Hi David,

David Welton wrote:
>  For those wishing to play around with something, I think that Hecl is
>  in some ways a bit simpler and easier to hack than Tcl, and might be a
>  fun sandbox for people.  Hecl currently has references, as we wanted
>  to be able to handle references to Java objects without using a hash
>  table to look them up.

I've had a look at Hecl a while back and I think it works remarkably well.

Contrary to Cloverfield, which keeps Tcl's copy-on-write semantics, Hecl 
uses references exclusively, as does Java, so it shifts away from Tcl 
semantics (for example, [incr $i] increments variable i, and [list $a $b 
$c] is a "live" list), but this totally makes sense since it's closely 
tied to Java. However I remember that an early version allowed both 
values and references, the latter using an ampersand instead of a 
dollar. Can you confirm this?

My goal with Cloverfield is to keep the existing Tcl semantics 
(copy-on-write, pass-by-value) while adding references. This implies 
that the string rep of a structure holding references may change over 
time. This would be grossly inefficient if the whole string rep had to 
be recreated over and over again, especially for large structures, so we 
need an efficient string representation that allows for selective 
invalidation of substrings. Unfortunately Java doesn't seem to be 
suitable for this task because of its everything-is-an-object, 
all-reference approach: locality of reference is poor and would kill the 
performances in this critical part of the implementation. That is why I 
plan to build the reference implementation from scratch using either C 
or (sanitized) C++, the primary runtime target being the LLVM 
infrastructure. However Java is a very likely secondary target because 
of its ubiquity. The string representation I have in mind is based on 
immutable, constant strings, so this is a good match with the Java model.

I'll take a look at the Hecl code base to see whether it can make a 
suitable experimental testbed. Having a prototype implementation of 
Cloverfield references would certainly help people grasp the concept and 
eliminate the design errors.

Cheers, Fred.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Tcl-Core mailing list
Tcl-Core@[...].net
https://lists.sourceforge.net/lists/listinfo/tcl-core