[erlang-questions] node.js vs erlang

Richard A. O'Keefe ok@REDACTED
Fri Jun 20 02:30:21 CEST 2014

In another thread, Thomas Lindgren mentioned ToonTalk.
From Ken Kahn's original paper about ToonTalk:
    Some of the design principles derived from good construction
    toys and video games include:
    1.  Make the initial experience simple and
        gradually increase complexity.
    2. Encourage exploration and curiosity.
    3. Provide and maintain appealing fantasies.
    4. Continually challenge without frustrating.
    5. Frequent use of animation and film
       techniques and principles (video games only).
    maybe the difficulty is not in the concepts per se
    but their lack of accessible metaphors.

We got a bunch of bills recently, and it occurred to me that
there's probably a good metaphor for Erlang 'receive' there:
messages *arrive* at our house in some definite order, but
we can look at the envelope and decide which one to open first.

The HTMLisation of documentation has made it much much harder
for me to comprehend.  It's like the difference between learning
UNIX programming from a jumbled set of man pages and learning
UNIX programming from Stevens' book.  What I personally need in
dealing with something like rebar or cowboy is something that
*starts* by explaining
 - purpose
 - models
 - metaphors
I'm still wondering where the rebar manual *is* and why it
wasn't in the release.

I should also say that my experience of npm has not been
positive.  There was a program that I very much wanted to
play with that was distributed that way.

Step 1: install npm and get it to work in my environment.
(a day later)
Step 2: install the package.
Step 3: contact developer to find out why that doesn't work.
Step 4: install the package.
Step 5: contact developer to find out why it still doesn't work.
Step 6: install the package.
Step 7: fish around inside the rubble to find a working piece.
        (npm was *supposed* to install it in a directly runnable
        way.  It didn't.)
Step 8: run the program.
Step 9: oh, it's not finished yet.

I'd like to thank Kennita Watson for the First and Second Laws
of Examples:

   ONE.  There must *BE* examples.  Lots of them.

   TWO.  The examples must WORK.  For the user!

More information about the erlang-questions mailing list