Implementing tables - advice wanted

Thomas Lindgren <>
Tue Jun 13 14:22:49 CEST 2006



--- "Joe Armstrong (AL/EAB)"
<> wrote:

> > I would instead recommend starting out by
> translating the 
> > notation into using dict or, even better, a
> well-defined 
> > behaviour. Then plug in more efficient
> implementations as needed.
> 
> I did that a long time ago - it's the more efficient
> implementation I'm
> interested in. 

OK, but why not settle the notation first?

For implementation of fast, pure arrays and hash
tables, I'd suggest looking at what could be called
the "exception list approach", e.g.:

ftp://ftp.csd.uu.se/pub/papers/reports/0040.ps.gz
ftp://ftp.csd.uu.se/pub/papers/reports/0042.ps.gz

or Eriksson & Rayner, 2nd Int'l Logic Programming
Conference, 1984.

The same approach was later used in Hipe
(pre-generational GC).

Best,
Thomas


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 



More information about the erlang-questions mailing list