[REBOL] The future of Rebol - achieving critical mass
From: geoff:productivity at: 21-May-2001 12:15
Hi folks
I have just been evaluating Rebol, (with the kind help of Carl and this
list) and it seems to have a lot going for it. But I am wary of making a
major commitment, as I have reservations about its future, given the RT
business model. At the risk of being flamed, I thought I would share my
concerns and see what the wider community feels about this issue.
I have been bitten before by technically superior products with the wrong
marketing model - I invested in the Amiga and NextStep to name just two. A
lot of developers lost their shirts on NextStep. I fear that Rebol may go
the same way. As Windoze, VB, ASP etc demonstrate so eloquently, technical
excellence and commercial success are not at all the same thing in the
software industry.
I have two concerns about the current approach:
- First, I fear that RT will fail to achieve a critical mass of users
- Second, that they will fail to achieve a critical mass of libraries and
APIs
It is hard to get a new language established, especially if it is based on a
paradigm that will be unfamiliar to most programmers (ie functional
programming). I suspect that it may be impossible if you charge hundreds of
dollars for the basic functionality in /Command.
Rebol is up against powerful, free environments like Perl / Python / PHP /
Ruby / Dylan / Lisp / Scheme / Guile etc etc. Most are Open Source. Rebol is
nice, but the source is closed and by charging for the basics it is entering
the market with a huge handicap, right from the start.
So far as I can see, commercial success in the software industry is all
about achieving a critical mass of users and generating Java style buzz. If
I were Carl and his backers, I would be giving away the full versions of
/Command, /View and the application wrapper, to get momentum going.
I appreciate that this would take commercial courage, but in the medium term
I think it would be much the safest strategy. With a large and vibrant
community the future of Rebol would be assured, and they could be generating
much greater revenues from selling advanced servers and applications,
specialised libraries, an integrated IDE, and consultancy than they will
ever generate selling /Core to a tiny community. With their current
approach, they may well be going up the same dead end as all the closed 4GL
languages that were supposed to revolutionise the industry (remember them,
anyone?).
And there would be a second, very major advantage. By charging for the C/C++
API they are creating a major disincentive for the community to integrate
Rebol with external libraries. Much of the success of PHP, for example, can
be attributed to its APIs to external libraries - almost 2000 internet
related functions available at the last count. I have an idea for an Open
Source project, but it would require /Core to run and this just does not sit
well with the Free Software ethos. If Rebol is to achieve a critical mass of
APIs and library code, Carl must somehow harness the energies of the Rebol
community. Compare what is available in Perl or PHP with the Rebol code
repositories and you will see the nature of the problem...
What do people think?
Geoff Caplan
