haskell-generalization-proposal.md

My response to https://docs.google.com/forms/d/1Rffj60hZXcg6zfu1dSGttHLTTsEcqXj8D66qIn9Grno/viewform?c=0&w=1

Your Name, and affiliation (if you use Haskell in a work context)

Mark Hibberd, Ambiata

Your preferred course of action *

 [Plan FTP] Implement the Foldable/Traversable Proposal in 7.10, matching the release candidate.

How you use Haskell

Significant amounts of production code at Ambiata is Haskell 
(data processing/machine learning/analytics). This is all mission 
critical for us. We currently deal with the existing "prelude" by 
pretending it doesn't exist. Getting the FTP proposal means a 
reduction in the amount of busy work our projects need to be 
developed at a professional level. We also hire and invest in 
training people with little to no Haskell experience, the current 
prelude situation makes this extremely difficult, where the difference
between "hobby Haskell" and "production Haskell" is significant. 
Anything that can be done to reduce this quickly would be extremely 
beneficial to us.

I also am a part of an ongoing effort at teaching Haskell and FP
across Australia on behalf of NICTA (https://github.com/NICTA/course)
who fund free courses targeted at improving the skill level of 
professional software developers in Australia. We have run this course
many times over the past 4 years, for 100s of developers. The FTP is 
also extremely helpful from a teaching perspective, for the same reason
as the training perspective above - basically a reduction in the tension
between "default hobby haskell" and
"functional programming with abstractions".

Thoughts and Concerns

I am concerned with delaying for two reasons, as mentioned above, FTP 
provides a much better default position for haskell - so the quicker we
get to that the better, but also secondly because I fail to see what
can be expected to "improve" in the delayed time frames. From an 
outsiders perspective, the CLC have done an amazing job of communicating 
this change, being extremely conservative in how it is approached to 
minimize negative impacts on everyone, and have done the hard, detailed 
work required to get this into a release. I don't see more being done 
in the extra time, only the current work bitrotting, and more fear and 
squabbling over what is just indecision. This is not a good position 
for Haskell or the people who rely on it. My feeling is a decision should 
be made, and if it is to generalize, then it should be in 7.10.