Regarding dupe Re: New Libraries Proposal process

Ben Gamari ben at
Thu Jul 1 13:58:35 UTC 2021

Carter Schonwald <carter.schonwald at> writes:

> I think wrt to the dup/dupe proposal the answer at the time was nope.

> 1) there wasn’t a clear name that wouldn’t collide with user code
> 2) it wasn’t clear that it provided new expressive power / reduced
> complexity.
> To be clear, there’s a higher bar for adding / changing exports of
> preexisting modules, and preexisting wide spread conventions in programming
> practice is often a good way to motivate including such operations. Or that
> it adds a meaningful abstraction that helps all users.
> Also there didn’t seem to be much wide spread support in the discussion
> that followed.
> @simon : that process doesn’t have clear buy in by those who participate
> actively in the libraries ecosystem. And raises questions around
For what it's worth, I agree that it would be beneficial to have a more
formal libraries proposal process. The current process has a number of

 * it has shown itself to be lossy; that is, proposals are sometimes

 * intransparent: I have found it difficult to locate clear conclusions
   on mailing lists threads when evaluating ghc core libraries MRs.

 * difficult to audit: when one does find a message that looks
   authoritative are carries a clear conclusion, it often isn't clear
   whether this is the opinion of the body as a whole, or simply one

 * may be poorly representative: recently most of my interactions with
   the CLC have been by pinging @core-libraries on GHC MRs, where I get
   at most one or two replies. My fear is that while the nominal size of
   the body is ~9 members, only a few members are actively engaged,
   suggesting that there may be a deeper problem with the structure of
   the body

In my opinion the opening of Carter's message ("I think ...") captures
the problem here rather well: there is no single place where conclusions
are clearly articulated and flagged as such. All we can do is trawl
through mailing list threads of yore and attempt to reconstruct the
collective state-of-mind of the committee. This is what Chessai's
proposal seeks to fix.

> 1) I don’t think we have enough volunteer effort to actually manage /
> support Such a process (it’s tantamount to expecting full time
> professionalization expectations for libraries core contrib... which I
> think our community and industrial sponsorship isn’t large enough to
> facilitate.  ). Plus that dramatically increases the participation
> expectations / requirements from clc members.
I don't believe anyone is suggesting that the process should require a
full-time employee to manage. Rather, I would point to the GHC Proposals
process as a process that is reasonably formal, effective, and
yet operates on relatively little time investment by its steering
committee. Given that the CLC's responsibilities are at least as
important as those of the GHC Steering committee, I see little reason
why this couldn't, or shouldn't, be replicated.

Of course, any more formal process will likely require time investment
than the status quo. To this end, I think it would be beneficial for the
CLC to discusss the expected time commitment required by the body and what
time individual members are able to provide. If the time investment
necessary for an improved process is greater than the volunteer effort
available then there are three options: (a) scale down the process, (b)
increase the size of the committee, or (c) kindly thank existing members
who aren't able to put effort into the body for their service and ask
that they step down.

> 2) it creates authority that frankly clc did not have before. (Ed and
> others who have been part of clc historically  please correct me if I’m
> wrong).  Clc is supposed to be a facilitator and tie breaker and guider of
> evolution of base and a mechanisms for engaging in community consensus of
> how to improve important libraries.
I completely agree with your last sentence. However, in light of this
I'm not sure I understand the point made in your first sentence: What
new authority are you seeing in Chessai's proposal?


- Ben

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 487 bytes
Desc: not available
URL: <>

More information about the Libraries mailing list