meejah <mee...@leastauthority.com> writes:

> Stepping back a little, there's two main things here:
>
> 1. Great Black Swamp (GBS)
> 2. non-Python
>
> These new pieces of Haskell code are about both things, so I want to try to
> draw the lines a little more clearly.
>
> Without GBS, a non-Python implementation first has to make a Foolscap
> implementation.
> While Foolscap is interesting in its own right, it doesn't have any users
> beyond Tahoe and hasn't received much attention. It doesn't have an
> implementation in any other language.
>
> The main point of GBS is to remove that prerequisite by running "the same"
> Tahoe protocol (with the same privacy features) over HTTP -- on the premise
> that basically every interesting language these days has an HTTP client
> implementation.

Thanks.  This thread is the first I heard of GBS.   I completely agree
that making tahoe have a protocol spec and multiple implementations is a
huge positive step.

> The other part, "a non-Python implementation", is essentially running with
> GBS and proving that the above claims are true.
> I personally would love to see "even more" implementations of GBS in other
> languages.
> So, I think it's exciting that we now have a subset of Tahoe working in
> Haskell -- it might not be a language for everyone or _every_ system, but
> it does serve the purpose well of being "a GBS implementation" and "not
> Python".
> Additionally, it helps prove that the GBS specifications are "good enough"
> to achieve an inter-operating implementation in a new language.

Agreed that regardless of opinions or where-it-runs concerns, it is a
huge step forward for Tahoe to be getting a second implementation.  Two
implementations and a written spec is vast different from "one
implementation and the code is the spec" (even if that hasn't been
fair).


Sorry if I sounded more negative than I should have and failed to
separate "this implementation isn't something I can adopt into my world"
from "regardless of portability concerns, it's a huge step forward."

And, I'm glad to see separate libraries for the separate pieces.  I
think that kind of software separation is also helpful.
_______________________________________________
tahoe-dev mailing list
tahoe-dev@lists.tahoe-lafs.org
https://lists.tahoe-lafs.org/mailman/listinfo/tahoe-dev

Reply via email to