Mailing List Archive: 49091 messages
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

[REBOL] Re: lost the case ...

From: maximo:meteorstudios at: 4-Nov-2003 21:15

> -----Original Message----- > From: bryan [mailto:[bry--itnisk--com]] > Sent: Tuesday, November 04, 2003 8:27 AM > To: [rebol-list--rebol--com] > Subject: [REBOL] Re: lost the case ... > > It would be nice to see some group projects. I would like to see a > project to put together a real library of xml functionality for Rebol, > something which I think might be too much for one or two to accomplish > realistically. >
I'll butt in here, as I am one who is trying to create a concerted development effort. why is rebol not as widespread as it should? its scary and easy to put down for IT managers who often have less knowledge about IT than about their dog's digestive proteins. All they want is a sense of security. rebol does not provide that. python does, perl does, .net does. what mega corporation claims and is proud to say that it uses ios or even rebol/core? ......... none I heard of. python uses google... and that is a powerfull statement if any. Problem is that many of us don't use rebol as our main development package, working a lot of our stuff at home in a hobby-like environment, so many projects are stale or move on only slowly. France might be the one exception to this, I hope it succeeds in becoming a beacon which light up the rest of the world to rebol. I have used rebol commercially, but when projects aren't comming in, I have less justification to put time on my tools. Take the steel project, for example, its always advancing. But I'm currently only putting a few hours on it every week, usually that's not enough to call it an update. Documentation also takes soooo much time, that it slows down development a lot. I started steel in the mindset that people might join me with specific tools which all complement themselves (my main contribution being glass). Steel is a core to which anything can be grafted, but I need to actually get that core to a release state and all that is quite tedious. releasing QAD code tidbits is easy, but implementing liquid-vid, for example, means that I must re-engineer 50 types of gadgets (30 are done ;-), make them stable, some quite hard to taim (even for the most advanced of us), in their current form... And that's just one of 20 pieces of the puzzle. I've shifted the development of steel so that low-level constructs are done which are released, which can be used by anyone. This is instead of making the code editor, which is what everyone expects, which would be rebuilt every few months, making any of your projects corrupt... lets take GLASS. I worked on it almost full time for 2 months. Man, it has its own dialect, complete dynamic nested layout, it supports run-time switchable localization (which even supports multiple datatypes, not just text), run-time skin swaping, all in a pretty fast and REALLY easy to use package... but there you go, I lost that job (because the company got stabbed in the back) and glass development all but stopped (at least I was officialy given the rights to all the code :-). I know have an advanced experimental layout engine, "in the closet", ready to be be put into an official release state (read as: completely re-coded), a complete and true VID alternative which goes beyond anything any of you have seen in rebol, but, again ... time is against me, steel must be done before I start again on glass. What's the point in releasing code if no one's going to use it cause its not usable enough, even if already great... liquid is a different issue... it is release stable, everyone seems to like what they hear of it, but I have not gotten a lot of feedback of people actualy trying it out... and trying to grasp its subtleties. Liquid.r really is a core tool which is meant to be used as a data flow architecture, also as the basis to create other tools, which plug into that data. Liquid-vid is a working example of such a thing, liquid-net and liquid-sql would be others... Do you have time to learn it, and code a module for it? probably not, unless you think that its advantages will outweigh the time spent on learning it, but how can I explain/convince people without real examples!? again, time. And I understand. I have a release plan which is pretty immutable. everything will be there in time, but sometimes I have so much work at home, that I can't do 1/10th of what I wish I'd do. The only thing that is certain is that steel is a project built on solid foundations with a tenacious and stubborn person at its head. I WILL see it through. It is a project 10 years in the making, and When I started steel in rebol, I though about it long and hard... and I decided that if I was going to start, it had to be finished, even if it would take years. I want all of those years designing and implementing obscure ideas on my old and trusty A1200(in language E), to be of some worth. If steel does not get to a certain point, it will be like I've been working hundreds of hours for absolutely nothing. ON ANOTHER NOTE: rebol.org is a great team effort. All of you please use it when you can. Also please give them feedback. I know how fun it is to get questions and even criticism... Nothing is worse than NOT knowing what people think. cheers! -MAx ------------- Steel project coordinator http://www.rebol.it/~steel