On 2024-11-14 21:52 +0100, Daphne Preston-Kendal wrote:
> The only comment I have is to echo one sent to me when I solicited
> some outside opinions on this SRFI: the design has changed so much
> from Common Lisp and MIT Scheme that we need significant experience
> testing the new design.

I’d say the same thing about every SRFI I’ve worked on, since they’ve
all ended up being novel in one way or another.

The only thing I can do right now to remedy this is to see how SRFI
255 works in my own Scheme programs. But this may not help. In my
experiences with library design, “it works for me!” never convinces
anybody. If someone wants to volunteer to acquire “significant
experience” with SRFI 255, of course, I’d be grateful for their
opinions.

> One false assumption that set this SRFI off in the wrong direction
> early on was that restarts are primarily a tool for debugging, and
> not for e.g. building application programs with features allowing
> users to correct incorrect inputs.

I don’t see any evidence of this assumption in the current SRFI.
The only occurrence of the word “debugging” is in a sentence (copied
from SRFI 249) about Common Lisp’s restart system. Furthermore, the
‘restartable’ forms (which have been present since the first draft)
were added specifically to allow users to correct incorrect inputs.

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