Apologies for the previous mailer-mangled "draft"... On 10/21/2015 07:55 AM, Herbert Valerio Riedel wrote:
Hello,
On 2015-10-21 at 02:39:57 +0200, Geoffrey Mainland wrote:
[...]
In effect, only those who actively follow the libraries list have had a voice in these decisions. Maybe that is what the community wants. I hope not. How then can people like me (and Henrik and Graham) have a say without committing to actively following the libraries list?
We have a method to solve this: elected representatives. Right now the Core Libraries Committee elects its own members; perhaps it is time for that to change. [...]
Proposal 1: Move to community election of the members of the Core Libraries Committee. Yes, I know this would have its own issues. How exactly do public elections of representatives address the problem that some people feel left out? Have you considered nominating yourself or somebody else you have confidence in for the core libraries committee? You'd still have to find somebody to represent your interests, regardless of whether the committee is self-elected or direct-elected.
Here's some food for thought regarding language design by voting or its indirect form via a directly elected language committee:
Back in February there was a large-scale survey which resulted (see [2] for more details) in a rather unequivocal 4:1 majority *for* going through with the otherwise controversial FTP implementation. If the community elections would result in a similar spirit, you'd could easily end up with a similarly 4:1 pro-change biased committee. Would you consider that a well balanced committee formation?
Thanks, all good points. It is quite possible that direct elections would produce the exact same committee. I wouldn't see that as a negative outcome at all! At least that committee would have been put in place by direct election; I would see that as strengthening their mandate. I am very much aware of the February survey. I wonder if Proposal 2, had it been in place at the time, would have increased participation in the survey. The recent kerfuffle has caught the attention of many people who don't normally follow the libraries list. Proposal 1 is an attempt to give them a voice. Yes, they would still need to find a candidate to represent their interests. If we moved to direct elections, I would consider running. However, my preference is that Proposal 3 go through in some form, in which case my main concern would be the Haskell Prime committee, and unfortunately nominations for that committee have already closed.
Proposal 2: After a suitable period of discussion on the libraries list, the Core Libraries Committee will summarize the arguments for and against a proposal and post it, along with a (justified) preliminary decision, to a low-traffic, announce-only email list. After another suitable period of discussion, they will issue a final decision. What is a suitable period of time? Perhaps that depends on the properties of the proposal, such as whether it breaks backwards compatibility. That generally sounds like a good compromise, if this actually helps reaching the otherwise unreachable parts of the community and have their voices heard.
My hope is that a low-volume mailing list would indeed reach a wider audience.
Proposal 3: A decision regarding any proposal that significantly affects backwards compatibility is within the purview of the Haskell Prime Committee, not the Core Libraries Committee. I don't see how that would change much. The prior Haskell Prime Committee has been traditionally self-elected as well. So it's just the label of the committee you'd swap out.
In the recent call of nominations[1] for Haskell Prime, the stated area of work for the new nominations was to take care of the *language* part, because that's what we are lacking the workforce for.
Since its creation for the very purpose of watching over the core libraries, the core-libraries-committee has been almost exclusively busy with evaluating and deciding about changes to the `base` library and overseeing their implementation. Transferring this huge workload to the new Haskell Prime committee members who have already their hands full with revising the language part would IMO just achieve to reduce the effectiveness of the upcoming Haskell Prime committee, and therefore increase the risk of failure in producing an adequate new Haskell Report revision.
My understanding is that much of the work of the core libraries committee does not "significantly affect backwards compatibility," at least not to the extent that MRP does. If this is the case, the bulk of their workload would not be transferred to the new Haskell Prime committee. Is my understanding incorrect? The intent of Proposal 3 was to transfer only a small fraction of the issues that come before the core libraries committee to the Haskell Prime committee. In any case, we would certainly need to clarify what "significantly affects backwards compatibility" means. Perhaps we should consider direct elections for the Haskell Prime committee as well as changing their mandate to include some subset of the changes proposed to libraries covered by the Haskell Report. My understanding of the current state of affairs is that the Haskell Prime committee is charged with producing a new report, but the core libraries committee is in charge of the library part of that report. Is that correct? Cheers, Geoff
Regards, H.V.Riedel
[1]: https://mail.haskell.org/pipermail/haskell-prime/2015-September/003936.html [2]: https://mail.haskell.org/pipermail/haskell-cafe/2015-February/118336.html