My understanding is that just "using Realm Object Server" means using Realm Platform Self Hosted, which starts at $2000/year. The hosted Realm Platform starts at $30 month, but I've read you need to upgrade to a similarly expensive $1500-2000 enterprise plan if you want to access the data from non-mobile-Realm clients. That is to say, I'm very open to just using Realm Database on the client for local persistence, because it's free and really nice to use, but don't think Realm for sync is viable.
AWS AppSync looks like it's worth investigating, thanks!
Agreed a custom data layer/cache might be the way to go, I think it gets a bit more complex once you have a bunch of objects being edited and deleted, and I've also spent some time with iOS background networking in the past and I know it can be buggy/frustratingly non-deterministic, which is why I'm reluctant.
I'm working on a very similar market for software development, where reward is granted based on completion of a unit test. Check it out at www.codetract.com
The search identifies venues with dishes that match your query and calorie criteria, and then sorts those venues based on distance, which is why McDonald's and Burger King will often show up first. Once we're able to support other search parameters (e.g., putting a limit on the grams of fat or the % of calories contributed by fat), the fast food results should filter out - or at least he "pinkslime" ones :). Another feature on the roadmap is providing some summary (e.g., a grade or score) of the nutritional value of a menu item - this score would factor in things beyond calories like macro-nutrient breakdowns and might help prevent "misleading" results.
In the meantime, the hope is that calories are useful as a jumping off point when it comes to discovering healthy dishes nearby, and as I mentioned before you can see the full nutritional info by clicking on a particular item.
Thanks for the feedback! Totally agree that calories != nutritional value.
Farther out, the plan is to allow users to conduct structured searches along all dimensions of menu items - macro-nutrients, sodium, etc. Due to the limitations of one of the APIs I'm using, I can only support a calorie parameter in queries for now. Note that once you drill into a particular menu item, you are provided with in-depth nutritional data like carbohydrates, protein, and fat.
Thanks for pointing this out! There was an issue with point total calculation that has been fixed, and all users' totals have been updated appropriately.
One thing on my "to-do" list is adding a test suite builder, that allows you to specify inputs and outputs and then generates the test code for you. That would allow people who can't code or just aren't familiar with the syntax of a particular language/framework to still post contracts.
OOC, why do you say "as a software engineer, I know these queries won't work"? What is it about those queries that leads the engineer in you to believe they are hard (not to say that they aren't, it just seems to me that Google provides remarkable results for many objectively tougher queries).