Thoughts on Open(Geo|Place)Reviews


Hi, regular OSM contributor and OsmAnd user here :slight_smile: Glad to have found this forum in my search for news on this project. Having gone through all the discussions about this project so far, I would like to share some thoughts.

My personal use case for place reviews is almost exclusively restricted to eateries (restaurants, cafes, bars). I would like to see OPR become a viable alternative to services like Zomato, which have more or less become the Facebook of this field (official websites replaced by Facebook/Zomato pages, and a walled-garden approach - in case of Facebook, inability to view content without registering, and in Zomato’s, an extremely aggressive push for their proprietary mobile app, resulting in e.g. a thoroughly unusable mobile website, which seems to exist for the single purpose of getting people to install it.)

Therefore, my first two points are for the use case of eateries -

  1. Along with reviews and ratings, I would like to see menu and pricing information, too. It is of immense utility when deciding on a restaurant - at a single glance,
    • a high rating [1] tells you that the restaurant is generally well-liked (without having you wade through reviews)
    • an approximate “price for two (with/without alcohol)” tells you if it is within your budget (without having to go through the menu)
    • and the menu itself gives you complete information of what to expect there and at what price (without having to go to the place itself)

It is, of course, possible that such information be added to OSM instead. I have discussed it in the IRC channel once, and found people agreeable to the idea, although expressing concerns of maintainability.

  1. Perhaps we can split a rating into different components? Hygiene, ambience, food, and staff could be rated individually, and a final score derived from that.

  2. I happened to read about the blockchain being used in the architecture. As contributors to FOSS and open data, we desire (…I hope this isn’t just me :smiley: ) to make efforts to improve the world. In the current situation of the world, with anthropogenic climate change already bearing its poisonous fruit, I strongly object to the use of blockchain, with its wasteful PoW computation, in this project.

[1] It needs to be adjusted for number of ratings, of course - a single rating of 5 shouldn’t rank higher than ten ratings of 4.5


Additionally, a friend of mine - when I informed them about this project - made a suggestion which I like more and more…that of rating individual dishes of a restaurant. One’s experience at a restaurant can vary wildly depending on what one orders. People often do this informally in their reviews anyway, but this can let us (as above) skip the review-reading and get to the conclusions - “people like place X for their Y and Z.”. This can extend to non-eateries too - replace dishes with services.

To avoid duplication and inconsistency, we can use dish names from the menu (volunteers would have to take up the task of obtaining the menu and adding the dish names and prices to the database, ideally with a wiki-style workflow) which would then be selected by the reviewers.

We can also encourage a structure for the reviews, based on per-aspect rating (described above) - a reviewer rates each aspect (hygiene, ambience, staff, food) and writes their remarks on it; each aspect gets a section; and the food section lists the dishes they had. This can give a common structure to all reviews and hopefully prevent runaway, rambling reviews.