Email list hosting service & mailing list manager

Hate post on Reddit about SRFI Jakub T. Jankiewicz (10 Nov 2022 22:08 UTC)
Re: Hate post on Reddit about SRFI Daphne Preston-Kendal (10 Nov 2022 22:19 UTC)
Re: Hate post on Reddit about SRFI Jakub T. Jankiewicz (10 Nov 2022 22:29 UTC)
Re: Hate post on Reddit about SRFI Lassi Kortela (11 Nov 2022 21:12 UTC)
Re: Hate post on Reddit about SRFI Arthur A. Gleckler (10 Nov 2022 22:31 UTC)
Re: Hate post on Reddit about SRFI Marc Nieper-Wi├čkirchen (11 Nov 2022 06:58 UTC)
Re: Hate post on Reddit about SRFI Lassi Kortela (11 Nov 2022 08:50 UTC)

Re: Hate post on Reddit about SRFI Lassi Kortela 11 Nov 2022 21:12 UTC

> I feel the need to write something reasonable why there is a need for SRFI.
> But I don't know what to write. I was hoping that someone will be more
> reasonable and write something. Not necessary to start arguing.  But to write
> canonical answer what is the need for SRFI. I was hoping of someone came with
> proper answer.
>
> I wanted to write something about recent discussion about SRFI being like
> standard libraries that can be installed on different Scheme systems, but
> this is not what SRFI are for.
>
> I would love to read good argument in favour of SRFI, if not on Reddit then
> maybe here on this list. Note that if you comment, other Redditors will read
> it not only the author. And there are a lot of Scheme users there.
>
> This can be start of great discussion.

The best argument is to count how many Scheme implementations ship with
SRFIs, and how many Scehme programs use them. The answer in both cases
is probably most of them.

 From experience, I would advise against starting a "standardization vs
non-standard implementation" discussion. It's like a dynamic vs static
typing discussion - lots of people are firmly in one camp, and will not
change their mind. (The current types discussion on this list is dynamic
AND static, not dynamic VS static.)

The troll we had earlier was convinced that less standardization will
save Scheme, without realizing that less standardization is where Scheme
came from. The situation he wants has existed for decades and it doesn't
work too well. More standardization has worked somewhat better. I
expect, but cannot prove, that much more standardization would work much
better.

SRFI is obviously not ideal for every problem, and may not be ideal for
any problem. But its detractors should stop endlessly repeating the same
criticism and build something better. That's not easy. I know what I
want to build, but it will be a huge and risky effort for which I need
to make a lot of preparations that take time. We need people who want to
work, and ideally tolerate risk. That's a basic attitude which is hard
to cultivate by discussion.