merge with paulp/policy

[jdegoes]

A community-maintained fork of the Scala compiler will have substantially greater chances of success if the entire community of developers looking for a "better Scala" is united around a single compiler.

To that end, I suggest the rather obvious: that typelevel/scala and paulp/policy should be one and the same project. Both are already philosophically united around the idea of a "better, simpler Scala" and differ primarily in requirements for backward compatibility, and in my opinion, typelevel's requirements need revamping.

Moreover, the current community-driven process of development on typelevel/scala will run into issues wherever there are substantial disagreements in the proposed direction of the compiler. This is why most successful languages have a "BDFL". A community-maintained fork of the Scala compiler needs a BDFL and the person most qualified for that role is @paulp.

So there you have it, my 2 cents on uniting behind a single community-maintained fork of Scalac.

[paulp]

I admit that I'd rather we were all working on the same thing, and @jdegoes astutely if only implicitly identifies the major reason we aren't. I would rather rule in hell than serve in heaven, as the saying goes. I served in heaven for quite a while and it wasn't that heavenly.

And if it doesn't seem like a good idea at this point (which I would completely understand) it might not take too long before it does, because I'm not sitting on my thumbs.

[propensive]

We spent considerable time prior to the announcement looking at whether Typelevel Scala and Policy could work together, but Typelevel's goals are actually more aligned to Typesafe's than @paulp's, for now at least.

One problem we have is that we're inexperienced and unproven, and we're trying to produce something of a sufficiently high standard that people will rely on it enough to use it in production. The more ambitious we are, the less likely we are to achieve that goal. We've had to take the conservative route, because we believe it has the greater chance of success. Once we all become experts in Scalac development, we can reevaluate our ambitions and goals.

Regarding having a BDFL, we decided months ago that we didn't want that; not @paulp, not @milessabin, not anyone. That may be detrimental at times, but it's a community fork, and we really want to keep all discussion and decisions open and public, not within a committee, or within one person's head.

I'll keep the issue open for now, because I think it's a worthwhile debate to be having - thanks for raising it, @jdegoes!

[milessabin]

+1 to the debate ... this is the wrong place for it though ... it should be on the typelevel mailing list.