[erlang-questions] Chicago Erlang Conference --- a dabbler's take

[Garrett Smith]

On Thu, Sep 25, 2014 at 12:00 PM, Lloyd R. Prentice
<lloyd@REDACTED> wrote:
> This post may be a stretch for this list but I humbly post it on grounds that it deals with Erlang questions in the large rather than nitty gritty "how do I..."
>
> First off, I can no longer refer to myself as a newbie since I've been hanging around too long. But I can't pass off as a wizard since I'm not. I do strive to plunge deeply into Erlang, but dabbler sounds about right.
>
> So, I spent a brain-frying day in Chicago last Monday breathing the fumes of exciting ideas and shaking hands with an impressive posse of Erlang leading lights. You missed much if you missed out.

Very kind words - thank you!

> Since, no doubt, the videos of the presentations will soon pop up on the web, I'll just say that, for the most part, they  dove deeply into the arcana of keeping complex Erlang systems and clusters running in top form. I learned much--- mostly how much I have yet to learn.
>
> Irons Guberman -- Maximizing Throughput on Multicore; How to get more bang for the buck from Erlang.  Super slides.

Irons Guberman is the toughest, most ruthless Erlang warrior to step
onto the stage. She delivered a high impact, Iron-Fisted talk that was
brilliant - even more so given it was her first conference
presentation!

Irina Guberman, Irons' soft spoken alter ego, was equally excellent,
engaging, warm and thoroughly entertaining.

> Garrett Smith --- Writing Quality Code in Erlang; As an old Forth guy I felt right at home. Factor, factor, factor until you distill out the essence of your solution in all its crystal clarity. Garrett was definitely channeling Chuck Moore.
>
> Brian Troutwine --- Monitoring Complex Systems; I hope I'm never in Brian's shoes. But if I ever have a complex system to wrangle, Brian would be the kind of guy I'd want by my side.
>
> Jesse Gumm --- Building Web Scale Apps with Nitrogen; Jesse exploded on the stage and wrote a non-trivial web app before our eyes in less than 40 minutes. Yeow!
>
> Other speakers were equally engaging, but not as close to my personal needs and interests.
>
> The big questions rumbled to the surface during the after-conference party.
>
> - Several Erlang newbies took on Joe Armstrong: "Why are libraries so inconsistent, documentation so sparse, and Erlang so hard to learn?" Joe held his own, but no doubt these questions will persist.
>
> - A group of Elixir groupies made a strong case for their new puppy. Thinking about learning one more language makes my head hurt. But the converging?/diverging? paths of old and new will be fun to watch.
>
> - And, finally, during waning minutes of the party, Garrett Smith told me that Erlang should be completely rewritten from the ground up. He's tried with e2, but there's no interest.

You forgot to mention that Garrett was stammering drunk and shouting
for Vinoski, saying he wanted to "teach that dirty scheduler writing
scoundrel a lesson he'd not soon forget!"

But to clarify, I did not say that Erlang should be rewritten, but
that some other new distribution, built atop the Erlang VM and
toolchain (similar to what Elixir has done), could be created that
cleaned up the core libraries and was free from legacy restrictions.

I personally like Erlang as it is -- I don't want a new language (like
Elixir). But I recognize that the legacy of Erlang prevents it from
moving aggressively into a more newbie friendly and accessible form.
That's a matter of physics, not politics, religion, or personal taste.
Erlang will continue to improve incrementally, but it will never enjoy
the freedom to build-from-scratch that Elixir has. But something new
*could*.

The problem, obviously I think, is the breathtaking cost of "rewriting
Erlang". This is a vision that is not likely possible. But one can
drink and ponder.

I have proposed a new language called Lamprey:

https://github.com/gar1t/lamprey

This doesn't exist. It's mostly a joke and I don't even know if it's
possible. But if it came to be, it would be epic in the history of
languages.

The idea is to take the stunning and amazing work of the Elixir
community and bring it back to Erlang syntax as a new language --
Lamprey. Lamprey would enjoy clean libraries, documentation, tools,
etc. without doing any work. When a new version of Elixir is released,
Lamprey would just run its bash scripts to convert everything back to
Erlang, and whamo - a new release of Lamprey!

Parasitism at the computer language level.

Robert Virding btw gets credit for the name :)

Thanks Lloyd for attending on Monday. It was a pleasure to finally meet you!

Garrett