World: r3wp

all the other internals, well, I'd implement completely differently... 
and I'm not talking Dataflow here.

I mean in making it as "open" and obvious as the face object is.

decision - difficult to say .... some form of more involvement, better 
organisation of efforst. Last two weeks I exchanged two emails with 
Carl. My opinion is, that we need to start working on several parallel 
fronts. DevBase, DocBase, new RT's site structure (both .com and 
.net). We don't know anything concrete about those. And those are 
importan - we need to start to think too, how to structure docs, 
if/how we allow them to be translated, etc.

it seems that the face is very well tought out, and with a simple 
guide you can grok most of its cincepts in an evening.

May I suggest to give a timeline to the decision and discuss and 
"fight" about it between the small group (choose your preferred discussion 
"place") and, as Gregg suggested, collect the resulting deliberations 
to be submitted to Gabriele and Carl?

VID dialect is the same.

VID internals are so obscure and limiting that in most parts, its 
easier to replace them than hook ourself in.

I would too, start with initial formal meeting with some agenda - 
including Carl if possible. One hour would not kill anyone imo.

GLayout has many things like scrollwheel support for all faces (without 
focus), but it was hellish workaround.

Things like Max is mentioning - top-down mechanism - setting goals 
- what do we expect from new VID ... yes, no, voting, reasons, or 
veto by Carl

Should VID has this, or that, etc., how to aproach layout, on-something 
handlers in upper level? etc.

I'd just like more hooking capabilities right in the api.   just 
like Amiga allowed many things to be extended without needing to 
replace them.

once done - 1 - 3 ppl implement.

It would be nice if RT communicated to us in here, we who are very 
interested in working as closely as possible with RT, but cannot 
do work directly on the core, on what it would be a very good idea 
for us to consider building.


Like: "RT thinks you should look at building a GUI system" or "RT 
would like you to build a test case suite for R3" or "RT would like 
you to work on making OpenGL work well with REBOL as a dialect like 
VID" or "RT would love to see you building protocols for this and 
that kind of communication" or "RT needs a very good multithreaded 
webserver, that can handle X users" and have those efforts officially 
endorsed by RT, similarly to how MUI eventually became the GUI of 
choice on the Amiga to build your applications on. Perhaps put out 
hard specs and see if anyone will pick it up.


Right now, many efforts feel like they are there, not because RT 
felt they were a good idea, but because some individuals thought 
they were good ideas. Most of us here speak highly of our own ideas, 
but without much dialog with RT. AltME feels like it's the only non-RT 
effort that is endorsed by RT and perhaps also Cheyenne. Such directions 
would also mean that perhaps a lot of people would flock to the same 
official project, rather than starting 2-3 separate projects.

being an ENTJ, I'll always want to start my own stuff though  ;-)

Maxim - that is OK. I think that the problem is, that they state 
being open, yet so far, they don't communicate.

for some reason, I always thought that once the R3 core was released, 
we would be allowed to measure it, come to grasp with IT.  then one 
or two updates later, we could better see how it might allow gui 
to be worked in.

IMO we are very few here already. What is the other closest group 
to ask for review in rebol community than here? There is RT ... then 
there is few ppl working closely with them - Ladislav, Cyphre, Gabriele, 
then few skilled here (I don't belong there) .... it is pity Carl 
can't find his way to top developers here.

I personally don't see the rush for view in R3.  frankly, no one 
in the community (or outside) will be wanting to fork their "serious" 
development on it, until a little bit of testing and all.

Communication is main barrier of RT imo.

brian tiffin needs to become the ultimate evangelist  ;-)

we all need to vote for him!

he has all the skills, the drive and the perfect personality ... 
someone who can tell Reichart to STFU and make Reichart laugh... 
well has impressive communication skills and intuition  :-D

well, I think Gabriele might feel demotivated by our lamenting here, 
while we don't mean it in personal way :-(

hehe pekr and me in the same room will have that effect usually. 
 ;-)

well, my understanding is, that it is also few others concert - read 
Henrik's reactions - he things along similar lines ...

but above all else, I see that what people are bringing back is the 
lack of support FROM RT into team work. In many different ways.

so... since Carl will not read all this, for communication to happen 
there has to be a message coming from the community too. what should 
i (or Gregg) tell Carl? telling him to read all of this will not 
work.

I really want it to be constructive, but the lack of Carl's communication 
is problem even for future imo. Some official organisation as rebol 
user group might work, as it could be more readable to Carl. I suggested 
him via email, that one hour per month or two of guru session here 
would help to keep spirit high .... it can fantastically motivate 
ppl ... and those motivated here can motivate down the way to the 
computing world ...

I would wait for Anton or Volker to add their opinion ...

My main question is why the rush?  why not let the community come 
up with a few example prototypes, why not let people like cyphre 
take care of the low-level architecture and make it as perfect as 
I know Cyphre can make it.

Well, generally I vote for - Anton, Henrik, Ashley, Volker, Maxim 
to form a close group to Gabriele and Cyphre, and discuss new VID 
- basic concepts ...

and why not give people like Anton, Ashley, Henrik, me and others, 
the chance to try out the proposed technology first.

I believe they are backing my opinions too.

I would hope that Anton will be included in any discussions on View, 
GUI planning. He knows it inside and out, and has been a great help 
to me through many GUI problems.

Maxim - whatever, but my undestanding is, that Carl really does not 
want all your technology. Start a dialect - start closed group of 
few named persons and at least outline solution - publish agreed 
or problematic bullets .... the rest will add opinion ... meet once 
again, close design doc, ---> implement

We can even come-up with main question to adress ...

that's Gabriele's POV, Carl did tell me "Tell me what you would need 
changed in R3 to allow DF natively"

so he is not specifically against the idea.

that will not take much time, and Gabriele might feel safe, that 
his implementation or proposal adresses most of wishes ...

I even said to all there that I share their angst WRT DF

(DF - DataFlow)

I fear that noone will understand deep DF thoughts, and that novices 
extending GUI might feel once again lost, like they were when looking 
at VID internals for the first time :-)

hum, I am an expert at VID... I have rewrote the wake event from 
scratch, I don't use ANY of the popup code for modal windows, etc 
and I'm still mystified by how VID's design came to be  ;-)

rememeber that I'm not trying to push DF into people throaths.  I 
know how its disruptive... its not just a cool thing, its another 
paradigm... its changes how you structure your applications.

but I think that DF used through a dialect is actually VERY simple. 
 my liquidGL demo show how simple it can become.  my DF view engine 
for elixir is another example at how the higher-levels can become 
simplified by adding a little bit of in-between stuff.

but when you see how it scales and how easily it can tackle the most 
complicated tasks, then it starts actually being much more simple 
than coding all the stuff by hand... that is the problem.

Max, do you remember what Carl said about Smalltalk?

I've faced the  same problem within the Visual FX industry where 
dataflow now has a solid grasp in the field.


when you want to do simple stuff, its often more complicated to use 
dataflow... but the moment you start building complex scenes... without 
dataflow... you might as well paint your movie using frame by frame 
oil paints  ;-)

yes, and its part of the angst.