Comments on: Alpha Release - Status Update

Here is a status report for the current release of REBOL 3.0...

Almost 200 bug reports and enhancement requests have been logged by the alpha team. Of those, 6 bugs have been dismissed and 94 have been fixed and marked as tested; 18 items are enhancement requests. So, while there are still a number of outstanding issues, a lot of things have already been resolved, and we're almost keeping our heads above water. I say "almost" because we decided to delay the beta by a couple weeks while we iron out some of the more important outstanding issues.

You have to keep in mind that large parts of the REBOL kernel are new, or changed to a great extent, so it isn't just minor tweaks we're making here. This is our chance to fix things we know can be improved, at the cost of some compatibility and, most unfortunately, some of the hard-won knowledge the community has acquired over the past seven years.

In order to offset that "experience deficit", we are committed to improving documentation. DocBase is already being used by the alpha team. It is currently based on MediaWiki, so others can contribute, but there are still a lot of open questions about how it will work long term as well.

We also take for granted all the infrastructure in REBOL related to view (VID and graphics), networking, and datatypes. There are big changes in those areas.

For example, with the change from faces to lightweight GOBs, we're starting fresh in many ways. There are a lot of notes underway on what the new VID should support, and a micro-VID dialect already exists as a starting point for discussion. Rich text is functional, Cyphre's old SVG-inspired tiger demo runs, and some testers are focusing their time on the graphics engine to make sure it's stable and fairly complete for the beta.

Schemes will be written differently in R3, with the goal of making it easier for more people to write them. TCP is native and an HTTP scheme has already been written. We need to prioritize scheme development, and may focus on providing documentation about how to write them first; that means the beta probably won't support all the protocols available in R2, but the community will be able to help create them in true open source fashion.

Datatypes, ah... datatypes. They are the lifeblood of REBOL and the blood is flowing fast these days. Some datatypes are going away (e.g. hash!), some are new (e.g. dictionary! and percent!), and some are getting a new lease on life (money! changing from a floating point to a BCD implementation). There are still a number of datatypes in development, and a great deal of testing is being devoted to type compatibility. Consider that there are more than 50 datatypes in REBOL; each needs rules about how to be made, molded, and work with other types via various ops and actions.

In that vein, there are many design discussion on certain subjects. For example, the percent! datatype seems like it would be easy to design, right? Wrong. There have probably been 200 messages exchanged, public and private, about how it should work. Behavior decisions are not easy to make, but they are critical; the time spent discussing these issues is not wasted.

I need to get back to work here, but I wanted to let you know that things are progressing rapidly, though the task is large. We don't expect to go beta on the 15th as originally planned, but we may add some more testers soon. But, you know that "Too many chefs spoil the soup", but "many hands make for a light load" (to mix the metaphors.)

21 Comments

Comments:

Brian Tiffin 6-Jul-2007 19:25:34	Thanks for the update. I may not speak for the majority, but I'd prefer to wait and end up with a better product. In my opinion, take your time. Make the technical core of R3 as robust as possible. Cheers
Louis A. Turk 7-Jul-2007 0:57:54	I agree with Brian, making sure that R3 works properly is the most important thing. But I'm also glad to read about the commitment to improve the documentation---that is the next most important thing. Clear but also complete documentation is essential if we want to see the number of REBOL users skyrocket. When people can easily learn what REBOL can easily do, they will love it and be hooked!
Paul 7-Jul-2007 8:17:30	Sounds great. Thanks for the update.
-pekr- 7-Jul-2007 8:58:16	Month here or there does not matter, if the result will be more feature complete/bugfree release! R3 is inevitable, it is coming, and that is nice. Of course, it would be nice to also prepare on some other front - docs, their translations, "marketing", internet evangelism, etc. Is new "marketing" guy already selected? We could start to exchange some info of how to aproach those things ...
ChrisD 7-Jul-2007 15:57:26	There should be no rush to market. From my little corner of the Mac world, good documentation and truetype font use are imperative. Recently many script languages (like Apple script etc) are starting to appear as clickable applications on the Desktop, and I see incredible potential for R3 in this role since it is cross-platform.
Dave Cope 9-Jul-2007 5:02:07	Echoing the previous posters, thanks for the update. I'd rather wait for a robust R3 then one that is rushed to meet a previous release date. I too will welcome the documentation - that is vital to the equation. Thoughts are with you - keep up the good work!
Luis 9-Jul-2007 5:25:39	Glad the docs are being seen to, but I'd suggest incorporating something in pdf, possibly generated off the mediaWiki? At least for beginners, printing off the Intro and a couple of chapters to have for reading on the go, keeps Rebol 'in their hands' and in their minds. Cheers, Luis.
Greg 9-Jul-2007 11:24:33	To the R3 team: Work hard! Thanks for the update.
Robert Shiplett 9-Jul-2007 17:39:38	With the delay, could the alpha test group expand as fewer issues are thought to be present? Early in alpha it had to be kept quite small ... and a slightly larger group would still be a select group ... unlike a beta
Jack Seay 11-Jul-2007 20:20:59	How about releasing documentation to all Rebol programmers as it is written, so we can be ready to dive in when R3 is released?
-pekr- 12-Jul-2007 0:08:43	I second Jack's opinion :-)
Art 12-Jul-2007 22:01:45	Just want to say thanks for bringing us a great product. It continues to amaze me how you are able to produce this kind of technology almost single-handedly. All I can say is that many have much to admire and learn from you. WOW
Vitaliy Sergeev 13-Jul-2007 10:19:28	Super cool but "single-handedly" technology frighten me cause amply years have passed since rebol was itroduced in rebol.com. Rebol needs more forums more talks more developers
Nicolas V. 13-Jul-2007 17:01:32	Since I discover REBOL in year 2K, I was really impressed and excited... Exactly the same thing with BeOS... The size of REBOL was incredible... REBOL provide fantastic concept... On/Off line internet based program... Exactly the same concept that everybody is talking about now... but with 6 years in advance... I'm waiting for R3... But it is very hard because the "pensée unique" (one way to think) is very strong. Maybe if we associate REBOL with AJAX, Web 2.0, Eclipse, RIA... and some other commercial concepts to make more $ for big company... For me, REBOL is a beautifull thing... The documentation is clear and the design is very good... Everything produced by the REBOL Team is "well thinked"... Except that the REBOL scripts are very hard to read for non REBOLer... or strange, weird... but so powerfull ! Unfortunatly, for me, one thing is missing since year 2K for REBOL : a true WYSIWYG editor for REBOL/View and more generaly a All In One product with RAD concept, test, documentation, help, debug... but it was not the goal of REBOL... Good luck for the RT. Viva la R3bolUTION !
croquemitaine 16-Jul-2007 14:03:07	hello ukiah,after r3 it is mix of ios and altme? Look here,a track to follow certainly: http://www.simplifymedia.com/download.html And ,you know in France, thousands of societies swear by Windev.It isn't a Rebol product,IT SHOULD BE! You can do much better ,it's not a matter of amount of work,matter of choice.
croco 22-Jul-2007 9:31:06	I'd be extremely pleased if it runs on my (UIQ3) phone.
Liviu Antoniu 2-Aug-2007 16:15:11	Where can I get R3? I want to see it.
Henrik 5-Aug-2007 5:31:53	It's currently an internal alpha, not for public distribution. It will be released as beta soon.
Pier Johnson 5-Aug-2007 18:14:31	Carl, thanks! There's a authentic buzz in your written voice, a buzz I have not read from you in awhile. You have lit my jets with your spirited words!
Bo Lechnowsky 19-Aug-2007 13:13:20	Carl, it was nice talking with you the other day, especially about R3. I am looking forward to R3, especially the rich text and GOB components, but many many other parts as well. I also like the large percent of R3 that will be open-source (not that I will have the time to pursue it in the near future, but I have no doubt that the Rebol community will make good use of that aspect which will benefit us all). Also, I don't mind having to update my scripts to work with R3. If it's too much of a hassle, I can always keep R2 running for those scripts.
popper 20-Aug-2007 14:57:12	iv said it before and ill say it ago, you guys are too tech (and thats a good thing) but to attract the masses someof you 3rd party guys are going to have to make some basic GUI apps that can be shown as fully working to these masses. that way thay can see first hand Rebol does make a great GUI app and they can modify it as they please. 1: make a GUI builder that outputs fully working Rebol code that you can just run and work. there is a script or two by carl that shows GUI grag&drop, use that as a base if needs be. 2: make several fully working GUI apps that have skins already made that users can drop in a dir and have these apps pick it up and look good. menu-system-demo.r can be the generic code to work from and build out perhaps ?, that is always impressive for first time showing to end users, it needs more work to use as a base for all these new apps that users can be shown. 3: Video Encoding is just as strong today, if not more so since Divx and the latest x264 codec started taking off, so make at least ONE SINGLE GUI app that can use interact with the shell Mcoder/VLC/ffmeg Encoders and make it plug-in based and user skinable to really impress these new users. these 3 things are what is going to impress users and in doing so will make Rebol3 used by the masses and perhaps taken up by them as word of mouth and GUI function moves them. if rebol3 is suposed to be easy to interdace with the shell/cli apps today than prove it by at least some everyday AVC/H264/x264 and divx Encoding/streaming working GUI apps and docs to hel your average newby learn and modify these scripts to re-create their own new GUI apps with that rebol GUI builder ;) (that doesnt exist yet). as already said, so much times passed already so why not hold back a little longer and produce something even beter. my pet wish as always LOL, and put that last little Multicast tweak in the new codebase that make your average newby want to expand the *fully working example* multicasting tunnel code that just works and create far more new innovative apps than we have now.

Post a Comment:

You can post a comment here. Keep it on-topic.