[Haskell-cafe] who's in charge?

Steve Severance steve at medwizard.net
Fri Oct 29 00:06:32 EDT 2010

> 2) This is a "doocracy" (man do I hate that word!).  If there is a problem,
> here's what you should do about it, in descending order of attractiveness:
> a) Fix it yourself
> b) Pay someone else to fix it
> c) Motivate or politely encourage others to fix it, providing moral support,
> etc.
> The key point is: you haven't paid any of us for this, and you have nothing
> even close to some sort of support contract.  I perceive a sense of
> entitlement on your part that people owe you no-cost coding.  That's just
> not how the community works.  Whether or not you really have that sense, I
> don't know, but your messages convey it nonetheless.

I understand why this and have accepted it. That being said the lack
of certain well documented high quality and widely written about
libraries are probably the biggest challenge to Haskell adoption after
the IO Monad. As a beginner I really have struggled to understand what
the right libraries to use are for common problems (specifically
databases). However other problems like the lack of a high resolution,
well documented date time library were also a source of friction. The
reason I am using haskell is to solve specific types of difficult
problems in machine learning and it is very good at that. Gluing a
fully working stack together has been challenging.

I agree that people need to be proactive about solving their own
problems but whenever I here any open source community
(yeah...everyone not just haskell) tell beginners to contribute a
package I always scratch my head with a little bit of wonder. Would
you really want a package that someone like me who is still trying to
figure out how to utilize haskell's features would build? Do you want
my outrageous non-use of the Monads that haskell offers?  Haskell
adoption is a frequently discussed topic and I personally believe that
haskell will and should remain a niche language for solving certain
types of problems. However if others in the community take the
position that success is determined by adoption then Haskell needs to
have a solid and well documented story about all the problems that
developers that write regular applications face including database
communication, ORMs, XML, Web Services, web pages (I know there is
Yesod just making the point),model view controllers (hits self in
head), etc... If "regular" developers are the target then they want to
be told what to do and what to use (with something like MSDN or some
of the big java toolkits), not discover it for themselves. I say this
with superior confidence after spending 2 years in a large (2000+)
developer organization.

I believe as a relative newcomer that haskell is best utilized when
solving complex and difficult to reason about problems that involve
small teams of highly skilled developers. Most software does not meet
this criteria. There are a wide variety of inferior technologies ;-)
(ahem java) that can be used to build most applications in the world.
Haskell should not try to compete because it can't (marketing
problem), and there are many things that it enables to be done better.

Not intending to flame (or otherwise ignite the fires of passion) at all.


More information about the Haskell-Cafe mailing list