Tomas, thank you for your helpful comments. I just wanted to comment on
a couple side points right now.
Tomasz Rola wrote on 12/29/18 1:48 PM:
I understand Racket community no longer considers themselves as part
of Scheme landscape, but I am unable to say where I get this
impression from.
I've been one of the people propagating that, for two reasons:
1. That was the impression I got of the intention of the name change
from PLT Scheme, and I guessed it was a time to be a team player. :)
2. Many of the most savvy developers will immediately want to use
"standard Scheme", not "embrace&extend proprietary variants". But
standard Scheme was a really great baseline foundation -- just not
sufficiently as a complete general-purpose production language. So
people were shooting themselves in the foot by trying to get by with
R5RS and SRFIs.
(I say this as one of those foot-sharpshooters. When I picked Scheme as
my next language platform for rapid R&D, I started writing missing
essential libraries and tools for it, and I went to a great deal of
trouble to keep them portable, and to test with maybe 10 different
Scheme implementations, with some scripts to package for a few of them.
Eventually, I decided I was handicapping myself too much, so I switched
to developing in PLT Scheme only, but always knowing how to switch to
Gambit, Kawa, Chicken, Guile, or some other Scheme implementation with
different strengths that I might need. I announced this, maybe a few
other Scheme and other Lisp people followed me to PLT, and some others
who focused on non-PLT Schemes took up repackaging some of my libraries
for those. The choice of PLT turned out to be a good one.)
(That said, RnRS is done by great people, and I think it would be good
if all the Scheme variants could find a common svelte baseline again,
and to continue to explore a diversity of different approaches and goals
atop that baseline. Then, for a comprehensive "standard library" today,
I'd favor a more decentralized model. Some of it perhaps much like
SRFIs, but maybe a little closer to IETF RFC and STD, combined with de
jure "standards" of the day based on current popularity of open source
libraries, some of which might eventually might turn into something like
RFC/STD. For example, given a baseline language, many libraries from
Racket, Scheme48, and elsewhere might simply be used, and that's enough,
and some might go through an RFC or STD process beyond that. In any
case, standardization efforts people care about tend to be a massive
amount of work, especially when there's community involvement, and I do
not envy those who heroically take on that task.)
BTW, maybe jobs using Racket will also encourage a lot more quality contributions of
packages, when there's the additional motivation of open source "auditioning"
for jobs, in addition to the current community participation, platform promotion, and
love of craft.
Forget it. I think this is standing a problem on its head.
That "BTW" was not suggesting how to create jobs, but speculating a
major reason why Rust-like contributions growth might happen once jobs
exist (and why Racket hasn't yet seen that degree of sustained
contributions).
(I have a holistic development model in mind that should give you
optional open source contributions almost for free (doing things you'd
do anyway for your project's and organization's immediate needs). But
that kind of organizational altruism/alignment/PR contribution, together
with individual purely altruistic or collaborative contributions, are
not the only kind or motivation of contributions -- another motivation
is marketing of individual developers. Individual marketing via open
source (and blog posts, etc.) has been a thing since at least the start
of Web popularity, but seems to be more of a standard practice than it
used to be.)
--
You received this message because you are subscribed to the Google Groups "Racket
Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.