World: r3wp

ah, ok then ....

or maybe carl made it like make function!

make function! [[spec] [body]]

(i'm going from memory... i think modules were described as i said 
above so tasks should be similar, but function-like makes sense too 
if the header is mandatory)

anyway these are the kind of things that are going to change during 
the alpha phase in june

ie. we let some developers play with it just so that we can make 
final decisions on this stuff.

hmm, at least 4 of my msgs lost ....

seems ok now ... well, I just said that the syntax looks ugly - too 
many brackets :-)

but what is definitely issue to me is that it is not imo consistent 
with how we construct scripts ...

well it's basically going from make task! spec body to make task! 
reduce [spec body]

and task for me is kind of higher level construct as script or modules 
are ... well, kind of "instance" of "environment"

so you just write a wrapper task: func [header body] [make task! 
reduce [header body]]

if you look at it this way... i'd actually want it to not support 
the above example at all.

ok, so then take my note for the discussions .... pekr did not like 
task syntax :-)

:)

btw - could scheme be seens as a "class" for port?

yes, i think carl described it exactly like that.

or is scheme different things? It seems to me it contains mostly 
definitions?

aha, ok .... at some point, Carl blogged about classes/objects in 
R3, but maybe I am confused? What happened to that concept? IIRC 
it was related to some low level stuff, to save memory or something 
like that ...

carl hasn't mentioned that at the devcon.

imho it's not really that important once you have user types.

was it that one? http://www.rebol.net/r3blogs/0035.html

we will have user types?

yes... carl said he had some issues while implementing them

implementing what? classes, or custom user types?

if we will have custom user types, isn't it what you wanted? :-)

I am sure you will not like following, but I would rename mutex - 
that is typical IT related term, which makes sense to programmers, 
but which has no particular meaning to me.  Why not 'lock? or anyting 
else?

Mutex is an ugly term. But it is precise. Lock could mean several 
things, including shared locks rather than exclusive locks.

custom user types (question above)

Semaphore? Longer, I know.

Semaphore sounds more logical .... it at least means something ....

and most amiga coders already know about what it means ;-)

which is probably about 95% of rebolers, I guess.

well, I used Amiga Basic once...

not exactly the pinnacle of programming :-)

the one by Microsoft, distributed with OS1.3? :)

yes

that was before I hated Microsoft, but I couldn't understand quite, 
why my old C64 had a faster Basic

If 95% of rebol user are really former Amigans, then it is sad. This 
looks as if they have chosen Rebol just because of  the merits of 
Carl S for the Amiga and his popularity under the Amiga users .

sqlab - maybe that is why REBOL did not get popular enough - the 
Amiga curse :-)

I'm not an Amigan, though I do have one book on 3D graphics programming 
in Amiga BASIC, and I have a good friend who had an Amiga and a Video 
Toaster. I didn't know who Carl was until I found REBOL.

That's the same for me. I came to Rebol looking for an easy solution 
to a communication project.

Ditto on Greggs statement, almost exactly, (except for the book part)

I was just joking btw :-)  REBOL just attracts a lot of Amigans, 
on its own merits.  Most of these are FORMER amigans, who liked the 
simplicity of amiga's design.

Never did get the chance to use the Amiga. Admired it from afar, 
mostly based on magazine articles about its internals.

My story is about the same as BrianH's. Really wanted an Amiga, read 
up on it a lot. By the time I could afford one (busboys don't make 
a lot of money) the Amiga was obviously done for.

I understand, that REBOL3 won't be 100% compatible with earlier versions, 
so some scripts will eventually crash. I also understand the reasons 
for this decision, the urge to get everything as 'right' as possible, 
that the language concept for REBOL is so inventing as it is, etc. 
But for larger projects, developers making libraries etc., this is 
not a good thing. Are there any plans around this problem for the 
future? Maybe compatibility modules is a possibility? Or will that 
be very hard to do, because it might lead to a mixture of old and 
new code?

I imagine combatibility libraries would have to come from 3rd parties. 
RT is distracted enough as it is

Geomol - let's really forget the compatibility for now ...

and - I would like to see proposed improvements to library interfacing. 
Once interface is convenient, why should it change?