project & script ids shouldn't be numeric - should be random hashes

[johrstrom]

Using numeric ids can be troublesome because they're based on the past so they have that dependency on time. What's worse is when you delete a project and create a new one, you're back to the same ID when we've likely modified the filesystem already for that id.

So - instead of coupling the current project/script with the past scripts, let's just use random strings (maybe 6 characters) for IDs instead of counting integers.