A request for Meteor Group.. Adopting meteor-kitchen

Meteor-Kitchen is awesome tool that helps cut the time and efforts for creating new Meteor projects. It’s fantastic how a little json file can turn into a fully functional project!
But it looks like it’s not maintained anymore… I think MG should adopt Meteor-kitchen and make an official tool.

Any thoughts on this?

Hi @MJO can you post some links for reference?

The site itself looks more like something that the community could take over. as for the actual code that can either also go to the community, but I think we could discuss a merit of it being integrated into the core codebase or something along that style.

1 Like

I helped Petar Korponaic some time ago with Meteor Kitchen, and I was always on at him to make it open source. At the time he was reluctant to do it, as he had some opportunities to commercialise it. Those are probably long gone now.
He said there was a fair amount of C++ code which he used in the code generation, and he said he wasn’t proud of that code. I did offer to help port it to Javascript, but nothing ever came of it. His interests are now in quantum computing.

Some of his code is open source. He even wrote a blaze to react convertor (and vice versa).

His repos are here: https://github.com/perak

We could approach him now (if there is interest in maintaining it) to ask if he would open source it.

Are code generators worthwhile?

Meteor Kitchen is a code generator, which saves from writing a lot of boiler plate code, and lets you get an app up and running quickly. There are other generators which have utility (create-react-app and various yeoman generators).

The biggest challenge with them is that they are basically one-shot, ie you generate your code, and then you maintain it by hand after that. If you improve the code generator, doing an upgrade isn’t always that easy.

A great solution would be to be able to keep the generated code separate from the custom code, so that you can get the benefit of both worlds. This takes a lot of effort.

Personally I have moved away from code generators in favour of writing more generic code that is data driven. Ultimately this is a more flexible arrangement, if you can structure it to allow plugins and custom code. I am thinking this can be the subject of my Meteor Impact talk this year.


I agree with you.

I believe we need to create better packages, easy to use, plug & play instead of code generators.

@MJO what are the use cases that you believe code generators would be than packages?