World: r3wp

inside the function you do bind/copy blk 'local-var

local-var is defined but the other ones, tag-title, tag-x, tag-y 
... I don't want to have to define in /local local-var ...

the bind should probably give an error for variables that are not 
set

All referenced variables would include ones like 'for, 'if and 'and. 
It seems to me that you want an object, not function local variables. 
That would be much easier to do in R3 with the improvements to bind 
but you can do it in R2 as well.

what I am after is in using rebol as a PFL (pure functional language), 
 in which all references are copies. I read something in the wiki 
about a new function keyword which allowed this but couldn't find 
it again in the wiki . Anybody knows this func2 keyword ?

No such thing. You would have to rewrite most of the mezzanines and 
natives to make a practical pure functional REBOL, as almost everything 
causes side effects and aliases. Pure functional languages tend to 
only be efficient when compiled - when interpreted they are dirt 
slow.

You can manually use REBOL as a pure functional language, but it 
will be much slower and have more overhead.

On the other hand, there should be nothing stopping you from making 
a pure functional dialect that compiled to imperative REBOL, and 
the result might be faster than all but the best hand-coded REBOL, 
once you get past the compiler overhead. If you can compile ahead 
of time that won't be much of a problem.

thanks, I am looking for a practical solution that doesn't require 
much work. I've moved away from rebol but I have a lot of legacy 
code i'd like to evolve, if possible

Well, code that is written with aliasing issues in mind tends to 
look pretty much like mostly pure functional code, and function result 
chaining gets rid of a lot of variables. We have been trying to make 
functional programming styles easier in R3 and some of that is getting 
backported to R2 starting with 2.7.6.

Many of the series functions are modifying though, and reassignment 
is used a lot.

is there any doc on the functional programming styles in rebol ? 
thanks

It's 20 days since that last update to Docbase; is that a good sign 
or bad?

not much to write right now

Plenty going on

New Core elements to support VID, and REBOL/Services being ported 
to R3

Peter; It's bad in a non-bad kinda way.  REBOL 2 seems to be keeping 
quite a few rebols in pay checks.  That's good for coders but bad 
for ongoing public buzz.  Buzzing about a loved hobby is completely 
different than buzzing about (still loved) work.


Gabriele did me a huge favour getting the REBOL in mediawiki extension 
plugged in and it sits waiting.  There is still work to get DocBase 
REBOLized, even if there is a slow down in the "fast pace, visible" 
R3 developments.  DocBase is really for us, the community, so it 
goes in spurts.   Next step (6 months later) is to make it so posting 
up MakeDoc is simple and easy, keeping in the REBOL headspace, including 
the document markup.

Ok, it's a joke we'll prove wrong.
What bet can't be won?
The alpha bet.

Gabriele today posted, that basic rebservices client is working with 
R3, being able to connect to DevBase. I think it is a good milestone 
....

I see Carl has now rewritten VID for R3

Hello. I want to start working on upgrading some code from R2 to 
R3. The only public alpha release is still far the beta release because 
some basic features changes or are not implemented. The last alpha 
releases seems much more interresting for studying R2 - R3 migration. 
Somebody knows when another public version will be released ?

Hello Karim - we can't say yet. But - what is your main area of usage? 
Core features? or GUI?

In fact both. I worked on a multi-user contact management application 
that connects to a SQL database (mysql and sql-server). There are 
some main areas I want to investigate into : DB access (Trebase could 
be a sexy alternative to other db engine), GUI and business logic 
processing. In the last area, I think that rewrite some code in R3 
can produce more elegant and more efficient program. I'm one of the 
guy who used hash! datatype ;-)

OK I also have to admit that I cannot wait to play with a beta release.

Karim, then I think it is a bit preliminary to consider porting:


- mysql, postgress - while those schemes are built upon raw TCP R2 
stack as a schemes, someone would have to port them to async R3 networking 
kernel

- sqlite - requires mostly DLL access. There is no DLL access in 
R3 and maybe there will not be one like we know it in R2. Carl is 
not decided yet, if we go via DLL layer like in R2, or via standardised 
plug-in interface, which is not ready though, as it needs modules, 
which are not ready either

- GUI - Gabriele was working on new VID. It is now called VID 3.3 
(third prototype). It is still not complete. But - Carl is trying 
to follow slightly different obectives, so he decided to rewrite 
it and bring us VID 3.4 prototype (probably called GIDI)

- View, unless Cyphre fixes some harsh bugs in its kernel, is pretty 
much buggy and makes VID3.3 crash in few secs to minutes.

That is the reality. I will try to talk to Carl, so maybe we will 
be able to plan on next release and define featureset, which gets 
completed and released.

OK. Thank you for the informations. I think the release notes of 
next public version needs to define what can change and what's will 
remain quite the same until the beta or final release. Then it will 
be easier tu us to know on what we can focus.

Assume that much of it can change, though most of what has been backported 
to R2.7.6 is not likely to change much.

Karim .. do you have some screenshots or a demo of your CRM?

Does anyone know anything internal about R3 development? I am really 
disappointed, we are down to 0 - 2 persons on r3-alpha channel per 
day. That is really disappointing ...

I noticed that too.  Creepy quiet.  I don't think it's disappointing 
(yet) ... feels more like a calm before the storm.

can't chat and code at the same time, I think. ;-)

hey, there's even a comment on the wildfire blog entry from Carl 
that says: "Sorry that I do not have time to do more -- too busy 
with work (REBOL Tech)."

In the meantime I re-read port documentation, and try to study http 
scheme, but this is hardly usefull for the community, as I am not 
the one able to bring in e.g. ftp or other scheme anyway :-)

suits me well. won't have time for R3 work until July anyway. :-)

I have 75% of a dict:// scheme written.  (First thing I did when 
I got in on the alpha.  Took about 6 hours.  Then I got confused 
on the new I/O completion rules.  Haven't looked at it since.)  Soon. 
 I use the R2 dict-demo all the time now.  Nothing beats using the 
REBOL console as a dictionary/thesaurus.  From using KDE, there are 
all kinds of schemes on the wishlist.

Well, Carl effectively killed Gabriele's VID3 by saying he was changing 
it ... and so we wait.

The bad thing is, that while we could have more than one UI, Gabriele 
does not seem interested in finishing it, and noone apart him is 
probably technically skilled to finish it ....

There's no point in finishing VID3 if VID3.4 functionality strongly 
overlaps, so we'll have to wait and see whether it does. If it does 
not, then VID3 can be finished as a separate project.

That assumes that the technical ability (of which there is lots) 
and the programmer time (of which there is little) is available.

not just that, it usually has to be the same person starting it and 
finishing it. that's why many open source projects fail. :-)

What happened to Cyphre?  I thouht he was involved in some of that 
stuff?  He always had the coolest REBOL View apps.

He's working on QTask.

He works with Reichart?

yes, I believe he does.

ahhh, I didn't know that.

Gabriele has worked on QTask as well, but I don't know if he's done 
with his stuff yet.

So who all is working on R3 right now?  What does the team consist 
of?

I'm pretty sure Carl is very deep into VID design now

So VID is the last major hurdle before release?