[gedrap]

In my university (Manchester, UK) there's a whole year course in the first year when you are assigned to a group of people and you have to make a website, anything you can come up with.

Every year, someone makes a web site for searching recipes by ingredients :)

To be more on topic, I like the idea but I think implementation (i.e. manual work required) is a bit too... demanding, involved? We are in the world of 1 click. I would like an app which would allow easy input (scan bar codes?) combined with machine learning to discover new meals (taste, effort required). Google Now for food?

[Theodores]

The dreaded group project. So in your typical group of six you have a project manager, someone writing the functional specification, someone writing the technical specification, someone that does testing, someone that writes the manual, someone that parties rather than show up and nobody that codes!

I remember it well, five people waiting for someone to write some code before they could do their bit. Of course that code had to be in Ada rather than a vaguely useful language. Under such doomed-to-failure circumstances it is irrelevant what the project is, it might as well be 'Hello World'. So if some group does actually hand in code that allows for recipes to be searched for then I think that course is doing jolly well!

'Google Now' for food exists to a certain extent, if you do want to order pizza or get some other fast food in. I would like 'Google Now' for food that would factor in the fridge, the corner shop, the local super markets, their inventory levels and opening hours. It would be particularly cool if you could rustle up something amazing using just the local corner shop, perhaps someone should write one for the cash and carry company that all the corner shops use.

[buckbova]

Exactly. I've been pitched this same idea literally hundreds of times over more than a decade, most recently about a month ago.

And yes, it is a great idea, if the implementation were great.