[erlang-questions] concurrency developments

Bob Calco bobcalco@REDACTED
Fri Jan 18 03:49:00 CET 2008

Thanks for the tip. I am definitely going to take this advice! 


Any more information about this strategy anyone can shed light upon would be
greatly appreciated!


-          Bob



From: Robert Virding [mailto:rvirding@REDACTED] 
Sent: Wednesday, January 16, 2008 5:17 PM
To: bobcalco@REDACTED
Cc: alex alvarez; erlang-questions@REDACTED
Subject: Re: [erlang-questions] concurrency developments


On 16/01/2008, Bob Calco <bobcalco@REDACTED> wrote:

I have recently decided to experiment writing a Ruby-like front end to
Erlang, code named Emerld, which would generate Erlang code that would then
be compiled (at least until I can learn the BEAM file format well enough to
generate it directly). It could theoretically be retargeted to .NET or Java
when either of the two got their act together for concurrency. That seems to
me more a matter of when than if. Until then though I have committed myself
to mastering Erlang, because I think it will give me a huge competitive edge
in the new multi-core world we live in as a software designer and architect.

If you are going to compile down to something other than straight Erlang
then the best target is Core erlang. It a pure relatively simple and
standard functional language which is used inside the compiler. The AST is
defined by a set of records and there are tools to read/check/print it. As
far as I know there is nothing *legal* you can do in the BEAM code which you
can't do in Core. Also you have the benefit of not being forced to modify
your compiler as the BEAM engine is improved. If I remember correctly HIPA
also uses Core for one pass in its compiler. Someone who knows can comment. 

This is what I am doing for my LISP front-end to Erlang. Soon to be ready.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20080117/be0d8d7d/attachment.htm>

More information about the erlang-questions mailing list