Early Direction

Zack is ambitious, we make things easier for us at first by making some simplifications.

Target Linux first

Zack will integrate with the OS to offer all its features. Doing this in a cross platform way, is difficult.

As bandaid, we might offer something like easy execution via docker and/or just use a devcontainer.

Sandboxing does not have to be strict

Zack is not primarly security focussed, so a pragmatic approach to sandboxing is fine.

Restricted set of programming languages

An obvious intermediate milestone would be to self-build Zack. On that way, we could create more simple examples with Rust.

Targeting reasonably sized monorepos

Zack should be fast for typical monorepos, that is the point.

But rather than aiming at Google scale, we contend ourselves if Zack works well for monorepos for communities / companies with <= 100 regular contributors. Another way to think about it is that

• the source should still fit comfortably on a developer laptop
• a full rebuild is possible on a developer laptop in a few hours