[REBOL] Re: oss revisited (real world examples)
From: tim::johnsons-web::com at: 9-Feb-2004 11:11
> Tom:
> > My wishful thinking forces me to bring up what may be a dead horse
> > issue, so I apologise in advance. What's up with rebol and open
> > source these days? Rsharp seems stalled.
>
* [SunandaDH--aol--com] <[SunandaDH--aol--com]> [040209 08:54]:
> Just a few comments to throw into a general discussion.
Good writing here Sunanda!
> Open source is not the only model for software distribution. It is a very
> trendy one
And sometimes, it seems, more trendy than useful.
<...>
> A potential open-source product will sit somewhere between those two
> extremes. Depending where we each draw the line, you and I may disagree about the
> open-sourcedness of a few products that straddle our personal lines.
A concept that I have had for some time and is really influenced by a
interview I once heard with Carl S:
Open architecture (as distinguished from open source):
A rebol distributed with binaries that could be linked by either a
POSIX- or ANSI-compliant C compiler and has the necessary hooks for
'plug-ins':
Real world example:
A large programming shop that I know of has literally hundreds of small
apps running. These apps are either perl or rebol, depending on whether
REGEX features are needed. Since regex components are available via
open source, if a regex 'engine' could be 'plugged in' to rebol and
a human-readable regex interface written, most of the perl programs
would be replaced by rebol.
Now, think "Rebol for Applications":
CAD-related.
Enabling rebol to run inside of a CAD system, (as AutoLisp does now),
would thrill me to death. Rebol is *so* much nicer to work with than
AutoLisp. (for my taste, that is)
And, as an alternative to VBA:
Enabling rebol to run 'inside of' Open Office as VBA runs 'inside of'
Microsoft Office.
All of these scenario are very attractive to me and to many of those
that I collaborate and communicate with ....
<....>
> This is both a trust model and a meritocracy -- you've got to earn your
> rights to be a committer (able to check code in). Non-committers can only suggest
> changes (or spin off their own code tree and run the risk of it becoming
> unintegratable with the main tree).
>
> I guess that Carl is extremely picky about what code gets checked in to the
> REBOL code base, and even if REBOL were open source, there'd be precious few
> people with the right to do that to the official tree.
Honest to goodness, I mean no irony here, but that might be one method
to overcome the binary skip bug, (as an example) which by Carl's
admission (if I read him correctly) has been so hard to fix.
Bottom line:
!! --- additional resources --- !!
Just my (and I repeat *my*) two cents worth.
--
Tim Johnson <[tim--johnsons-web--com]>
http://www.alaska-internet-solutions.com
