Mihai Budiu

In this case it's probably easiest if you start with the text input/output. Again, the simple.dl program is a good example. The dat file is a text file with DDlog...

The manual still does not explain how the evaluation of expressions inside interpolated strings is made.

`Error: "unknown constructor Fri of type \'evaluation::adt_enum::souffle::TDay\' in PosStruct(\"Fri\", [])"` While parsing a dat file.

BTW: this error message should also probably have a line number.

Where should we publish the jar? We actually have a 1-line script in the sql directory which installs the ddlogapi.jar in the standard maven loca repository; it is called ./install-ddlog-jar.sh....

This is an open-source project, we cannot have here code that is specific to internal company repositories.

I don't see what we can do about this. DDlog consumers should implement this functionality.

What is the process you use to deploy a new version of ddlog? I believe that installing the jar in a repository should be part of that process.

That's exactly the kind of proof we should be able to make once we have a calculus.

It seems to treat the multiset as the distinct version of the multiset.