Learn Erlang in 5 seconds - competition - win a prize

David Hopwood david.nospam.hopwood@REDACTED
Sat Aug 26 05:10:54 CEST 2006


Garry Hodgson wrote:
> "Joe Armstrong (TN/EAB)" <joe.armstrong@REDACTED> wrote:
> 
>> I'd like you to send me your course suggestions, in both the 5 and 10
>>second category.

  "Anything that can be said in less than 10 seconds (including this), is
   an oversimplification."

Sorry to be the one to pour the cold water, Joe, but I think this
5 or 10-second sound-bite thing is a bad idea.

> Today's software requires concurrency, from multicore processors
> to massively parallel distributed systems.  Concurrency is hard,
> but most languages make it harder than it needs to be.  
> Erlang makes it easy.

Plenty of today's software does not require concurrency (at the programming
language level).

The second sentence is fine, but it doesn't say anything about Erlang.

Erlang does not make concurrency *easy*. Programming in general is not
easy -- even quite trivial programs often have errors.

It is not clear that making concurrency easy is a primary goal,
regardless of its possibility. "Reliable but difficult" would trump
"unreliable and easy".


(I could have made similar kinds of criticisms of most any of the
suggestions so far.)

----

Programmers are pedantic, skeptical people -- or else they wouldn't be
good programmers. They/we have a tendency to pick apart anything that
is said, often concentrating on the bits that are wrong, or wrong in some
situations. Therefore, don't oversimplify when speaking to programmers.

It is possible to do more damage to Erlang's reputation in 10 seconds
than it is possible to do good. You may argue that the rest of the talk
would be spent explaining the sound-bite(s) -- but why put so much
emphasis on something that may cause a bad impression, then potentially
have to spend considerable effort fixing that impression?

When explaining technical concepts to an audience most of whom are
unfamiliar with them, IMHO a slower, incremental approach is preferable.

This does not mean that short, pithy "slogans" cannot be effective, as
mnemonics for an argument. A slogan does not attempt to accurately convey
the argument by itself -- it only has to be sufficient to stand for the
argument *for someone who already understands it*, acting as a reinforcement
to their memory.

Here are two examples:

  "In the future we shall know more." (Lee Smolin on quantum gravity)

  "Well-typed programs do not go wrong." (Robin Milner on type theory)

In each case, you wouldn't attempt to explain the associated argument to
an unfamiliar audience by first stating the slogan. First make the argument,
including any caveats, and *then* associate it with the slogan.

-- 
David Hopwood <david.nospam.hopwood@REDACTED>





More information about the erlang-questions mailing list