REBOL 3.0 mentioned on the blog!
[1/20] from: petr::krenzelok::trz::cz at: 10-Mar-2006 10:19
Hi guys, hopefully good times ahead for us rebollers! Just go and read "Time for big changes" blog article: http://www.rebol.net/article/0259.html Cheers, -pekr-
[2/20] from: fergus4:bellatlantic at: 10-Mar-2006 8:37
Any idea what kind of changes we will be seeing... Petr Krenzelok wrote:
[3/20] from: petr::krenzelok::trz::cz at: 10-Mar-2006 14:54
Alan Macleod wrote:
> Any idea what kind of changes we will be seeing... >
well, not yet, but I expect REBOL 3.x to introduce incompatibilities, but for the good. I expect it to: - have multiasking/threading high concurency, maybe message based, dunno - have modules support - being just one .exe, the rest as dynamically loadable components - some parts even - as for gfx, sound, will support plug-in architecture (for codecs e.g.) - RT will keep Core (kernel) of the language and the rest will be open-sourced to some extent - other features already discussed - RIF, Rebin, language plug-ins for language extensibility - I heard also View will be overhauled - faster blitting, new compositing engine, better event system Simply put - all things we wished for in the last few years will hopefully come. If RT is concentrated, I expect Core 3.0 alpha till the end of the year personally, but that is just my guess based upon how quickly was Contra (1.0 to 2.0 rewrite) brought to life - it was actually few months IIRC. Now language as a technology is already pretty much in good state, so I expect RT "just" putting it into more robust surrounding framework .... Once again - that is just my opinion .... -pekr-
[4/20] from: hallvard::ystad::gmail::com at: 11-Mar-2006 20:46
On 10/03/06, Petr Krenzelok <petr.krenzelok-trz.cz> wrote:
> > well, not yet, but I expect REBOL 3.x to introduce incompatibilities, > but for the good. I expect it to: > > [...] > Simply put - all things we wished for in the last few years will > hopefully come.
[5/20] from: gchillemi::aliceposta::it at: 12-Mar-2006 11:05
>> Simply put - all things we wished for in the last few years will >> hopefully come. > Unicode?
Multitasking ? 3D support ? NiceUI ? Good development enviromnent ? I have never had the needing for those features. I can still live with View 1.3.2 ;-) Giuseppe Chillemi
[6/20] from: SunandaDH::aol::com at: 12-Mar-2006 5:49
> Simply put - all things we wished for in the last few years will > hopefully come.
The main thing I wish for is compatibility -- some of us don't want to have to rewrite code because the old code breaks the new interpreter. It is not efficient to keep having to reformat or rejig data from one version of REBOL to another. Same as it is not efficient in the more physical world to have to unload goods from a truck to repack them in a train or boat; that sort of stuff is early 20th Century, If code is just data, and data is everything, I ought to be able to stick it in the equivalent of a TEU and have it go anywhere in the REBOLsphere. Sunanda.
[7/20] from: massung:g:mail at: 12-Mar-2006 9:40
I agree. REBOL, interestingly enough, is the first language I use where I don't feel a compelling need for syntax highlighting or some uber-IDE to program in. The desktop and console work just fine. As per 3D support, I would personally like to keep REBOL as small as possible. I understand that I'm in the "noob" crowd still, but that is one of REBOL's most appealing features, IMO. The fact that I could actual read about (and retain 90%) of the language in 24 hours was a very pleasant experience. I'd just leave it to others to create game engines or such as dynamic libraries and import them into REBOL using a wrapper. I think multi-threading would be nice, but it doesn't need to be over-the-top. Just a way to execute a block in another thread, and return a thread handle. Something as simple as:
>> foo: make thread! [ forever [ print "Hello, world!" ] ]
== Thread Handle: 0x02020 The list of functions needed would be fairly small: thread? kill pause resume and perhaps a way of setting priority. One example of where this is useful would be opening multiple consoles from the same environment. So far, I haven't found a way to run the desktop and the console at the same time. Hardly a problem worth adding a feature over, though. REBOL is so small that opening 10 instances isn't a big deal. Still, it could be useful. Most of what I'd like to see involve little, minute additions to the language that, while I could simply write the mezzanine functions myself, would be worthy additions. Some kind of MAP function is my big one at the moment. The one thing I can say I miss from the VID is the ability to right justify controls, but there's probably a way to do it already that I'm missing. My 2 cents. Jeff M. -- massung-gmail.com On 3/12/06, Giuseppe Chillemi <gchillemi-aliceposta.it> wrote:
[8/20] from: massung:gm:ail at: 12-Mar-2006 9:44
Something I forgot to mention, though, about multi-threading. This is potentially a gigantic can of worms to open. If we suddenly had to start worrying about critical sections, mutexes, locking global variables, etc, then I'd rather not go down that road. The language could handle most of this for us, but those features would unnecessarily slow down the language by default, just by needing to handle the "worst case" senario. KISS. Jeff M. -- massung-gmail.com On 3/12/06, Jeff Massung <massung-gmail.com> wrote:
[9/20] from: greggirwin::mindspring::com at: 12-Mar-2006 10:23
Hi Jeff, JM> I agree. REBOL, interestingly enough, is the first language I use where I JM> don't feel a compelling need for syntax highlighting or some uber-IDE to JM> program in. The desktop and console work just fine. I came to REBOL from VB, and many MS BASICs before that, all of which are IDE based. I think the IDE thing has been part of why other languages haven't grabbed me, but with REBOL, I'm OK. That said, I think some higher level "environments" will be very nice for building larger projects, and domain-specific tools will be even better. JM> I think multi-threading would be nice, but it doesn't need to be JM> over-the-top. Just a way to execute a block in another thread, and JM> return a thread handle. Something as simple as: What about something like tasks in Erlang, or Vats in E? I don't know what Carl has in mind, but I think there are some workable models out there. JM> The one thing I can say I miss from the VID is the ability to right JM> justify controls, but there's probably a way to do it already that JM> I'm missing. There are some resizing and other layout toolkits that can help with this (assuming you're talking about setting buttons and such against the right edge of a layout). If you're just talking about right justifying text, simply use the RIGHT facet for the face. -- Gregg
[10/20] from: lmecir:mbox:vol:cz at: 12-Mar-2006 22:36
Jeff Massung napsal(a): ...
>Most of what I'd like to see involve little, minute additions to the >language that, while I could simply write the mezzanine functions myself, >would be worthy additions. Some kind of MAP function is my big one at the >moment. >
... I may not understand your wish, but e.g. http://www.fm.vslib.cz/~ladislav/rebol/map.r proves, that it is trivial to define. If you feel a need to "enforce"/"make standard" one specific version, then it may be a little problem, but I do not see it as a big deal. (Why do you think such a "simple" function has to be "official"?) -L
[11/20] from: nick::guitarz::org at: 12-Mar-2006 13:52
What attracted me to Rebol was that you could just do so many things immediately, without a long learning curve. The fact that it's not limited in the same ways as other simple-to-use tools, is what makes it useful beyond the initial learning curve. We all appreciate that Rebol works easily with Internet protocols, so many data types, GUIs and graphics, etc. - all natively. Those things are neatly wrapped up in the language so that you don't have to work at a low level, or use an OO approach with lots of libraries and layers - simple things are done simply . The main thing I'm hoping is for that trend to continue - for RT to continue to wrap up more things that _have_ to be done at a low level at this point, using the OS directly or some other development tool. Keep the language interface straightforward, consistent, and useful. Make the current version work across more modern platforms and keep the distribution as small as possible - that's what attracted us all to Rebol in the first place, and what keeps us devoted to it. It's powerful now. I don't think we need any new paradigms, just new native abililities. If something is _not possible_ in Rebol at this point _without accessing native OS calls_, then those are the things that I hope are added. Most of those things are related to hardware control: better multiplatform support for sound, CD music, maybe a simple native multiplatform interface to webcams, etc. Those are things I could really use, and which would make Rebol ultimately useful. Anything that can be implemented in Rebol - as it exists - doesn't need to be added by RT. Protocol extensions can be the responsibility of the community. Everyone has different needs and desires, and it's clear that Rebol can't stay small if everyone's wish is granted. Doc Kimbel's MySQL interface, Gabriele Santilli's PDF dialect, Christian Ensel's menu system, Andrew Hoadley's 3d engine, Oldes' flash dialect, etc. are so useful, but those things don't need to be part of the interpreter. Clearly, they were added as third party creations. If it's possible to implement common video codecs in Rebol as it exists, for example, then that can be the responsibility of the community. If it's not _possible_ to implement a video codec in Rebol, then we need native tools in Rebol to make it possible. Speed up the things that need to be sped up, provide access to the hardware in a way that makes it possible, etc. If RT wants to provide that sort of thing modularly, that's great, but not at the expense of a small minimum distribution (of course, those who are happy with 1.3 could just continue to use 1.3...). Beyond that, I hope RT really works at marketing. A bigger community means a more productive and capable language via more third party tools. That's where Rebol's really lacking. No one uses it. Carl and many of the Rebol gurus are design geniuses - my sense is that they should devote more of their energy and vision to promoting the superior design of Rebol, and making it commercially acceptable to use it! Success in that area would be success for everyone involved :) We use language as an interface to the machine's hardware and OS. Rebol lets us control many of those things natively, with a language design that's brilliant. RT implemented GUI and graphics beautifully. They added things like serial I/O natively. We need more native abilities like those - anything that has to be done directly with the OS and/or some other tool. I don't see anything wrong with the language design - I only see a few capabilities missing - continuing the trend so that Rebol isn't incapable in certain areas will attract more people to use Rebol, and will continue to make it more productive for all of us. Quoting Giuseppe Chillemi <gchillemi-aliceposta.it>:
[12/20] from: gabriele::colellachiara::com at: 12-Mar-2006 23:56
Hi Jeff, On Sunday, March 12, 2006, 4:44:59 PM, you wrote: JM> Something I forgot to mention, though, about multi-threading. This is JM> potentially a gigantic can of worms to open. If we suddenly had to start JM> worrying about critical sections, mutexes, locking global variables, etc, JM> then I'd rather not go down that road. The language could handle most of JM> this for us, but those features would unnecessarily slow down the language JM> by default, just by needing to handle the "worst case" senario. Our plan is indeed to add multitasking, not multithreading. That is, no (as far as possible) shared data; communication via a port (if possible, without copy, so that it is very fast). We don't yet know what the compromises will be - but we don't want synchronization headaches in REBOL. Easy rule: if it's not simple, it's not REBOL. ;) Regards, Gabriele. -- Gabriele Santilli <gabriele-rebol.com> --- http://www.rebol.com/ Colella Chiara software division --- http://www.colellachiara.com/
[13/20] from: carl::cybercraft::co::nz at: 13-Mar-2006 22:56
On Sunday, 12-March-2006 at 13:52:45 Nick Antonaccio wrote,
>Beyond that, I hope RT really works at marketing. A bigger community means a >more productive and capable language via more third party tools. That's
<<quoted lines omitted: 6>>>acceptable to use it! Success in that area would be success for everyone >involved :)
I agree. A re-write of REBOL to add new features would be nice if it's a success, but I see no reason to suspect it will be, given the amount of unfinished REBOL stuff that's about. The Mac View version is still "pending" on their website, though in alpha or beta if you look deeply. And REBOL Services? It's still in alpha or beta too I believe, as is Rebcode. And the REBOL plugin is still only on IE. I'm worried that while they're trying to make REBOL perfect, the world's leaving them far behind. I wasn't going to comment on this thread, but then I read this from Dave Winer... http://www.scripting.com/2006/03/13.html#whyIWillStopBlogging ---8<--- When I started blogging, depending on how you look at it, either in 1994, 1996 or 1997, I had different goals, and happily the goals have been accomplished. Billions of Websites now no longer seems an outrageously ambitious goal. We're pretty close to a billion, I suspect. The goal was also to create tools that would make it easy for everyone to have a site, and then more specifically a chronological one. That's done. I wanted programming to turn upside down, to have the Internet be the platform instead of Microsoft and Apple. That worked too. APIs on web apps are now commonplace, and a basis for comparison between offerings. While user interfaces have gotten better, of course, there's been a steady flow of new ideas in how my work connects with yours, and vice versa, and we're doing it without a platform vendor controlling it. ---8<--- That second paragraph's the important one. REBOL should be part of it and would be way better than the horrible mix of stuff used to put an AJAX site together. It isn't though, and I don't see how improving the core language will make it so. At the stage it is now, it needs two (or perhaps three) things done: To have the plugin available for all major browsers and platforms and for it to be promoted! (The third thing is going fully open-source, but I'm unsure of just how much difference that would make.) REBOL's promotion now is all "come to me". "Join AltME" instead of "plug this into your existing website". Carl has a blog, but there's no public commenting allowed, so no conversation takes place around it. You can just leave a message, which no one except those at RT will see. The traffic's all going the wrong way, a black hole instead of a sun. Hee, and to prove my point, I've just searched Technorati. Just one REBOL hit in the last three days, and this is it... I just killed a couple of days reading Stevey's Drunken Blog Rants, which is a fascinating account of a talented engineer spending a year searching for better languages, and finding Lisp in the end. Lots of interesting insights -- I found his essays on practicing programming and math every day to be particularly inspiring, since I've already found Lisp, and Ruby, and OCaml, and Erlang, and pretty much all of the other languages that are mentioned. Except Rebol. That one was new. And having posted this, how do I link to it on the net if I want to point someone else to it? Wait until it appears in the mailing-list archive, right? Trouble is, that doesn't happen instantly - and sometimes not for a day or more. Which is usually after you've moved on to something else. Which is no way to promote anything. Umm - rant over. :-) -- Carl Read.
[14/20] from: petr:krenzelok:trz:cz at: 14-Mar-2006 0:49
Carl Read wrote:
>On Sunday, 12-March-2006 at 13:52:45 Nick Antonaccio wrote, >>Beyond that, I hope RT really works at marketing. A bigger community means a
<<quoted lines omitted: 13>>>I wasn't going to comment on this thread, but then I read this from Dave Winer... >http://www.scripting.com/2006/03/13.html#whyIWillStopBlogging
And I don't agree at all guys. First - if you would still insist on original rebol, you would be 10 times or more slower, had some features which showed as not so important later, respectively speed gains from different implementation was right. Carl did reimplementation in few months and we got 2.0 branch. Incompatibilities? So what? It is still rebol and imo it is normal. Look at Perl 5 to 6 and I don't expect Rebol to be so much incompatible ... However - in my opinion there was chosen wrong path for rebol from the very beginning - it should be async, tasking, multimedia ready from the very beginning. You know - we've got async for a while, but it showed not being so much stable. And Carl clearly stated that some things are not possible without slightly different implementation. I don't understand you at all - you can get much better lower level implementation and you are scared? those things don't necessarily mean your rebol code will break every 2 lines ;-) It is imo like inserting rebol language into more powerful underlying framework. My vote for Rebol 3.0 - definitely - no more compromises ... -pekr-
[15/20] from: tim-johnsons:web at: 13-Mar-2006 15:18
* Carl Read <carl-cybercraft.co.nz> [060313 13:17]: This thread seems to have become about restrospection and expectation. So here's MTCW:
> I agree. A re-write of REBOL to add new features would be nice if > it's a success, but I see no reason to suspect it will be, given the > amount of unfinished REBOL stuff that's about.
As in the long-lived binary skip bug... is it still there after 5 or 6 years? As to retrospection: ------------------------- I began learning rebol in 2000, after 11 years of C programming, along with some Assembler and VBA. I also began learning python at the same time, as my assignment was to provide semester 1 and 2 of a learn-to-program cirriculum for a now-defunct online high school. I choose for the first semester rebol because 1)I had heard of it 'cuz my niece was working for RT at the time. 2)I like the easy help mechanism and absense of system calls for first-times I choose python for the second semester for 1)it's higher level of discipline 2)and greater level of usage. Since that time I've programmed about equally in rebol and in python, and I feel that each inform the other. I think that rebol is about 50% more efficient than python, line-for-line, but I have also found that python scales better and is easier for me to maintain. Python has more features available from the download community, but rebol is growing in leaps and bounds thanks to rebol.org, which in my opinion is one of the great redemptions of this PL. Further redemptions: I use Andrew Martin's html dialect along with Doc Kimbel's mysql protocol, they are my bread-and-butter and are *far* more efficient to use than comparable python resources. On the other hand, there are many features in python that I like, but many of them can be emulated by rebol (like built-in special methods and selective error handling ) and the concept of enforcing *rules* with rebol - as opposed to python's strict engineering syntax, can make rebol more scalable. This should be explored more, methinks. But here's the kicker: the Expectation part:
> The Mac View version > is still "pending" on their website, though in alpha or beta if you > look deeply. And REBOL Services? It's still in alpha or beta too I > believe, as is Rebcode. And the REBOL plugin is still only on IE.
> I'm worried that while they're trying to make REBOL perfect, the world's leaving them far behind.
[16/20] from: carl:cybercraft at: 14-Mar-2006 7:01
On Tuesday, 14-March-2006 at 0:49:03 Petr Krenzelok wrote,
>I don't understand you at all - you can get much better lower level >implementation and you are scared? those things don't necessarily mean >your rebol code will break every 2 lines ;-) It is imo like inserting >rebol language into more powerful underlying framework.
I'm not scared by any minor inconsistencies with previous versions, but that we'll just enter another wait and see mode. I'm not talking about the language, but its promotion. There's zero buzz about REBOL, despite it being useful enough now. There's no point in writing something for the plugin when it won't work on any browser other than IE or any OS other than Windows, but at least it's very good for CGI work. This doesn't seem to be known about though, as hosting services that actively support REBOL are next to non-existant. You'll get PHP, perl, python, Ruby on Rails and so on, but no REBOL. Compare the results of a search for "ruby on rails hosting" to one for "rebol hosting", then ask yourself, which has been around the longer, Ruby on Rails or REBOL? If we'd managed to get hosting companies interested in supporting REBOL it might've been more popular now, but we didn't, and I think that window of opportunity is now gone. The buzz now is for online, in-browser apps, with the leading contender for dominance being AJAX. (Which just uses a bunch of existing tech.) The REBOL plugin could be a contender in that space, if only it was on all the popular browsers. Why will the proposed new features of REBOL 3 bring in more users? What will be its market? -- Carl.
[17/20] from: Paavo:Nevalainen:saunalahti:fi at: 14-Mar-2006 9:11
At 07:02 AM 3/14/2006, Carl Read wrote:
>The buzz now is for online, in-browser apps, with the leading >contender for dominance being AJAX. (Which just uses a bunch of >existing tech.) The REBOL plugin could be a contender in that >space, if only it was on all the popular browsers.
I guess getting new plug-ins is quite a heavy process. Would like to see more plug-ins, though. Paavo N.
[18/20] from: petr:krenzelok:trz:cz at: 14-Mar-2006 10:37
> If we'd managed to get hosting companies interested in supporting REBOL it might've been more popular now, but we didn't, and I think that window of opportunity is now gone. >
If windows of opportunities is gone according to you, then why to be afraid of some changes? RT can only surpirse us, no? :-)
> The buzz now is for online, in-browser apps, with the leading contender for dominance being AJAX. (Which just uses a bunch of existing tech.) The REBOL plugin could be a contender in that space, if only it was on all the popular browsers. > > Why will the proposed new features of REBOL 3 bring in more users? What will be its market? >
What is its current market? Look - it is clear now, that Rebol 3 will support some module/component architecture and that some of its parts will even be open-sourced. That may speed-up some things and get us rebol deployed into more real-life end-solutions. What is wrong with my-task: make task! [some-code-here]? It could as well be just called Core 2.8, but the changes will be so big, that it deserves to be called 3.0. That is of course my speculation, but from what was told during the last few years, Rebol 3.0 will give us features we wanted for the long time already! And guys - who spreads the View OS-X non being available? From last I heard, even if beta, it seems being stable enough, that e.g. Carl uses AtlME on OS-X machine IIRC. Well, of course there are some things unfinished, but if RT is focused, why to be scared of wait-and-see mode? Core 2.0 was brought in 6 or so months? Or even less? Don't forget please, that Rebol as a language is already here. We are mostly talking new framework it is being deployed into. Tasking, probably new event system, open interface to language extensibility, rewritten View, ready for plug-ins etc etc. I think we should wait for Carl to tell us more technical details, but from what I heard, there will be more than one dedicated coder working on Rebol 3.0 full-time ;-) I hope it is not just noise but actually truth .... and if so, I expect first beta to be awailable at the end of the year, and it is imo the pessimistic variant! And if Rebol 3.0 underlying architecture opens rebol for the community to help with some open sourced components, if it allows better async, tasking behavior, new things, which are not even easily possible now, will come! And if so, it is imo worth it. As for browser plug-ins - just give me ONE C coder with a bit of free time. I gathered MOST of the info into plug-ins Altme group, and I can repost it here (maybe I even did, do not remember now), about how to get NS compatible plug-in implemented. It would cover Mozilla/FF, Netscape, Opera at least. And it is imo not more than wrapping some 20 - 30 funcs at max. It can be done, but RT wanted us to help here. It is us who was not able to do it yet ... -pekr-
[19/20] from: tim-johnsons:web at: 14-Mar-2006 10:41
* Carl Read <carl-cybercraft.co.nz> [060313 20:05]:
> I'm not scared by any minor inconsistencies with previous versions, > but that we'll just enter another wait and see mode. I'm not talking > about the language, but its promotion. There's zero buzz about REBOL, > despite it being useful enough now.
Hugely useful. It runs rings around python! See my previous post. Sell "freedom". Rebol is the most adaptable language that I have ever found and that can be very appealing to programmers who like freedom.
> There's no point in writing something for the plugin when it won't > work on any browser other than IE or any OS other than Windows, but at
<<quoted lines omitted: 6>>> If we'd managed to get hosting companies interested in supporting > REBOL it might've been more popular now, but we didn't, and I think
Rebol's advantage is that it is very easy for a sysadmin to install. It is like the "sqlite" of languages. and BTW: I think sqlite and rebol are potentially a very good fit for small projects. I haven't yet tried btn-sqlite, but plan to soon.
> that window of opportunity is now gone.
Nah! Always will be new technology and the opportunity to "open" a "new window".
> The buzz now is for online, in-browser apps, with the leading > contender for dominance being AJAX. (Which just uses a bunch of > existing tech.) The REBOL plugin could be a contender in that space, > if only it was on all the popular browsers.
Bingo! (see above).
> Why will the proposed new features of REBOL 3 bring in more users? What will be its market?
Good point. tim -- Tim Johnson <tim-johnsons-web.com> http://www.alaska-internet-solutions.com
[20/20] from: charles:mougel:spinodo at: 14-Mar-2006 21:00
Tim Johnson a =E9crit :
>>The buzz now is for online, in-browser apps, with the leading >>contender for dominance being AJAX. (Which just uses a bunch of
<<quoted lines omitted: 3>>>> >Bingo! (see above).
I'm really agree with that. VID is cool, for layout description. But, a dialect for web-apps in "AJAX", for rebol in cgi will be a good thing. A webpage, rebol-generated or not, and call to a server-deamon, written in rebol. A dialect for describing web-dynamic-interface, as easy as in VID ! Here is what I dream now. Charles.
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted