Contribution vs quality, and a few notes on the Platform process

Duncan Coutts duncan.coutts at
Thu Nov 11 16:47:44 EST 2010

On 11 November 2010 17:40, Ian Lynagh <igloo at> wrote:

> Should the HP attempt to provide:
>    A set of packages that are popular and meet certain quality
>    standards
> or:
>    A set of packages that are popular, meet certain quality standards
>    and have a consistent API
> ?

"has a consistent API" is a level of detail of package requirements
that is not specified in the overall HP goals.

When the HP steering committee set up the procedure for adding
packages we discussed how we would agree what standards should be
required. We discussed trying to work them out in detail beforehand so
that reviewers would just go through a checklist. We decided just to
set up a list of what we felt were minimal and non-controversial
requirements. For more detailed requirements we decided to take the
approach of doing some actual reviews to help us work out what these
standards ought to be (we feared that discussing standards in a vacuum
would not necessarily lead to good decisions and would block proposals
in the mean time).

Note in particular that the section says:

   It is expected that the list of requirements will be adjusted over
time by further agreement of the libraries list.

So certainly, the libraries list can and should discuss and agree on
standards / guidelines for HP packages. The HP steering committee
probably ought to take the initiative and kick off such discussions
about specific proposed standards/guidelines, but in principle anyone
can do it (we've had some good suggestions in these recent threads
about documentation standards). We can then record the decisions in
the page above and package authors/proposers can take them into
account when preparing packages and proposals for inclusion.

BTW, as for whether they are requirements or guidelines, the policy
page above states:

  Every package should fulfil the following requirements. Any
requirements that are not met must be clearly explained and justified
in the proposal.

It is then left up to the reviewers.


More information about the Haskell-platform mailing list