World: r3wp

Which one is harder?

At the first sight, none is, I bet, but at the second one:

probe reduce dynamic [] 3

probe reduce definitional [] 3

The answer is different

So, which one is harder?

Btw, for DYNAMIC you will get different answers depending on whether 
you are running it in R2 or in R3

R3 
>> probe reduce dynamic [] 3
** Script error: i word is not bound to a context
** Where: reduce
** Near: reduce dynamic [i i i i] 3

>> probe reduce definitional [] 3
[3 2 1 0]
== [3 2 1 0]

So, any favourite?

I'm trying to understand what's going on in R3...

(in the dynamic version)

I could show you a similar example with CATCH/THROW, but it would 
require more code

Dynamic variables in R3: they are "not bound to a context" when the 
function is not running.

(that formulation is a bit unfortunate, meaning, that I would not 
have used it, but, in essence, it means: "don't ask what the value 
of dynamic variable is, when the function is not running")

Nevertheless, you can insert something like


if i = 0 [print mold block] to see the result, and compare that to 
the result in R2

wow, I think you've just illustrated a big part of the issue and 
you should add the above to your exceptions page.

it does a better job at explaining the underlying issue than the 
R3 error page IMHO.

The thing just illustrated were "dynamic variables versus definitional 
variables"

yes, but the error document tries to illustrate it too, and I didn''t 
see it as plainly as I do now.

need to check that

I find that the above is a good intruction which makes the  "Exception 
Marker Scoping"  more understandable in detail.


its not exactly the same, since the error document talks about unwinds, 
but I feel that the above helps me understand *why* the unwind does 
what it does.

I really need to fill in the Exceptions page with the real differences 
and advantages of dynamic and definitional scoping. There are advantages 
that either has that the other does not. It would be a real loss 
to not have both in different circumstances. The different circumstances 
being definitional for returns and dynamic for breaks and throws.

And anyone who thinks definitional is always better needs to look 
at #1521.

with the above example, I'm thinking about the past discussions and 
they are starting to make sense in a new light  :-)

I have discussed a new page organization strategy for the Exceptions 
page with Ladislav. As time permits, I will try to reorganize. It 
will make these issues much more clear.

Max, the errors page you mentioned tried to illustrate the difference 
between dynamic and definitional return, which looks more subtle, 
than the difference between dynamic and definitional variables...

its just less complicated to build mezz code with dynamic returns

 - Depends on the code. Some obscure but important things are impossible 
 if you have dynamic returns. Some less-obscure things aren't even 
 defined as a concept if your returns are definitional..

The reason we can get rid of dynamic returns in R3 is because the 
critical stuff that we used to do with dynamic returns in R2 is *necessarily* 
handled by dynamic breaks in R3, so dynamic returns are no big loss.

A description of the difference between the definitional and dynamic 
return:


*the dynamic return returns to the dynamically closest function (i.e. 
to the function that was called as the last one)

*the definitional return returns to the definitionally closest function 
(i.e. to the function that was defined as the last one and contained 
the return in question)

That's not all. Definitional also requires the affected functions 
to have (and in some cases fake) lexical scope. For R3 this means 
nested blocks.

lad,  what I mean is that the illustration of the variable issues 
above helps us realize that there even *is* a dynamic vs definitional 
distinction. 
it then allows us to more fully understand the return issues. 


you might be surprised that I never even realized there where dynamic 
variables, in the way you present it above.  Even if I've grown to 
understand binding pretty deeply.

aha, OK

in a sense... you can say that I might even begin to understand your 
guru code now  ;-)

that is you who should say that ;-)

hehe

Fortunately, functions are built with nested blocks, though that 
is partly enhanced in R3 by tasking issues. And while loops are also 
build with nested blocks, BREAK also applies to PARSE, which relies 
on dynamic scope (and yes, I can point you to mathematical proofs 
of this, though there's no point), so we still need dynamic BREAK. 
Dynamic THROW is just a good idea because of all the capabilities 
that would give us.

but, nevertheless, there isn't any reasonable defnition of dynamic 
versus definitional concepts in the article, so I take that as a 
hint

Absolutely right, Ladislav. There needs to be such a definition, 
or at least a comparison. I was going to add one.

Not to your page of course :)

Probably an example (or two) is preferable

This would be part of the issues section, so it would be mostly prose. 
This is part of a strategy to remove most of the prose from the other 
sections, so there will be more room for examples.

btw, the dynamic variable, just solved a long standing binding problem 
riddle I had with classed based instantiation of objects in R2.  
when I had talked to Carl about it at the last devcon, even he couldn't 
explain it... so its a very nice thing to get "out in the open".

With parse being a dialect on it's own, BREAK in PARSE can be separate 
from BREAK in loops. So that's not really an argument to keep dynamic 
BREAK (for loops).

That way the issues can be laid out clearly, so the rest of the page 
can just link to the appropriate issues sections instead of repeating 
things.

PARSE doesn't bind the code in its parens - that's regular REBOL. 
Parse also can't know ahead of time which blocks it will be treating 
as rules, because it uses dynamic scope for the whole dialect.

Definitional returns or escapes rely on lexical scope. If you don't 
have lexical scope, you don't have the ability to do definitional. 
So what PARSE needs is for the top-level BREAK to be dynamic. And 
if one level is dynamic, we are better off with all levels being 
dynamic, at least for the same escape function. Same goes for definitional.

Thanks for restating the facts. Once again, there's nothing intrinsic 
in PARSE that forces it to use the same BREAK used in loops.

Yes there is: Definitional breaks redefine BREAK to a definitional 
function, and PARSE relies on BREAK being dynamic because PARSE is 
inherently and necessarily dynamic and doesn't rebind anything in 
the parens, nor should it. For that matter, PARSE rules can be reassigned 
in their productions, so PARSE can't even rely on the rules being 
the same thing the next time round.

note that there are two usable BREAKs in parse.    parse [break] 
  vs   parse [(break)]

Use LEAVE-PARSE instead of BREAK for exiting PARSE.

Maxim, there is also parse [return].

See also the change log notes for A98, where the BREAK handling was 
actually introduced:
http://www.rebol.com/r3/changes.html#section-23