Here are some comments regarding Haskell and the website. I think
there are some things we can do to make it a little easier on people
that enter haskell.org from outside the sheltered halls of
academia.
I think the comment about the organization of haskell.org is fairly
accurate - we have been adding more and more content without a really
systematic way of searching for things. I would really like to be
able to sift through the information at haskell.org in a more
useful way. Much of the best content on haskell.org is hidden in
Mailman archives - I also have problems digging through old topics.
I'm not sure we should change over to different discussion software
but maybe we could do something better to access the archives.
John
------- Start of forwarded message -------
Subject: Haskell web
To: "peterson-john@cs.yale.edu"
From: "Jim Lewis"
Hi John
I would like to offer my input on how to make Haskell more successful.
In running a software publishing company over the years I got a bit
behind in language technology and only discovered FP and Haskell a few
days ago. FP in general and Haskell specifically appear to be a great
advance over OOPs (we use c++). So it's surprising that Haskell is not
more mainstream. With years of experience in software publishing, I
thought it might be helpful to offer some strategic feedback (aka
constructive criticism), knowing well that I am probably not the first
to make these points. Actually I am excited enough over the potential
of Haskell that I would consider trying to help commercialize the
language but unfortunately my plate is rather full at the moment.
In summary, the problem is that the site and tools nearly stopped me
from learning Haskell. I had a lot of trouble getting over the initial
hump - not so much because of the concepts but because of the site,
tools, and docs. (I still only know about 10% of the language but I
can now get some trivial apps (a few pages of code) to run.)
Here are some of the problems I encountered:
1) Site organization - I wasted hours finding SOEGraphics as I missed
the link about older versions. It seems unlikely I am the first
with this problem and must conclude that someone is not carefully
listening to feedback. I almost gave up with Haskell due to this
small glitch.
2) The documentation on the site and in two books I have starts out
slow and then picks up speed too fast leaving the reader in the
dust.
3) The archaic forum UI makes it impractical to find anything and the
zip archive is a broken link. And the topics are mostly too
advanced for a novice. For about $150 one can buy a modern forum
management package. Downloading zip archives went out of style
close to a decade ago.
4) The search options are usually unproductive. For example I could
not find how to resolve how to handle the Maybe construct in my
app. I only found advanced discussions that did not help. My use
was very simple.
5) Someone is not listening to bugs. Hugs crashes on startup after
changing any value in the options dialog. This also wasted hours. I
can't be the first to experience this. I also almost gave up after
this experience.
6) The language reference has numerous functions with no description
or examples. I know you know this but find it hard to understand
how it can remain so. The Prelude Tour on the other hand is a model
of the proper approach - I found it very helpful in learning the
language as each function showed the description, code, syntax, and
examples.
7) There is so much complexity to the language that it should be made
clear what minimal subset one needs to learn to start writing some
code. It took a while to figure out what I needed. Once the user
starts writing code successfully he will be much more hooked.
I realize it must be difficult in a non-commercial environment to
depend on volunteers to fix such problems but these types of glitches
will stop all but the most dedicated newbies. I would like to see
Haskell succeed and offer any advice on setting a better strategy for
managing the spread of the language.
No doubt you've heard most of the above many times before but perhaps
I've added something new from the perspective of someone with almost
two decades of experience in commercializing software.
Regards,
Jim Lewis
President
Micro Logic
New Jersey
201-447-6991 x 113
------- End of forwarded message -------
