Mangrove.reviews support #1964

gimse commented 2026年04月03日 05:46:59 +02:00

Décrivez votre problème, votre suggestion

Hi. It would be really awesome it the apper showed reviews of places from https://mangrove.reviews/

The project has not many reviews currently, but it is growing. Mapcomplete have added support for it, and more are coming. It thing this is a better solution compared to BlueSky

Système d'exploitation

No response

Navigateur

No response

### Décrivez votre problème, votre suggestion Hi. It would be really awesome it the apper showed reviews of places from https://mangrove.reviews/ The project has not many reviews currently, but it is growing. Mapcomplete have added support for it, and more are coming. It thing this is a better solution compared to BlueSky ### Système d'exploitation _No response_ ### Navigateur _No response_

yoyo34 commented 2026年04月04日 03:23:53 +02:00

Well, if I'm allowed to put in my two cents, I'd say that Mangrove is still very niche. And in general reviews are a sensitive issue, because the potential of abuse and spam is very high and they have to be costly moderated.

Well, if I'm allowed to put in my two cents, I'd say that Mangrove is still very niche. And in general reviews are a sensitive issue, because the potential of abuse and spam is very high and they have to be costly moderated.

gimse commented 2026年04月04日 08:48:19 +02:00

I understand @yoyo34 :)

Maybe in the beginning it would be possible to send reviews to mangrove.review, but show reviews from bluesky? Or a weighted average of blusky and mangrove?

Just trowing out ideas :)

I understand @yoyo34 :) Maybe in the beginning it would be possible to send reviews to mangrove.review, but show reviews from bluesky? Or a weighted average of blusky and mangrove? Just trowing out ideas :)

laem commented 2026年04月04日 17:43:37 +02:00

and more are coming. It thing this is a better solution compared to BlueSky

Why ? Lots of people told us this, but did not give us a compelling argument to go for a niche network instead of a growing decentralised network of 40 million users. See this initial comparison here.

Atproto/Bluesky is built for commenting things. Why build something on its side ?

OpenStreetMap is built for making a common database of geographical things : why build a small database on the side ?

> and more are coming. It thing this is a better solution compared to BlueSky Why ? Lots of people told us this, but did not give us a compelling argument to go for a niche network instead of a growing decentralised network of 40 million users. See this [initial comparison here](https://codeberg.org/cartes/web/issues/767#issuecomment-5892802). Atproto/Bluesky is built for commenting things. Why build something on its side ? OpenStreetMap is built for making a common database of geographical things : why build a small database on the side ?

hlrx commented 2026年04月14日 17:29:32 +02:00

What about https://lib.reviews/ ?

What about https://lib.reviews/ ?

laem commented 2026年04月15日 12:07:43 +02:00

Same reasoning as Mangrove. It's just a new db struggling to emerge :/

Same reasoning as Mangrove. It's just a new db struggling to emerge :/

laem commented 2026年04月16日 10:02:09 +02:00

Also, note that the current review user experience is not good enough. E.g. photo uploading. Let's focus on providing a good UX on the current solution for now, and then add providers.

Well, that's what I have in mind. This is an open source project of course, so anybody can propose the features they like :)

Also, note that the current review user experience is not good enough. E.g. photo uploading. Let's focus on providing a good UX on the current solution for now, and then add providers. Well, that's what I have in mind. This is an open source project of course, so anybody can propose the features they like :)

contrapunctus commented 2026年06月16日 10:15:06 +02:00

@laem Bluesky is a centralized network which is cosplaying decentralization. I was really disappointed to see Cartes.app fall for the bait and suggest ATProto as a preferable alternative to independent, community-run open data review platforms.

The only reason Mangrove Reviews and Libreviews are "struggling to emerge" is because more apps need to integrate them. It's a chicken-and-egg problem.

Meanwhile, as someone who advocates for people to switch to libre software and OpenStreetMap, I have to recommend a whole bunch of different apps to replace Google Maps -

1. CoMaps/OsmAnd for offline maps,
2. Cartes.app for online maps, indoor maps, and satellite imagery, and
3. MapComplete for ratings and reviews (via Mangrove Reviews).

If cartes.app added support for Mangrove Reviews, at least one of those applications could be removed.

Also, I believe Mangrove Reviews now has support for attaching photos to reviews.

@laem Bluesky is a centralized network which is [cosplaying decentralization](https://rys.io/en/167.html). I was really disappointed to see Cartes.app fall for the bait and suggest ATProto as a preferable alternative to independent, community-run open data review platforms. The only reason Mangrove Reviews and Libreviews are "struggling to emerge" is because more apps need to integrate them. It's a chicken-and-egg problem. Meanwhile, as someone who advocates for people to switch to libre software and OpenStreetMap, I have to recommend a whole bunch of different apps to replace Google Maps - 1. CoMaps/OsmAnd for offline maps, 2. Cartes.app for online maps, indoor maps, and satellite imagery, and 3. MapComplete for ratings and reviews (via Mangrove Reviews). If cartes.app added support for Mangrove Reviews, at least one of those applications could be removed. Also, I believe Mangrove Reviews now has support for attaching photos to reviews.

laem commented 2026年06月16日 10:49:32 +02:00

@contrapunctus I'm not sure its relevant to give us a link to an article written in 2023 just some days after Eurosky published mu.social and a full replication of the Bluesky infrastructure, making most of the Fediverse's claims about Bluesky obsolete. This should be celebrated. I know the Fediverse (or more precisely, the Mastoverse) well enough to know that the news will be at best silenced, at worst criticised as being another fake decentralisation.

I don't want to waste my time on these petty squables.

Some people want "full" decentralisation. Some do not. It's OK, the latter shouldn't be received as such an offense by the former.

As for Mangrove, we don't think adding a new independent database dedicated to reviews is a good idea to gain traction. We believe existing commenting platforms should be used. ATproto and the Fediverse are commenting platforms.

Mangrove decided to create something anew. We respect this choice, and will integrate Mangrove if it gains significant traction. For now, the website barely works. I don't see any "place" result when I type "restaurants". I don't see any map. The last blog post was 6 years ago. Meanwhile, the ATproto ecosystem is thriving and enabling well more than reviews, it's a whole user-owned database.

Integrating Mangrove without complexifying the interface is a hard task. We're not prioritising this, too many obvious things are lacking (like offline maps, the 1 on your list, very relevant).

Commenting places on cartes.app using Atproto is possible and works well. We're proud of it. The open source community should celebrate it and relay the news instead of shooting its own feet because of turf wars :/

@contrapunctus I'm not sure its relevant to give us a link to an article written in 2023 just some days after Eurosky published mu.social and a full replication of the Bluesky infrastructure, making most of the Fediverse's claims about Bluesky obsolete. This should be celebrated. I know the Fediverse (or more precisely, the Mastoverse) well enough to know that the news will be at best silenced, at worst criticised as being another fake decentralisation. I don't want to waste my time on these petty squables. Some people want "full" decentralisation. Some do not. It's OK, the latter shouldn't be received as such an offense by the former. As for Mangrove, we don't think adding a new independent database dedicated to reviews is a good idea to gain traction. We believe existing commenting platforms should be used. ATproto and the Fediverse are commenting platforms. Mangrove decided to create something anew. We respect this choice, and will integrate Mangrove if it gains significant traction. For now, the website barely works. I don't see any "place" result when I type "restaurants". I don't see any map. The last blog post was 6 years ago. Meanwhile, the ATproto ecosystem is thriving and enabling well more than reviews, it's a whole user-owned database. Integrating Mangrove without complexifying the interface is a hard task. We're not prioritising this, too many obvious things are lacking (like offline maps, the 1 on your list, very relevant). Commenting places on cartes.app using Atproto is possible and works well. We're proud of it. The open source community should celebrate it and relay the news instead of shooting its own feet because of turf wars :/

contrapunctus commented 2026年06月18日 07:53:24 +02:00

@laem I am extremely disappointed by your attitude towards decentralization.

The question with protocols like Matrix and ATProto is not whether decentralization is "possible", but "practical". The claims will never be "obsolete" as long as the protocols remain the way the are.

These are not "turf wars" or "petty squabbles". It's the age-old question of decentralization and user/data independence. You'd think people would learn after Reddit, Twitter, et al, but I guess with enough VC funding it's possible to fool even technical people.

IME the best way to evaluate Mangrove Reviews is not the website (analogous to OSM and osm.org?), but MapComplete.org. It works well enough for me to regularly recommend it in OSM communities and the audiences of my OSM talks, and to use it myself.

I really can't understand what Cartes.app stands to gain from using ATProto instead of Mangrove Reviews/Libreviews. Is there an existing dataset of reviews present on ATProto that Mangrove Reviews/Libreviews do not have? The potential issue of "a separate service needs a separate account" does not even exist for Mangrove Reviews, since "account creation" is done transparently and the user doesn't even need to be aware of it.

Finally - is there a way for the ratings/reviews created through Cartes.app to be shared as a single dataset? If so, what license is it released under? And is it possible for other applications to query this database?

@laem I am extremely disappointed by your attitude towards decentralization. The question with protocols like Matrix and ATProto is not whether decentralization is "possible", but "practical". The claims will never be "obsolete" as long as the protocols remain the way the are. These are not "turf wars" or "petty squabbles". It's the age-old question of decentralization and user/data independence. You'd think people would learn after Reddit, Twitter, et al, but I guess with enough VC funding it's possible to fool even technical people. IME the best way to evaluate Mangrove Reviews is not the website (analogous to OSM and osm.org?), but MapComplete.org. It works well enough for me to regularly recommend it in OSM communities and the audiences of my OSM talks, and to use it myself. I really can't understand what Cartes.app stands to gain from using ATProto instead of Mangrove Reviews/Libreviews. Is there an existing dataset of reviews present on ATProto that Mangrove Reviews/Libreviews do not have? The potential issue of "a separate service needs a separate account" does not even exist for Mangrove Reviews, since "account creation" is done transparently and the user doesn't even need to be aware of it. Finally - is there a way for the ratings/reviews created through Cartes.app to be shared as a single dataset? If so, what license is it released under? And is it possible for other applications to query this database?

laem commented 2026年06月18日 11:25:01 +02:00

I won't comment the subject of decentralisation anymore, sorry. I'll just say that OpenStreetMap works, despite being totally centralised. It shouldn't be an objective that takes priority on all others. We've got 1000 other priorities.

IME the best way to evaluate Mangrove Reviews is not the website (analogous to OSM and osm.org?)

Yes and it's an error I believe. We know mapcomplete.org, but 90 % of people won't and will think the project is dead. I know it's hard to maintain such a website, of course. Well that's why on our side we don't want to maintain a proper review DB, nor its website, nor do the enormous communication effort needed to impose it as a successful DB : we want to use a protocol and (decentralised) DB that covers the use case with a 40 million user base and that has the particularity to let users own their data.

Again, as I told you above, if Mangrove had a dynamic community of reviewers, we would have considered the option, because even if ATproto is superior technically from our perspective, being able to show existing reviews to our users would be appreciated.

The particular thing about reviews is that they're quickly obsolete. Lots of places's owners change every 3 years, so a DB with a lot of records from 2010 is mostly obsolete. What's important is the number of active users on the platform.

since "account creation" is done transparently and the user doesn't even need to be aware of it.

That's a good point, but unfortunately, the OpenStreetMap user base, despite being quite numerous, is not active enough nor general enough (lambda people, not tech people) to make a review DB thrive. Note that we're building cartes.app on top of OSM, both the data and the oauth system, so we're happy to use it, but it's not sufficient for reviews.

If you really can't understand that we're appealed by the opportunity to use an existing, well maintained, well documented, widely used (hundreds of applications already) JSON database owned by the user and that can migrated in one click from a server (PDS) to another keeping the user's identity and records, I'm sorry we're not on the same wavelength and it will be hard to reconcile our views. ATProto is a brick that the Web has been needing since its beginning. We're very excited to use it.

On our side, we understand the idea behind mangrove reviews. In 2020, ATProto did not exist.

Finally - is there a way for the ratings/reviews created through Cartes.app to be shared as a single dataset?

Yes ! The full DB is here https://serveur.cartes.app/deno/records
The (awful) UI to read it is here https://cartes.app/bluesky

The licence used is ODBL. Sorry, the terms and conditions of our review system is only in French for now :/

https://cartes.app/blog/conditions-utilisation-avis

And is it possible for other applications to query this database?

Of course, it's not our database, it's in the atmosphere ! Any app.cartes.comment JSON bit are stored in the ATProto users's PDS, publicily available, explorable under any PDS explorer.

I won't comment the subject of decentralisation anymore, sorry. I'll just say that OpenStreetMap works, despite being totally centralised. It shouldn't be an objective that takes priority on all others. We've got 1000 other priorities. > IME the best way to evaluate Mangrove Reviews is not the website (analogous to OSM and osm.org?) Yes and it's an error I believe. We know mapcomplete.org, but 90 % of people won't and will think the project is dead. I know it's hard to maintain such a website, of course. Well that's why on our side we don't want to maintain a proper review DB, nor its website, nor do the enormous communication effort needed to impose it as a successful DB : we want to use a protocol and (decentralised) DB that covers the use case with a 40 million user base and that has the particularity to let users own their data. Again, as I told you above, if Mangrove had a dynamic community of reviewers, we would have considered the option, because even if ATproto is superior technically from our perspective, being able to show existing reviews to our users would be appreciated. The particular thing about reviews is that they're quickly obsolete. Lots of places's owners change every 3 years, so a DB with a lot of records from 2010 is mostly obsolete. What's important is the number of active users on the platform. > since "account creation" is done transparently and the user doesn't even need to be aware of it. That's a good point, but unfortunately, the OpenStreetMap user base, despite being quite numerous, is not active enough nor general enough (lambda people, not tech people) to make a review DB thrive. Note that we're building cartes.app on top of OSM, both the data *and* the oauth system, so we're happy to use it, but it's not sufficient for reviews. If you really can't understand that we're appealed by the opportunity to use an existing, well maintained, well documented, widely used (hundreds of applications already) JSON database owned by the user and that can migrated in one click from a server (PDS) to another keeping the user's identity and records, I'm sorry we're not on the same wavelength and it will be hard to reconcile our views. ATProto is a brick that the Web has been needing since its beginning. We're very excited to use it. On our side, we understand the idea behind mangrove reviews. In 2020, ATProto did not exist. > Finally - is there a way for the ratings/reviews created through Cartes.app to be shared as a single dataset? Yes ! The full DB is here https://serveur.cartes.app/deno/records The (awful) UI to read it is here https://cartes.app/bluesky The licence used is ODBL. Sorry, the terms and conditions of our review system is only in French for now :/ https://cartes.app/blog/conditions-utilisation-avis > And is it possible for other applications to query this database? Of course, it's not our database, it's in the atmosphere ! Any app.cartes.comment JSON bit are stored in the ATProto users's PDS, publicily available, explorable under any PDS explorer.