REBOL.org
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

Library member page for grshiplett


Now that we have Rebol Liquid it becomes much more interesting to 
look at implementing a smart Rebol tutorial/training environment 
within Oz if only for the IDE via QTk + AI.


And Rebol has more than just the 'word' idiom in common with Tcl; 
Rebol's use context and bind have their parallel in Tcl upvar. See 
my 'Tcl the Rebol'.


What is also interesting is the high quality of the XOTcl IDE called 
XOTclIDE which retains the same degree of 'prototype-based' as the 
TiScript variant of JavaScript ( of interest because both Rebol and 
TiScript are able to use DyBase )  XOTclIDE offers inspectors and 
workspaces on a par with any Smalltalk programming environment.  


That you should implement a Rebol IDE in Rebol may seem obvious: 
it is not.  Were it in Oz+, the transition from Rebol2 to Rebol3 
or from vanilla Rebol to Liquid Rebol would be trivial.  Ditto for 
other dialects.  An IDE built with Logtalk+XSB+XUL could also be 
as flexible.


The failure of VisualAge for Java was not because the IDE for Java 
was in Smalltalk.  That Eclipse is in java has not been an issue 
for AMZI! Prolog.


It can be reassuring to have the IDE built in the language ( QuickPascal 
comes to mind ) but our confidence in Rebol is not based on the elegance, 
stability and maintainability of the IDE.  Indeed, focus on the Smalltalk 
GUI framework gives the impression that class inheritance is the 
key to OOP and may not have served Smalltalk and Smalltalkers well. 
It also resulted in the impression that a single process was tolerable 
(no native threads.) In terms of widgets, Java AWT to Swing was just 
Smalltalk history repeating itself.  Ruby and performance in a VM 
has its parallel ( Ruby is Smalltalk without the IDE.)


Fudging the line between environment and OS is part of the evolution 
of Smalltalk and can be seen in the current Slate  project. It has 
a parallel in Rebol.  I plan to add a note on Tcl-DP and Hermes in 
this regard.

Framework is not language.  A framework may preclude some languages.

That language choice now largely defaults to a management strategy 
of Java or Microsoft C#.net does not look smart as soon as a language 
such as Scala appears.  If we see managers return to implementation 
language selection as a project milestone, Rebol and other niche 
languages will have some breathing room. AJAX and RIA should not 
dictate a language given what Seaside, Rails, Curl and Rebol can 
offer.  This was always true with Gemstone versus J2EE.  Gemstone 
was just never a buzzword. SOA is a buzzword.  But RIA requires implementation 
choices.  It puts alternative languages back in the game. AJAX as 
Javascript + CSS + HTML + XML+ some_kit + some_other_language + SQL 
does not look or sound risk-aversive.  


Most RIA strategies look to eliminate or constrain JavaScript.  First 
they underestimate JavaScript then try to fall back on Java just 
as Java6 embraces scripting. Go figure. As development platforms hamstring JavaScript, an opportunity opens for flexible dynamic languages.


If flexible, dynamic languages cannot facilitate adaptation so as 
to achieve competitive advantage, then they will not survive in anything 
but their very small chosen niche.  


But the evolutionary metaphor fails because access to resources is 
determined by shibboleth, not by phenotype.  


But social organization counts, which is why a backing of developers 
who produce frameworks which are implemented in reward-winning applications can mean not just surviving in a niche, but flourishing in the open.


It is not being in the open that counts but being seen to be in the 
open.


Most months no jobs post on any major job site in any marginal dynamic 
language even for demonstrations of project feasibility or RAD prototyping. 
 Most companies fail to innovate; they prefer to follow.


What would be refreshing would be to see a Curl front-end and a Rebol 
back-end.  Or Tcl/Tk with Rebol.  Or Rebol with a deductive database. 
Or to see a Rebol variant implemented in Oz.  Or to see Rebol doing 
RDF without XML. 

We will soon see a Rebol3 DLL which can be readily wrapped.


Rebol should be marketed with national language variants and interop. 
 Rebol should be the first Web2.0 language which is not a set of 
English keywords.  This option was always available to Tcl.


Rebol developer active with Io, Javascript/Self, Curl, MetaCard/ 
Transcript/ Revolution and Mozart-Oz with many years of Smalltalk 
and Prolog.
Active user of Jython, Logtalk and XUL
Learning LIFT for Scala on Jetty to compare to Squeak Seaside

Contributes to Ruby ( Mondrian IDE updates) and wikipedia computing 
stubs
Contributing 3.9+ Prolog/V for Squeak Smalltalk

Author of java classpath tool and various flight sim configuration 
tools

Author of Dynamo for FS4 ( dynamic aerobatic squadron generator - 
in PDC Prolog, compiled using Borland C++ for OS/2)

Accepts regular e-mail in French, German, English or shorter msg 
in any language.
Interest in console REPL interpreters and shells for the same.
home is www.logiquewerks.com
blog is eclectic-pencil.com and eclectic-pencil.blogspot.com