few major (!!!) roads.
Can you tell its “community” is nonexistent compared to OSM’s?
The difference is that OSM (the project, not the data) is not meant to be directly used by the end consumers who would be writing reviews.
I kind of thought that @SomeoneElse could remember, but ok Wikimapia - OpenStreetMap Wiki @spughetti seems to be of the opinion that the OSM wiki is a branch of wikipedia. It isn’t and we are f******* allowed to promote our project as much as we want.
Since lib.reviews has been mentioned, a couple of updates:
Aside from better OSM integration and easier invite-requests, some other things on my own wishlist include:
Lots to do! If you want to help, the main repo is on GitHub. Our IRC channel is #lib.reviews on libera.chat.
