Vincent,

Thanks for chiming in!

On 2024-11-14 18:08 -0800, Vincent Manis (he/him) wrote:
> I agree that the SRFI should be finalized. I'm less enthusiastic about
> restarters as a desirable language paradigm. My skepticism is based upon my
> experience, decades ago, with PL/I, whose exception-handling facilities had
> automatic retry defaults in many cases. As a result, programs that did not
> correct errors could easily run into infinite loops.

That does sound painful, and it’s one reason SRFI 255 doesn’t recommend
a default restarter discipline for any Scheme forms.

> It would be a good idea for people enthusiastic about restarters to try them
> in real programs. If in fact, they make programs more robust and readable,
> then we should consider them for inclusion in the Large language.

I agree. I wouldn’t suggest including them in R7RS until they’ve
seen more use.

The main contribution of SRFI 255, I think, is that it integrates
restarters into the R[67]RS exception system. The main text of SRFI
249, on the other hand, predates that system, so it’s not as good a
fit for “modern” Scheme as the new restarters-as-conditions model.

--
Wolfgang Corcoran-Mathe  <xxxxxx@sigwinch.xyz>