Hacker News Reader

Comment by est31 | Hacker News Reader

Most of the proprietary Google apps on Android are actually just frontends to their proprietary network services like play, maps, gmail, gms, firebase, etc. Giving a replacement that deserves the name means that you have to re-create the associated network services which is the actually hard part because you won't just get shop owners to upload their opening hours to your service as well as google maps. That's an immense competitive moat that Google has built. It's on the territory where they are strongest, providing network services. And it's part of the reason why they could open source the Android OS in the first place: no competitor manages to replicate all of those network services.

View on HN

> You won't just get shop owners to upload their opening hours to your service as well as google maps

Over the last couple of years, I have noticed some shop owners (mainly chains, but also individual proprietors) adding their opening hours to OpenStreetMap themselves. Certainly more needs to be done to compete with Google, but OSM is still a libre alternative with some level of uptake. Plus, anyone walking past an establishment can add its opening hours from what the sign on the door says.

I add opening hours to establishments I frequent. Since next time, I can easily answer the question "Is X open right now?" without needing to be online to do so.

OSM is fantastic. At least if you live in an area where people maintain it. The level of detail that OSM has on Artis, the Amsterdam Zoo, is astounding. On Google Maps it's hard to see which parts are inside and which are outside the zoo. Google Maps also very reliably sucks on bicycle routes in and around Amsterdam.

Is there any effort to define a customer-facing standard for information like this?

    /.well-known/location_list
    /.well-known/location/<code>/open_hours
    /.well-known/location/<code>/phone
    /.well-known/location/<code>/<other_metadata>

etc? Or even on-page microformats. Google used to do this until they got big enough that they didn't have to anymore. This would probably be the only way we could chip away at that moat, and ensure we don't fall into that trap again.

Yes. It was called the semantic web, but it never took off.

The semantic web failed due to HTML being a format for humans. Where humans need to enter for them unnecessary extra data so the machines could understand what the humans already understood. There should be a (URL maybe?) scheme for machines generated by machines with standard properties like open hours, site category, etc.

There's still some attempt being made, for example with json-ld[1].

Ultimately any kind of microdata has the same problems that OpenGraph or ye olde meta keywords etc. tags did, GOOG and Co. need to parse the human-readable content to confirm any extra data for bots is not a trick. Which tends to encourage them to just ignore it, at best it's a slightly more structured version of what they're scraping from the page, at worst it's a lie.

1: https://json-ld.org/

There is one: https://schema.org/

It's honored by Google (it can enhance your search visibility) and that's why it's slowly getting popular. Far from wide adoption, though.

There are other reasons why an open source implementation is useful even if it relies on some proprietary backend stuff (like being able to review the code running directly on your device).

And I think it's more interesting in the context of the recent Huewai/Android story.

The other moat is that Google will blacklist you if you try to scrap public domain information like operating hours from google maps.

And with good reason. While singletons of those are indeed trivial and not protected, once collected into a huge database, that database earns protection.

I don't agree. There should be no copyright claim or other legal protection status possible on public domain information. That is not the same as saying Google should be forced to hand it over. I'm talking about crawling their website.

Reply on news.ycombinator.com