what's wrong with REBOL?
Cameron Laird
claird at lairds.com
Thu Sep 4 08:25:59 EDT 2003
In article <mailman.1062519306.15064.python-list at python.org>,
Jeff Epler <jepler at unpythonic.net> wrote:
>For a language to be "right" for me, it must satisfy a lot of qualities.
>
>One of them is some notion of freedom. For instance, will I be left
>high and dry if I want to use this language on some new device or
>system architecture? (For instance, when I start buying Itanium or
>Opteron servers to replace my racks of decrepit Xeons..)
>
>Here are two ways I can be pretty sure this won't happen:
>* Can I have the source code? (C, C++, Python, Perl) I'll port it to
> my new platform.
>* Is there already a standard with multiple working implementations?
> (C, C++, C#, Python) This multiplies the chances that some
> implementation will be ported to my new platform.
>
>As far as I can tell, the Rebol folks don't include source code with any
>of their licensing options (SDK, etc), and it doesn't look like their
>documentation gives enough information for a third-party to write their
>own implementation either.
>
>Of course, for those who already rely on single-source software with no
>source-code availability (*cough*microsoft*cough*), maybe this isn't
>such a big deal. But having been prevented from upgrading to a modern
>version of a compiler by a no-source third-party C++ library recently,
>I don't relish the idea of being reduced to begging for a recompiled
>version of commercial software again.
>
>Jeff
>
Source code is available for some parts of REBOL, under
some circumstances. Discussion of REBOL's proprietary,
closed-source nature has dominated this thread, though,
and that's appropriate.
REBOL has a few technical failings, as well. While I think
it's quite rare for anyone to decide against REBOL because
of these, there's real value in being specific, rather than
treating the language as a rather cloudy abstraction.
Unfortunately, I'm not a good candidate for the task. It's
been a couple years since I was current with REBOL. Among
the blemishes I believe the language (including its
libraries) still exhibits are:
* an inflexible concurrency model
* cumbersome interfaces to modules
coded in other languages (conse-
quence: embedding and extending
are impractical)
* arguable scalability, in the sense
of what appears to be skimpy sup-
port for team-written projects
* no /APACHE for convenient scripting
with the market-leading Web server
* near GUIlessness
* inconsistent functional-vs-object-
oriented semantics in aspects of
program evaluation
Conclusion: there's a lot to like about REBOL; it's not
perfect; its licensing, rather than technical factors,
dominates most use decisions.
--
Cameron Laird <Cameron at Lairds.com>
Business: http://www.Phaseit.net
Personal: http://phaseit.net/claird/home.html
More information about the Python-list
mailing list