Would it be reasonable to suggest, as a part of R7RS Large process, to explicitly specify a superset of R7RS Small (we may informally call it R7RS Medium) having just enough extra primitives and mechanisms to allow for practical R7RS Large reference implementation to be distributed as a portable set of R7RS Medium-based libraries, with minimal cond-expand tricks?

This would allow small Scheme implementors to partake in R7RS Large game with reasonable additional effort.

-S

> On Sep 30, 2024, at 4:12 PM, Daphne Preston-Kendal <xxxxxx@nonceword.org> wrote:
>
> On 30 Sep 2024, at 21:26, Antero Mejr <xxxxxx@antr.me> wrote:
>
>> Daphne Preston-Kendal <xxxxxx@nonceword.org> writes:
>>
>>> I am happy to respond to constructive, actionable criticism of our processes,
>>> but ‘I don’t like it’ is neither constructive nor actionable.
>>
>> 1. It's uniquely annoying to write portable Scheme, because each
>>  implementation has inconsistent library support.
>>
>> 2. In order to find documentation for writing portable Scheme, you need
>>  to consult at least 3 different websites simultaneously (the
>>  standard, SRFIs, implementation).
>
> If only there were a standards effort attempting to provide a larger base for portable libraries and provide more accessible documentation for the base language.
>
>> 3. No reference implementations, so lots of duplication of work. The
>>  SRFI examples often are not portable (especially the older ones), and
>>  need to be put together piecemeal, because there is no standard for
>>  package management.
>>
>> 4. The largest program written in Scheme, the Guix package manager, does
>>  not have a way to install or (re)use R7RS libraries, ironically.
>
> I am not sure what you expect me to do about this.
>
>> 5. There's no Scheme Foundation, and no way to connect with other
>>  Schemers in a given geographic area, etc. Most of the different
>>  Scheme sub-communities are really just a few people, a webpage, and
>>  maybe a mailing list.
>
> I want a Scheme Foundation too, but the SC has rejected the idea as being too much bureaucracy for a small language community like ours. I’m not sure they’re wrong, honestly.
>
>> 6. The steering committee's activities seem completely opaque. What do
>>  they discuss? What decisions do they make? How will it affect the
>>  direction of the language? How is their membership decided? How
>>  active are they?
>
> The SC generally deals directly with me only, or with others if they want to ask them.
>
> What they discuss: Almost nothing. There were three SC meetings in total last year – two to deal with John Cowan’s resignation, and one for me to give them a status update. Prior to that, they had been inactive since 2014.
>
> How the membership was decided: There was an open, online election in 2009. Should someone currently on the SC resign, there will be a new election.
>
>> 7. As Marc mentioned, the difficulty in following what is going on
>>  where. The reason I didn't reply to your initial message for a few
>>  days is because I wasn't subscribed to srfi-discuss, and had no idea
>>  you responded.
>
> If only the person who started this thread had chosen some other forum than srfi-discuss.
>
>> 8. It's unclear who has the ability/authority to make any sort of
>>  change. Could you address any of the above, if you wanted to?
>>  Hypothetically, if most people agree that N=3 is not sufficient,
>>  would you have to get approval from the steering committee to change
>>  it? What would happen if they disapproved?
>
> N=3 is an informal rule of thumb, suggested by the SC when I became chair, and has no official status in the charter. You can read the charter here: <https://small.r7rs.org/wiki/WG2Charter/>
>
> In practice, I intend to get more strict about requiring implementation support and implementer engagement with particular features as the Foundations approach finalization.
>
>>> I am really not sure what we can do to make the process more ‘transparent’. But
>>> it’s especially hard to know when people apparently have ideas such as that ICFP
>>> is somehow part of the process.
>>
>> That is part of the transparency problem. I heard about ICFP from the
>> WG2 mailing list. Hosting meetings at ICFP is inaccessible and
>> problematic: the price and logistics of attending such a conference
>> keeps most Scheme users out. Unsure if that is the intention.
>
> My point is that ICFP is *not* part of the process. We don’t make decisions in meetings at ICFP. I have no idea where you got this impression, and I have no idea how you interpreted my previous mail as saying that we do.
>
>> And the WG2 mailing list appears to only be available on Google Groups,
>> disjointed from any of the other Scheme sites.
>
> Google Groups hosts that mailing list, yes. But it is a normal email mailing list, and can be subscribed to without a Google Account. I did not make the decision to use Google Groups but it can’t easily be changed.
>
>> It doesn't have to be this way. If a Large reference library were
>> provided, then Small implemetations would get Large for free. However, I
>> think Scheme currently lacks the coordination to create and maintain
>> such a library. I would be happy to be proved wrong on that though!
>
> If only there were a standards effort attempting to provide a reasonable standard library with sample implementations.
>
>
> Daphne
>