3 Replies to “Off the Beaten Nail”

  1. Sounds like a very cool workshop…but at theory-heavy POPL? Perhaps we could organize something like this at PLDI or ECOOP!

  2. It is a great idea, but it doesn’t seem like it takes itself seriously. There are no papers, just short summaries of a problem and an invited talk.

    I have a great deal of interest in the history of language design and various attempts to critique language quality. For example, the Shalishan benchmark for parallel computing. What ever happened to it? It seems like the only clearly defined benchmark that has stood the test of time is writing “99 bottles of beer on the wall” in your language in the fewest lines of code possible.

  3. By the way, what problems do you think are worth pursuing beyond those in the Subtext manifesto?

    My big interests: automatic synthesis of communication protocols from code. Rather than having to use a library that makes arbitrary decisions on things like message order, message format etc. it seems like it would be far more beneficial for clients and servers to negotiate these details on the fly based on: (a) what the client would prefer the most (b) what the server can actually reasonably provide (c) load (d) other things.

    This isn’t that wild of a suggestion, since it is basically how humans operate when building web-scale services. I just wish we could automate most of it so that we don’t have dependencies on cruft like “Windows Communication Foundation”, “SOAP”, WS-*”, “REST”, etc. We should be able to programmatically reason about every trade-off in Fielding’s derivation matrix, and more.

    Some language researchers, such as Cook and Kiselyov, have separately created languages where you talk to all systems using one model, but it is not clear to me this is the right answer. Then there is a whole swarm of killer apps that can derive from such a flexible substrate. For example, applications like Tableau Software’s database navigation software.

Comments are closed.