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" <peterson-john@cs.yale.edu> From: "Jim Lewis" <jimlewis@miclog.com> 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 -------