World: r3wp

Henrik - I can tell you one thing, which I smiled to :-) Carl is 
imo worrying, that releasing in current state could put us in situation, 
where we would not be able to handle the load. And I know, that you 
mostly back up everything Carl says ;-), but then I pointed him to 
R3-alpha release wiki page. He told me he forgot about that site, 
and that he forgot we already did something like that. So please 
- give me an answer - first public alpha was released in last january, 
and NOTHING bad happened. It was released only because of my constant 
push and ramblings. So - why should devs collapse under some heavy 
load now, one year later? Nothing like that will happen.

this bottom up, top down, and anything that grabs Carl's attention 
is not working.

Graham, it's working quite perfectly from my perspective: Bugs are 
being fixed as fast as they can.

Henrik - you can be sure, that many devs out there would share my 
opinion ...

Pekr, nothing happened, because no one is using it, because it doesn't 
have unicode.

Are you saying that noone was using alpha because of the lack of 
Unicode? That would contradict why ppl do use R2, no?

also the public release can't be used in production environments. 
R2 can.

Henrik - this was so nice. Release often, ppl felt the process is 
dynamic, ppl felt constant progress. We tried it several times in 
the past, but not longer than for one months, and then - another 
blackout: http://www.rebol.net/wiki/R3_Earlier_Releases

Henrik - we are dealing with human beings here. And you might be 
forgetting about it. If we want the community, not just customers 
to buy products, we have to support them, provide them ways to get 
involved. Even if the alpha was released to satisfy those unsatisfied 
ones, it served the purpose. Ppl will become familiar with it, play 
with it, maybe find some bugs. Who cares, if CureCode is loaded with 
thousands of tickets? The bugs are there, or they are not there. 
DevTeam can choose what to fix ....

The release early and often style doesn't hold up for early alphas. 
It works better when the product is more stable, so you can get reasonable 
feedback. I wonder how much time the KDE team would have saved, had 
they waited till 4.1 with a release.

But enough of that. I want to have some influence, as a suggesting 
voice. It is upon Carl to choose, how to further proceed. I am just 
stating, how I would proceed, and what consequences current aproach 
carries. If the majority thinks I am wrong, that I will stay unheard. 
But my experience is, that as for organisation of efforts, I am aften 
right.

I can understand how Pekr, Graham and many others feel about the 
lack of R3 releases especially given all the early announcements 
and the first public alpha. However, I really feel that Carl is still 
prototyopng R3, he is a long way from settling on the design of R3. 
There is too much missing for the current version to be considered 
an Alpha (e.g. No modules, no threads, ASCII GUI, Host environment 
& Runtime Core in a single executable,)


This is seems to be Carl's way of working and something that he has 
to work through step by step.

I prefere release later a better stable version instead of release 
soon a crap:) I believe we will not end up with the third version 
release later a crap. The fourth version: release soon a stable version 
is a best wish, but that's all.

Pekr, why Ladyslav isn't here anymore ?

noone really knows. He vanished from REBOL land suddenly, without 
any trace. No reply on other comm channels too. Hopefully he is doing 
fine.

I had a funny observation: You can edit rich-text with the DOC style 
due to a bug, by clicking over the pieces of text and start typing. 
It works nice and fast, although there is no cursor management, ways 
to move between different text styles with the keyboard or any font 
settings. But then rich-text editing should not be so far away.

i wonder if some great guys who help carl thought about Database 
access ? odbc? rebol native lib for mysql db2 oracle mssql etc....

thought about it, yes. who will do it and how it will be done? don't 
know. if anyone wants to, I'm sure Carl will approve.

ask again in another year ..

like we have "unless" for "if not", should we have "unempty?" for 
"not empty?" ?

I have 'something? for "not empty?"

thank you to henrik for those examples  of REBOL3 GUI ;-)  it seem 
to be faster than R2 and perhaps we could have some grid and db grid 
and others..........

something?: all [ not empty? blah found? blah ]

ita ha only sense as a native, should be pretty useful for GUIs

ita ha
 = "it make" .. 8)

who is the must skilled for this job ...................SoftInnov 
perhaps ????????????

dockimbel ??????

the same guy

don't see the point, if you want and SQL DB the you can  use any 
lib

and make routines

this solution is still cross platform ?

but if you want build an inside DB (like RIF) then you need other 
skills that just calling API

you can see how it's difficult to make efficent ch

... choices by following the TRETBASE thread

the major drawback is to find an efficient method to build indexed 
files (sorted data)

Rebol is definitly out for this type of job

for some project rebol can use a liittle cross platfrom db

actually, it's the case

but for insert R3 in some real compagny we need to access to all 
other db mssql etc....

but i think it's a shame that rebol doesn't deliver his own DB scheme

Claude all majors DB furnish api to do that , you just have to read 
documentation

i am agree

I think that DB support will be done via plug-ins. Plug-in is more 
formalised kind of DLL interface we have with R2. In fact, DLL interface 
will be probably done as a plug-in. We can then link to odbc for 
e.g. too ...

Steeve? DB scheme? Come on, guys, you must live in wonderland, no? 
So supporting sqlite, mysql, postgress, and others via ODBC is not 
enough?

ahah

for developpement and support R3 need more guys !!!!

but - to proceed in some unified system way, we would need Carl to 
return to Core development, after GUI is done. Without finalisation 
of main interfaces in R3, we still can't do much ...

Pekr, it's not me who made announces on RIF

on peut toujours rêver Claude

Steeve - but you know, what the situation with REBOL is like. So 
you either wait for RIF, or you use something even maybe more suited 
to the job - SQLite. Does everything need to be pure REBOL based?