Re: Scheme Review vs. SRFIs Lassi Kortela 04 Dec 2022 11:07 UTC

> No. The SRFI process is both effective and useful, and has been for ages. If it ain't broke, don't fix it.
>
> Discussion between_implementors_  remains strong. Discussion between end-users also remains strong - each implementation has mailing lists and IRC channels, and there's generally a reasonable level of activity.
>
> The major weakness of the language as a whole is the lack of consistent networking, which is the primary factor concerning portability of modern end-user code. This is a non-trivial thing, partially because the historical Scheme IO abstraction - that is, ports - is not exactly conceptually compatible with sockets. Discussion on this has been going on for some time and will hopefully result in some mutually-agreeable proposals soon.
>
> The Scheme ecosystem as a whole is fairly healthy. There are several major implementations, a low barrier to entry, and a general level of civility in discussion amongst very bright people working out technical problems in their free time. There are presentations on new research and new concepts in the Scheme world at every major programming languages conference. There are new things being done all the time, and new people seeing the beauty and elegance.
>
> That there are a few technical differences - which have NOT led to fracturing of the community as a whole, except for the R6 debacle - does not indicate a broken language or the need for an overhaul of the established systems, and, historically speaking at least, doing such is more likely to lead to fragmentation.
>
> That the R7-large process got bogged down does not indicate an overall deficiency - it does mean that the scope was perhaps too large and that certain things were overstressed, taxing people's copious free time - don't forget that most of us have real jobs and life outside of this and we're doing this on a volunteer basis.
>
> I've said it before, and I will say it again - the technical problem of networking is the ONLY major barrier between implementations. Given that there are schemes written in close to a dozen languages, targeting a half-dozen architectures and every major OS, this is a difficult - but not insurmountable - problem. This does not detract from the vast majority of code being portable and useful, and the established institutions serving their roles effectively.
>
> We've had about a year of the new committee structure - and I have seen absolutely no improvement in speed, efficiency or actionable output from what was before.
>
> This would be a good indication that perhaps a rethink regarding the "need" to change everything may be in order.

Most of this seems to be either misunderstandings or tangential to the
discussion at hand. I still don't see any major point of disagreement.