Every few months, the principals of the Racket world meet for a day to
discuss the state and near (and, occasionally, distant) future of the
Racket world. We met at the periphery of PLDI on January 6/7 in Chicago,
IL. Here is the list of major points we discussed:
1. Expander and Compiler
As many of you know, Matthew is in the process of re-implementing
the expander in Racket and re-organizing it on the way. At the
moment, the revised expander is 2x as slow as the C-implemented
one. Investigations so far point to Racket's runtime system and
compiler backend as the bottleneck.
We are considering two options:
[] continue development of the new expander and related backend
improvements on a separate branch, bringing neither the benefits
of the new expander nor the drawbacks of decreased performance
until the latter are reduced, or
[] swap in the Racket expander and take the performance hit.
Matthew is continuing to investigate and explore alternate
runtimes, and we'll request feedback from the user community in the
near future when the issues are more clear.
1a. Also noteworthy, Leif Andersen is working on a replacement
frontend compiler in Racket using a nanopass framework.
2. License
Over time, the license we currently use for Racket has become an
obstacle for some potential users and is problematic for continued
compiler improvements.
We decided to re-license the software (dual license Apache 2/MIT) as
much as possible. You have seen the relevant email and request for
signatures. (https://github.com/racket/racket/issues/1570)
We are still pondering license issues concerning the Racket run-time
system (e.g., Gnu Lightning).
3. Committers
We have decided to grant access to committers to individual repos
instead of all repos at once.
4. Nightly Builds vs Alpha vs Beta
Some of our users asked for more frequent releases, which we
interpreted as the desire to have access to an 'alpha' and 'beta'
version of the software. In response we have decided to turn the
Northwestern snapshot build into an 'beta release site', with a
turnover of several weeks (probably six), and to turn the snapshot
site at Utah into an 'alpha release' site, with a faster turnover
than Northwestern.
The actual releases at Northeastern and at several mirror sites will
stay on our normal schedule.
5. Slack
We decided to use the racket-slack channel as another way to communicate
among ourselves and with all other developers and users.
6. RacketCon (from our June meeting)
As you saw from the announcement, RacketCon will take place in
Seattle this year, in collaboration with the large Racket group at
UW. This move also ought to accommodate some of the industrial users
in Seattle.
In 2018, RacketCon will move back to St Louis, co-locating with
Strange Loop and the International Conference on Functional
Programming.
Finally we would like to thank Matthew Butterick for his efforts
to improve the looks of our web site. If you like it let him know.
If you think it can be improved, thank him for spearheading it
all and let him know. And if you really want changes, submit
pull requests.
We will meet again in a few months. If you think we should
discuss a particular topic, send us email.
--
You received this message because you are subscribed to the Google Groups
"Racket Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/racket-dev/6492C61C-2917-4DCA-BB70-6E90ADC21C90%40ccs.neu.edu.
For more options, visit https://groups.google.com/d/optout.
