Squeak and ST in general (3)

Torsten.Bergmann at phaidros.com Torsten.Bergmann at phaidros.com
Tue Nov 16 22:22:16 UTC 1999


OK - last try to finish this mail with the [*bad word removed*] MS Outlook.

Let me add my 2 cents to squeak and Smalltalk in general: 

While being a student I've tried to get familiar with all the various
Smalltalks out there, but there were only a few systems I like most. 

1. Smalltalk/X
==============
   ST/X is one of the most powerfull ST's around. It's cross plattform,
   has a lot of tools, goodies, and it also integrates Java. With the 
   ST2C Compiler you can compile to native format, you can write inline
C,...
   What's the problem with ST/X: only a few people know about his power.
   Personally I don't like the dark gray User Interface, but everybody can
   change that  ;)

2. SmalltalkMT 
==============
   A lot of Smalltalkers think it's not a true Smalltalk, because it has
   no virtual machine and you have to know a lot about Windows and COM.
   But it's very fast and the only ST around that can create small
   runtime executables (no big image). With COM you have the ability to
   access components built in Java, VisualBasic, C++, Smalltalk, ...
   What is the problem of COM: it's Microsoft, it's complicated and not
   so well designed. But COM is a good way to communicate with other 
   applications, integrate your own applications into others, divide tasks
   into components, ...
  
2. Squeak (of course)
=====================
   A lot of people in the Smalltalk community think that squeak is like
having
   a nice game. But a lot of squeakers think about creating a system that
can
   beat Media directors, Java, Windows or whatever ...
   A system that can boot, run on a host OS, on a PDA, that can behave like
   VMWare or Wine, ... 
   After I've seen squeak booting with User Interface I would'nt laugh
anymore 
   about the idea of a squeak operating system. 

I don't like the following systems:
==================================
VisualAge:   Too big, to slow 
VisualWorks: It's a good and stable IDE but nothing really happend in the 
             newer versions, also a big image

The power and also the problem in the ST-World is diversity. You can use
Pocket 
Smalltalk for creating Palm applications, ST/MT or Dolphin for Win
applications,
ST/X for Unix, STAgents for Mac/Win, ObjectStudio, VisualWorks, VisualAge,
STExpress,
....

But Squeak is a little bit different: it's not only a Smalltalk-IDE it is 
a SYSTEM. If I see Squeak growing I see a lot of people with different
interests 
working on this system. Some would like to use Squeak as multimedia
environment,
others would like to use it for web applications, games, programming, ...
Comparing to Java I would say Squeak is the platform and Java the language.

What should we do to keep the squeak platform/system running (some ideas):

 - get squeak closer to commercial Smalltalks (speed and tools)
   Fast bytecode with Jitter is good - but what about a way to compile
native,
   building native libraries from within squeak ?   
   Get squeak accepted in the Smalltalk community!

 - get squeak closer to underlying systems (hardware and software)
   Changing UI like VisualWorks is a nice thing, but squeak should have it's
own.
   Getting Squeak closer to Linux, drivers, ... ?

 - break the Smalltalk limits (big image, big object system with
dependencies)
   create interchangeable squeak components with stable interfaces, uniquely
identified
   code, Sandbox for downloaded components, code signing, class caches, ... 
   We can learn a lot from COM. In contrast to Java components it is not
limited to one 
   language. If you create interfaces like COM or CORBA you can change
implementation
   and interchange components without effects to clients. 

 - think about a change from one-user-per-image to multiuser-per-image

 - integrate other languages into squeak 
   Prolog is done, what about C/C++, ...
   Get squeak accepted in the developer community!
   Improve the multilanguage support (menue strings as interchangeable
resources, UNICODE, ...)

 - show people that we can also create serious applications    

 - find a way to install/deinstall squeak programs with different versions 

 - experiment with new ideas, get squeak accepted by teachers, universities,
industry, ...

 - reuse things that already been done
   UIUC Archive contains a lot of code! 

 - interface with application data formats (for instance import all the
different image formats
   used on different platforms)
 
 - give people what they wanna have on their computer:
   text processing, a good webbrowser, paint programs, media, games,
communication, ...
   Get squeak accepted in the user community

 - motivate others to take part on the squeak development

 - speak about the things already done, write articles, books, websites, ...


Like Dan said "Let's Get Bizzy" 

-Torsten


BTW: While looking today at www.javasoft.com I was a little bit surprised:
     Java is now "Turbo-Charging client-side performance" and is moving
     "from basic functionality to extreme performance". It is now 
     "super-high-performance" and "Swing (JFC/Swing) libraries have been
     super-tuned". The "performance has been boosted". - Wow ! 

     What about implementing a virtual machine for Assembler or C in the 
     Java language to get all these slow native CPU program's faster. ;)





More information about the Squeak-dev mailing list