World: r3wp

so to solve that i used fixed position with "at"

Wait, that would be in mezzanine code, the layout function - I am 
familiar with that function. Send me some test code that demonstrates 
the bug and we should be able to fix it.

The fix would make it into the next R2 release.

BrianH, notify Anton as well. I know he published some heavy patches 
to layout in DevBase.

Brian: sure, user defined types are interesting features, but do 
you seriously think that such low-level feature would be implemented 
(and finished) before 2010 if high-level features like VID take precedence 
? What would be the next priority, View's Desktop ? Looking at how 
R3 has evolved since the first alpha in june 2007, I see that only 
very few core vital features have been finished, like ports, and 
the rest of the time has been spent on less important things like 
unicode support (which is a *very* valuable addition, but not vital, 
because anyone can implement it at mezz level if required). Features 
that nobody, except Carl, can add to the language (because it's closed) 
like modules, threading, rebcode, user-types,... are still pending.

that how i told it  you just mix below and accross blocks

We should check to see if the bug still exists with Anton's current 
patches.

Doc: I agree totally.

BrianH here is the working sample that show the bug

view layout [ below btn "1" btn "2" across btn "3" btn "4" return 
btn "5" btn "6" below btn "7" btn "8" return b

tn "9" btn "10" across btn "11" btn "12" return btn "13" btn "14" 
below btn "15" btn "16" ]

that what i love with rebol  an sample is while discusing to pop 
out a bug is so easy to do  ^^

Doc, a lot of the module code is there already, mostly mezzanine 
and a few core changes that have already been done. Rebcode we can 
add ourselves after UDTs are there. Unicode was a lot more important 
than you think, because it required language design changes and had 
deep implications - it had to be done first. I agree that there are 
definitely some low-level things that I would like Carl to focus 
on next, like threading and the changes to the object! semantics 
(this affects modules and nearly everything else). Only then can 
the plugin interface be done, and with plugins come UDTs.

see you have buton 11 below the btn 8 instead of having it below 
the btn 10

expected position should be below the 10 and beside the 4

Got it, the return after the 8 is returning to the point set by the 
first below before the 1.

hum well if it was only a little hum spacing problem I wouldn't  
point at it but why the first column is rendering OK and the column 
with exactly same instructions is not

and where the bug arrive to the top of the mountain is when you add 
a 3rd colum you have the 2 first column rendered OK  and the bug 
show on the 3rd column ...

view layout [ below btn "1" btn "2" across btn "3" btn "4" return 
btn "5" btn "6" below btn "7" btn "8" return b

tn "9" btn "10" across btn "11" btn "12" return btn "13" btn "14" 
below btn "15" btn "16" return btn "17" btn "18"

across btn "19" btn "20" return btn "21" btn "22" below btn "23" 
btn "22" ]

Brian: I agree with you, but Carl seems to be more interested in 
"chasing windmills" than implementing vital core features.

but hum ???

look the 3 column code ... that's freak

It looks like some variable in layout isn't getting set correctly. 
This may be tricky - there may be code out there that depends on 
the broken behavior. R2 is in compatibility mode, you know. We don't 
make breaking changes if they will break too much, or if we do we 
break towards R3 compatibility. Still, this seems to be bad enough 
that the cde pattern would be avoided, so it should be safe to fix.

cde -> code

that why i proposed in R3 dev blog tomaybe change the way to handle 
block of widgets using grid to suport  the widget like in GTK+ (yes 
i'm not original shame  on me )

VID3 is completely different. I doubt the bug would even apply there.

hum i didn't tested the bug with teh R3  alpha i have but i will

Doc, I'm not assuming that Carl is chasing windmills at this point; 
unfortunately I don't have enough info to make that judgement. All 
I can tell is that the Unicode stuff was deep and far reaching, and 
required many core changes, so much so that it had to be done first. 
That was definitely not chasing windmills, that was bare minimum 
functionality for a modern programming language, something that all 
of the other language rewrites going on right now have had to do.

AFAIK, there is no above or below system in VID3. It layouts more 
like traditional systems like QT or GTK.

above = across, sorry

but since VID3 is in limbo right now, it may not be useful to do 
tests with it.

From what I can tell though, the work Carl is doing on VID is having 
implications for low lever stuff like binding and the object! model.

lever -> level

Modules are built on objects, plugins are built on modules, and UDTs 
relate to plugins. He is working on the core.

Even threads are affected by binding issues.

ass you stated VID3  is really different from VID2 ....  the code 
the way it is doesn't work on vid3

no it will absolutely not work in VID3 :-)

That's not a bug though :)

since vid2 is still existing yes ... VID3  who can tell when we will 
put our hands on it ?

BrianH - I noticed you several times mentioned changes to object 
semantics. Did I miss any discussions on that? I can't remember it. 
Or are you just expecting them to happen?

The discussions happened a year ago, mostly on Carl's blog and the 
R3 development world, but were put on hold until enough of the R3 
infrastructure was in place to do them properly (including binding 
changes and Unicode). The discussions will be resuming after Carl 
resurfaces.

Don't worry, it was just a year ago. I can still remember all of 
the major arguments on every side.

No classes though, so no worries there.

Much easier to fake class-based or other inheritance structures with 
the recend changes to BIND and IN that came out of Carl's VID work 
recently.

He's using those changes to do CSS-style inheritance for VID styles. 
Sometimes you have to do the high-level designs to get an idea about 
improvements to the low level that will have pervasive changes.

CSS style based that reminds me the desktop google and opera applets 
....

Not with actual CSS, mind you :)

The inheritance will work like CSS inheritance.

rebol wlll loos in originality .... and therefor ppl will say why 
to do GUI application in rebol since i can do them in desktop google 
or opera widgets with more functionalities

Um, no. VID is intended to beat HTML in functionality.

well I have to use it to see the benefit of it ...