A small R3 FAQ
26-Oct-2007 12:32 Filed in: REBOL 3
I don't work for RT and don't have full access to the
juicy details of RT's source code for R3 and all the
future plans, but I have some insight that many
people don't have. It makes me feel very special
and important.
What I answer here are observations, not hard facts, since some plan details may be shuffled around as they usually are in such projects.
RT says there will be an alpha available some time after DevBase is completed. This is necessary for the adoption of DevBase among users.
As it has been until recently, only people who are willing to contribute have been allowed in, in order to reduce noise to an absolute minimum. At this moment, that access has been restricted a bit more to focus on DevBase. I would wait until the alpha appears.
No. You have to remember that R3 alpha is indeed alpha. It's fairly stable, but a lot of things are missing, that R2 can do, such as a finished VID3. R3 can process gigabytes of data more efficiently than R2 can, but you'd have to suffer under:
I'm running all my scripts in R2 and won't start migrating anything until after beta.
There was an early time frame given, which may have been treated as official or not. It was cause for much controversy, since it ignited the discussion whether R3 would ever be released or not. The alpha came out on 1st July 2007 and the beta was somewhat scheduled for one month later, 1st August 2007. However after a few days, I saw that as completely unrealistic. With the current speed of development, it's probably fair to say some time in the first quarter of 2008.
This continues to be a moving target. I flat out don't know, but there was an early indication that R3 beta would contain a similar feature set to REBOL/View minus the Desktop. I don't know if that holds up now.
None of the developers have the luxury of being available to work on R3 full time. Not even Carl. That is an unfortunate part of life, as we all have families that from time to time, pull us away from the computer.
As another reason, DevBase is the main focus right now and RT is going 100% forward with that. When DevBase is done, R3 development continues. In the meanwhile, developers are still finding bugs and adding new mezzanines for later inclusion.
You can't be sure from what you see, but from what I can see here, the motivation is pretty damn high. Dropping R3 on the floor now would be an absolutely crazy thing to do, I would have wasted my time blogging about it, and I'd have to find another job, because all the things I develop is technology based on REBOL.
You have to remember also that RT is AFAIK not financed by a large company that have the desire to reap RT of it's technology after R3 is finished or to drive R3 in a particularly narrow direction and away from the public.
RT is also not managed by a technologically clueless executive with no vision, who will cancel R3 development when it's 97% ready to save a few bucks. RT is managed by one of the world's finest software developers. While the marketing skills may leave a lot to be desired, having the main developer at the company helm makes R3 much harder to kill.
Imagine if Carl had developed R3 under Commodore just prior to it's demise. R3 would have been just a memory today, buried under unreasonable licenses and patents, even if it had been hailed as the perfect programming language during its (short) lifetime. That's one of the reasons why, the Amiga is so hard to resurrect.
This is my personal opinion: As long as there is no alpha, there's no reason to sit around and wait for R3 anyway. Some people believe that people leaving the community causes permanent damage to the community and that it in the end will consist of 4 people instead of the current amount of ... 7 or 8 people.
My theory is this: You have seen R2 in action and would be strongly interested in R3, because R2 is great, but there's juuust this one thing missing. Some kind of dealbreaker that prevents you from using R2 for a big project.
But when R3 comes out, you will still be interested in the product and you will come back to see it in full glory. This is why I don't worry about people leaving.
Since you can't have R3 in your hands yet, there's no reason to twiddle your thumbs and if you don't want to work with R2 (which you should, to make it easier to learn R3), there are always other languages to explore, to make you appreciate R3 even more, when you get your hands on it.
I will keep posting information as best as I can, but it would be stupid to sit and wait around.
This is not a good idea right now. First, R3 is still alpha and while it looks stable to us, more bug reports are sure to come in quickly, when a public release is made. This may affect the stability of your product in short term until the bugs are fixed.
Second, the beta date is not known. If it gets pushed further back to, say, winter 2008, even for really good reasons, your boss may not have confidence in RT's deliverability to allow the use of R3 in your company in the future. Please keep quiet about R3 from your boss, until the beta is out, to give him/her a hands on demonstration of R3.
Third, don't plan your personal products around R3. If you are asked by a customer to deliver an R3 product by March 2008, I would say no.
R3 will be entirely free (as in beer). It will be more open (as in open source) than R2. The partial openness is an intricate plan to allow R3 to grow much more than R2 ever has and this collides with R2's licensing scheme of allowing certain features to be Pro only. R3 users would be able to easily circumvent that, so there is no longer a paid Pro scheme. RT will generate revenue in other ways, using technology built on R3.
For an SDK and encapper, the license and price is still unknown.
There will be much more information available on this on Carl's blog, but basically it's R3 source management in an iTunes interface (sans coverflow). It will be a multi-user system
with read-only guest access. You can submit
source changes that skilled developers (i.e.
better than me) will approve for getting into
the official R3 source tree.
DevBase is designed to allow RT to spend as little time as possible working on the open parts of R3. RT wants to focus on the closed core of R3 and let qualified developers work on the open parts.
DevBase is a community tool to allow you, the user of R3, to find a bug and submit it along with a fix, test case and documentation to RT for much quicker insertion into the official source tree. Similarly to how DocBase is updated by you, you can fix bugs in the FTP protocol with DevBase.
A few factoids:
Sure, ask away. Note though that I am limited in my knowledge on other parts than VID3, som graphics and generic console use. I know nothing of ports, tasks or protocols. But if you have code (small bits only, please) that you want to test, I can test it in R3 for you.
What I answer here are observations, not hard facts, since some plan details may be shuffled around as they usually are in such projects.
Will there be an alpha available?
RT says there will be an alpha available some time after DevBase is completed. This is necessary for the adoption of DevBase among users.
Can I get to feel special and important
too? Can I get access to R3 alpha now?
As it has been until recently, only people who are willing to contribute have been allowed in, in order to reduce noise to an absolute minimum. At this moment, that access has been restricted a bit more to focus on DevBase. I would wait until the alpha appears.
Can't I just drop R2 now? It seems pointless to use it.
No. You have to remember that R3 alpha is indeed alpha. It's fairly stable, but a lot of things are missing, that R2 can do, such as a finished VID3. R3 can process gigabytes of data more efficiently than R2 can, but you'd have to suffer under:
- Unfinished and buggy VID3 (only simple styles available, poor text editing)
- Windows only
- Ugly and cumbersome DOS console only
- Untested tasks
- Unfinished testing scheme
- Lots of undocumented features
- No security requesters
- In fact no requesters at all
- Still about 75-100 known bugs left, some are serious crashes.
I'm running all my scripts in R2 and won't start migrating anything until after beta.
When will the beta be available?
There was an early time frame given, which may have been treated as official or not. It was cause for much controversy, since it ignited the discussion whether R3 would ever be released or not. The alpha came out on 1st July 2007 and the beta was somewhat scheduled for one month later, 1st August 2007. However after a few days, I saw that as completely unrealistic. With the current speed of development, it's probably fair to say some time in the first quarter of 2008.
What will the beta contain?
This continues to be a moving target. I flat out don't know, but there was an early indication that R3 beta would contain a similar feature set to REBOL/View minus the Desktop. I don't know if that holds up now.
Why is it going so slow with development?
None of the developers have the luxury of being available to work on R3 full time. Not even Carl. That is an unfortunate part of life, as we all have families that from time to time, pull us away from the computer.
As another reason, DevBase is the main focus right now and RT is going 100% forward with that. When DevBase is done, R3 development continues. In the meanwhile, developers are still finding bugs and adding new mezzanines for later inclusion.
How can I be sure that RT wants to actually finish R3?
You can't be sure from what you see, but from what I can see here, the motivation is pretty damn high. Dropping R3 on the floor now would be an absolutely crazy thing to do, I would have wasted my time blogging about it, and I'd have to find another job, because all the things I develop is technology based on REBOL.
You have to remember also that RT is AFAIK not financed by a large company that have the desire to reap RT of it's technology after R3 is finished or to drive R3 in a particularly narrow direction and away from the public.
RT is also not managed by a technologically clueless executive with no vision, who will cancel R3 development when it's 97% ready to save a few bucks. RT is managed by one of the world's finest software developers. While the marketing skills may leave a lot to be desired, having the main developer at the company helm makes R3 much harder to kill.
Imagine if Carl had developed R3 under Commodore just prior to it's demise. R3 would have been just a memory today, buried under unreasonable licenses and patents, even if it had been hailed as the perfect programming language during its (short) lifetime. That's one of the reasons why, the Amiga is so hard to resurrect.
I'm leaving. I'm tired of waiting.
This is my personal opinion: As long as there is no alpha, there's no reason to sit around and wait for R3 anyway. Some people believe that people leaving the community causes permanent damage to the community and that it in the end will consist of 4 people instead of the current amount of ... 7 or 8 people.
My theory is this: You have seen R2 in action and would be strongly interested in R3, because R2 is great, but there's juuust this one thing missing. Some kind of dealbreaker that prevents you from using R2 for a big project.
But when R3 comes out, you will still be interested in the product and you will come back to see it in full glory. This is why I don't worry about people leaving.
Since you can't have R3 in your hands yet, there's no reason to twiddle your thumbs and if you don't want to work with R2 (which you should, to make it easier to learn R3), there are always other languages to explore, to make you appreciate R3 even more, when you get your hands on it.
I will keep posting information as best as I can, but it would be stupid to sit and wait around.
I told my boss that we're going to use R3 in our next project in december.
This is not a good idea right now. First, R3 is still alpha and while it looks stable to us, more bug reports are sure to come in quickly, when a public release is made. This may affect the stability of your product in short term until the bugs are fixed.
Second, the beta date is not known. If it gets pushed further back to, say, winter 2008, even for really good reasons, your boss may not have confidence in RT's deliverability to allow the use of R3 in your company in the future. Please keep quiet about R3 from your boss, until the beta is out, to give him/her a hands on demonstration of R3.
Third, don't plan your personal products around R3. If you are asked by a customer to deliver an R3 product by March 2008, I would say no.
What will the license be?
R3 will be entirely free (as in beer). It will be more open (as in open source) than R2. The partial openness is an intricate plan to allow R3 to grow much more than R2 ever has and this collides with R2's licensing scheme of allowing certain features to be Pro only. R3 users would be able to easily circumvent that, so there is no longer a paid Pro scheme. RT will generate revenue in other ways, using technology built on R3.
For an SDK and encapper, the license and price is still unknown.
What's this DevBase thing?
There will be much more information available on this on Carl's blog, but basically it's R3 source management in an iTunes interface (sans coverflow
). It will be a multi-user system
with read-only guest access. You can submit
source changes that skilled developers (i.e.
better than me) will approve for getting into
the official R3 source tree.
DevBase is designed to allow RT to spend as little time as possible working on the open parts of R3. RT wants to focus on the closed core of R3 and let qualified developers work on the open parts.
DevBase is a community tool to allow you, the user of R3, to find a bug and submit it along with a fix, test case and documentation to RT for much quicker insertion into the official source tree. Similarly to how DocBase is updated by you, you can fix bugs in the FTP protocol with DevBase.
A few factoids:
- DevBase is currently developed in R2, because VID3 is not yet capable of the things it needs to do.
- DevBase can hold any kind of source. DevBase will early on hold open R3 sources and DevBase sources.
- DevBase uses your favourite editor for code and your favourite separate editor for documentation.
- DevBase is about 40k of source code and uses REBOL/Services for communication.
- DevBase is the first Altissimo application.
Can I pound you some more for questions?
Sure, ask away. Note though that I am limited in my knowledge on other parts than VID3, som graphics and generic console use. I know nothing of ports, tasks or protocols. But if you have code (small bits only, please) that you want to test, I can test it in R3 for you.
|