Everybody:
I think if Mozart is to "take off" the powerful multi-paradigm concepts on
which the kernel language approach is built must be translated into a value
proposition IT organizations can take to the bank.
I've given this a lot of thought and it seems to me 3 things are needful
from a technology standpoint:
1. A better, more comprehensive, more systematically organized Standard
Library. I like the "YBL" - "Yellow Brick Library" for reasons that are
embarrassingly obvious. In order to compete with the "big boys" Mozart-Oz
needs to have a general approach to common problems like universal data
access, and a development framework for something at which it is
particularly well suited. I nominate "MIDDLE TIER BUSINESS LOGIC" as the
likely candidate.
2. Following the train of thought form pt. 1, it seems reasonable that the
"killer app" for Oz that would earn the language a following in the business
world is twofold:
a. An application server (I like the name "Emerald City") built for
interfacing to Java or .NET
clients and any database back end - the ultimate middle-tier tool
for building business
objects. The value of the business objects NOT being hog-tied to
a particular computational
model should then be obvious once a general approach to solving
specific kinds of business
problems utlitizing the constraints programming and concurrent
declarative programming
capabilitites of Mozart is facilitated by the Standard Library and
exposed in this environment.
b. A development tool for composing these business objects and
interfacing them to the target
front and back end technologies. My once-and-future "Amadeus" IDE
comes to mind here...
c There should also be a GUI application which allows the user to
configure app server properties,
add and remove Business Objects (i.e., Functors), and everything
else app servers do in the
real world, including load balancing and the whole nine yards.
"Mozart Maestro".
3. In order to facilitate 2, Mozart needs a top-quality GUI toolkit and set
of base components for building applications. I know Tk is popular on Unix
but it doesn't quite pass the muster for Windows users. I nominate wxWindows
as the cross-platform library of choice. Like Oz itself, it's build in C++.
It is feature complete and uses native widgets where appropriate on each
platform.
I'd like to put together a development team to start prototyping these ideas
and, once 1.3.0 is out (hopefully with the new ByNeed implementation and
other optimizations), start serious development of these add-on packages.
The POC I personally will be working on will include a .NET client (ASP.NET,
WinForms, etc.) on the front end and a PostgreSQL database on the back end.
(I have my reasons here at WebMD...)
Anybody interested in this besides me ... ? I am going to start taking
requirements and I'll probably create a new project on SourceForge (unless I
can just add it to the current CVS as a separate development branch).
There are huge implications in those tidy 3 points (for instance, a version
of the emulator that can be made emeddable may be needed - all kinds of
things may be needed - to support the effort). But I think the vision is a
good one and can take Mozart-Oz to the next level in terms of a.) industry
applicability and b.) real-world visibility.
Being able to talk to both .NET and Java, and interface to popular web
servers, would be a major strategic design objective, needless to say.
/**
* Bob Calco
* Senior Software Architect
* Medical Manager R&D/WebMD
* 15151 NW 99th Street
* Alachua, FL 32615
* 386.462.2148 x 27111
* 352.278.1140 (cell)
* bcalco@webmd.net <mailto:bcalco@webmd.net>
* "rcalco" handle (both AIM and Yahoo)
*/
-
Please send submissions to hackers@mozart-oz.org
and administriva mail to hackers-request@mozart-oz.org.
The Mozart Oz web site is at http://www.mozart-oz.org/.