World: r3wp

You mean compilation workflow?

yes and maybe light enlightment of design (ie metacode is created 
by compiler or linker?)

metacode? If you mean the native code generation, it is done by the 
compiler.

I'll see if I can make a short description on how it is organized.

and what linker cares about, glues together multiple compiled source 
files?

Compiled source files are glued together by the compiler itself. 
The linker currently simply generates the right binary from compiled 
code, global data and external libs imports. The linker will be later 
extended to be able als to statically link external libraries.

als => also

so emitter is only helper?

Emitter module purpose is to isolate as much as possible target-specific 
code from the compiler. So, it provides target-independent helper 
functions to the compiler and for target-specific code, it loads 
one from  %targets/ folder.

thx for description

The emitter module API is used by the compiler only when it can provide 
an added value, else the compiler taps directly in the target-specific 
code through a target-independant API. In other words, a new target 
module needs to implement all the current target API (all the emit-* 
functions in %IA.r) to be able to work with the compiler.

Doc - do you think we can get Red to the following page, or is it 
too preliminary?

http://en.wikipedia.org/wiki/List_of_programming_languages

somewhat premature .. it's not in use yet

I agree.

Probably too early.

Direct link to Red/System specifications draft changed, it is now 
available from: http://static.red-lang.org/red-system-specs.html

I just visited the Haiku OS website, and I can see, that it was accepted 
for the Google summer of code 2011. Maybe once Red becomes an alpha, 
we could apply too? Let's say for 2012 :-)

Google Summer of Code is a global program that offers student developers 
stipends to write code for various open source software projects
  I wouldn't qualify for the "student developer" part.

No, you would qualify for the mentor part. So you'd have to find 
students willing to write open source Red code, and then you have 
to mentor them

It's very hard to get into SoC, beginning with the rush to register, 
and if you don't have the organisation to conduct the mentoring, 
or your project is too eccentric for students to get into quickly, 
it's fairly pointless

It's also questionable if the students will go on to maintain their 
code. The biggest advantage if you do get in is simply the marketing, 
because people suddenly think you're associated with Google

(and google also gets to notice you a little)

Yeah, so they can steal your ideas... I figure the cost is well worth 
it to them

Kaj: thanks for the insights.

Questions about the declarations.  

1-  why are you using this syntax for *all* declarations? : 
word: [datatype value] 

ex:
my-val: [integer! 20]

woudn't just this do ?:
my-val: integer! 20

I find its a hell of a lot easier to read,


and when you add type detection, its back to 

my-val: 20   


meaning that the integer! word really is just a typecast operation 
here.

since you are compiling and pre-filtering the source code, the integer! 
word really is contextual,  what it means is inherently bound to 
where its found.   so I see no issue with this use syntax.

anyone care to debunk me?  am I missing something?

Where have you seen: my-val: [integer! 20] ?

word: [datatype value]  => I don't remember writing that in the specs? 
Maybe a typo?

Oh, I guess you're just refering to pointer! initialization?

darn... I'm completely off.. yeah... I guess the old pointer documentation 
got me all mixed up.

Doc, I have to say, the red docs are getting really nice.


its clear now, and I was talking nonsense... I should have gone over 
them... will make sure to visit each time. they've significantly 
updated since last I visited.

I was just taking a little break from all of the cheyenne dev.

they've significantly updated since last I visited

 You are too used to R3's monthly updates...it's Red here, it's updated 
 daily ;-)

hehe

yeah, I forgot that Red lives in the same space-time continuum as 
the rest of us humans  ;-)

If you have looked at Red sources since a long time (like a week 
;-)), you should look at the Quick-Test framework added yesterday 
by PeterWood: https://github.com/dockimbel/Red/tree/master/red-system/tests
(see the readme.txt)

have => haven't

cd %red-system/tests/
do %run-all.r

That's cheating, that was ported from Boron ;-)

hehe :-)

Kaj: BTW, Andreas added ELF sections table support a few days ago, 
so, does it help a bit loading executables on Syllable?

Yes, I saw that and want to test it, but I'm very busy. Maybe I'll 
get to it tomorrow

Kaj :That's cheating, that was ported from Boron ;-)


Not true. That was my initial intention but simple-test was felt 
to be too heavy for Red/System so I built a lighter one.

By the way, the boron framework was ported from REBOL ...
.... and back again :-)

OK. Sorry, I didn't mean it in the strict sense, more as predecession 
:-)

Given the potential licensing issue, it's best to be careful about 
that. *GPL is a one-way street.

simple-test.b is just a script written in boron. I don't believe 
it is GPL because of that just as all Java code is not GPL because 
Java is GPL.

Cool.

Though the "all Java code is not GPL because Java is GPL" thing doesn't 
apply to GPL2, which is why the Classpath exception was made. GPL3 
fixes this though.

Oh, I didn't realise that Java was released under GPL2. I thought 
it was only under Sun's proprietary licence until it was issued under 
GPL3.

(getting off topic) The main problem was that GPL2 code *written* 
in Java was illegal to *run on* proprietary JVMs and link to even 
the bundled Java libraries. The same goes for GPL2 code written in 
REBOL. LGPL2 is a little more legal for running on proprietary languages, 
barely.

There were and are a lot of popular GPL2 apps written in Java, but 
none of them are legal to run (ish, depending on distribution).