Red Herring (was: Bitten by dynamic typing...) 
Author Message
 Red Herring (was: Bitten by dynamic typing...)

Quote:

> Would you know offhand where VisualWave fits in with other web technologies like CGI,
> or Java servlets?  (I assume that it's a server-side system of some sort.)

> I actually just kicked some ideas around with someone about making a "Smalltalk
> Servlet" system.  If something like this already exists, that would be very nice.

What VisualWave does is similar to what servlets do: It provides dynamic
creation of html pages. What makes a big different - at least I
experienced it like that - is that you don't have to produce html
directly in VisualWave (although you can if you like). VisualWave
"translates" your normal GUI (as far as possible) into generated html
pages, i.e. it handles (client side) application flow tracking
automatically (either with cookies or hidden attributes). Thus you often
don't have to create a flow of pages which you manually string together
by handing attributes (or objects) between them. Instead you write a web
application like every other application...
I certainly was more productive with VisualWave then I'm now programming
servlets.

Now consider to additionaly use ClassicBlend for UI parts which cannot
be covered by html and you're having applets and dynamic html pages
directly accessible as ST objects.

Johannes



Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:


> > > [...]
> > > But IBM builds the VisualAge/C++ stuff in smalltalk. It makes me
> > > wonder if maybe it's better to use their smalltalk for GUI/integration stuff
> > > than C++.

> > IBM has an existing codebase written in Smalltalk.  I would wager that if they
> > needed to build an IDE from the ground up today, it would be in Java.

> I think we've already established that building a complete IDE in Java would be very
> difficult, if not impossible.  The reflective, dynamic, everything-is-an-object
> capabilities would lead IBM to choose Smalltalk again, IMO (unless the decision was
> made by a non-technical manager).

Where was that point established?

Quote:
> David
> --
> ***************************************
> David Whiteman
> Applied Reasoning
> whiteman AT AppliedReasoning DOT com
> ICQ# 12178974
> http://www.geocities.com/Heartland/3261
> ***************************************

        -Mike


Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:

> What VisualWave does is similar to what servlets do: It provides dynamic
> creation of html pages. What makes a big different - at least I
> experienced it like that - is that you don't have to produce html
> directly in VisualWave (although you can if you like). VisualWave
> "translates" your normal GUI (as far as possible) into generated html...
> I certainly was more productive with VisualWave then I'm now programming
> servlets.

Interesting.  It'd be interesting to see how an event model is mapped to http.  Like you
say, the free, no-frills servlet development kit doesn't come with classes that provide
this layer of abstraction.  If you're willing to pay for it, though, I think that there's
servlet libraries that do what you're describing.  Have you checked any out for your
project?  For example, VA for Java (v 2.0) lets you develop html pages in its GUI designer,
and you handle events like form submissions by wiring java beans together.  I developed a
couple of prototypes, which went very quickly, and I could deploy them to an Apache
server.  I need to check it out further, though, to see if I can everything I'd need to
through this library.

Quote:

> Now consider to additionaly use ClassicBlend for UI parts which cannot
> be covered by html and you're having applets and dynamic html pages
> directly accessible as ST objects.

Hmm - also very interesting.

Thanks,
Robb



Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:


> > An IDE is a pretty good test because it has tool bars, buttons, a
> > strong edit control, file reading and writing, and integration with other
> > programs like compilers and de{*filter*}s. If you can't do these things well with a
> > language, what is it going to be like for industrial and commercial
> > applications?

> IDE's have little or no concept of distributed programming.

Check out VisualAge Smalltalk with Distributed Smalltalk.  It includes
Distributed Browsers and Distributed De{*filter*}s/Inspectors for Browsing,
Writing, and Debugging code (and inspecting objects) in other images on
other machines.  An IDE with a *very* strong concept of Distributed
Programming.

(Using a Distributed De{*filter*} to use a client image on NT to patch a bug
in the server image on AIX while it was running was a nearly religious
experience for me)

Also, I'm not sure if this is the direction they are intending or not,
but Squeak's use of the Pluggable Web Server and Scamper come close to
making evey local instance of the development environment just a local
node in a distributed development environment.

Quote:

>   Senior Consultant
>   STR

Take care,

Jay O'Connor



Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:

> This is not a language issue, but rather an IDE issue, and is independent of the
> language that the IDE was written in.  For example, I use emacs for a lot of my
> development for multiple languages.  I can extend it to integrate and glue code as
> well.

One of the better general-purpose IDEs to ever have been written.
Wasn't it written in C for portability reasons? Notice that most
of the cool stuff was NOT done in C, but an embedded scripting
language (elisp) because it's not good to extend the system in C
and it's easier to patch and extend lisp. Why? Dynamic typing,
and interpreters are better for that stuff.

I earned my pay in C for a very long time, and earn most of my
pay these days in C++, but I'd welcome a change to smalltalk or
python (or maybe dylan) faster than a change to Java. That's just
because of the clean nature of the language and the bureaucratic
nature of programming in some of the others.

I can use and admire one language, and still see why others would
be more appropriate sometimes, or even more fun. :-)

Quote:
> IDE's have little or no concept of distributed programming.

We'll find that this isn't entirely true. Better IDEs are extendable
enough that there are ways to make it distributed. The advantage of
an extremely dynamic environment is that it's easier to stretch it
in ways originally not conceived. In statically-typed langauges, it
just takes a little more work.

Quote:
> > But IBM builds the VisualAge/C++ stuff in smalltalk. It makes me
> > wonder if maybe it's better to use their smalltalk for GUI/integration stuff
> > than C++.

> IBM has an existing codebase written in Smalltalk.  I would wager that if they
> needed to build an IDE from the ground up today, it would be in Java.

The original one didn't appear out of the ether. Someone
built it, and they built  it in smalltalk for some reason.

--
--------- There is no craftite {*filter*} -------------
Tim Ottinger         Object Mentor      OO Training and

-------------------------------------------------------
We can interpret a bad temper as a sign of inferiority.
                                        -- Alfred Adler



Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:


> > This is not a language issue, but rather an IDE issue, and is
independent of the
> > language that the IDE was written in.  For example, I use emacs for
a lot of my
> > development for multiple languages.  I can extend it to integrate
and glue code as
> > well.

> One of the better general-purpose IDEs to ever have been written.
> Wasn't it written in C for portability reasons? Notice that most
> of the cool stuff was NOT done in C, but an embedded scripting
> language (elisp) because it's not good to extend the system in C
> and it's easier to patch and extend lisp. Why? Dynamic typing,
> and interpreters are better for that stuff.

But my point is that the IDE doesn't have to be built in the language
you are developing
in.  Nor should it matter.

Quote:
> > IDE's have little or no concept of distributed programming.

> We'll find that this isn't entirely true. Better IDEs are extendable
> enough that there are ways to make it distributed. The advantage of
> an extremely dynamic environment is that it's easier to stretch it
> in ways originally not conceived. In statically-typed langauges, it
> just takes a little more work.

It is also easier to break in ways not originally conceived.  In theory,
since the
Smalltalk IDE is entwined with the language, you could tweak the IDE but
end up breaking
your production code.  Or you could introduce a subtle bug in some of
the more advanced
IDE features.  If you follow a rigorous coding practice, any change to
the IDE should
require a reapplication of the testing originally applied to the IDE.

Quote:
> > > But IBM builds the VisualAge/C++ stuff in smalltalk. It makes me
> > > wonder if maybe it's better to use their smalltalk for

GUI/integration stuff

Quote:
> > > than C++.

> > IBM has an existing codebase written in Smalltalk.  I would wager
that if they
> > needed to build an IDE from the ground up today, it would be in
Java.

> The original one didn't appear out of the ether. Someone
> built it, and they built  it in smalltalk for some reason.

Here's the unofficial scoop on VisualAge for Smalltalk heard through the
grapevine.
Originally, IBM developed the GUI building stuff using Digitalk
Smalltalk.  However,
Digitalk was busy working on Parts.  When IBM wanted to ship their
product bundled with
Digitalk Smalltalk, Digitalk pulled a non-compete clause.  So IBM bought
another
Smalltalk engine (Envy), and ported the GUI to that engine.  This effort
took over a
year (cheers to all the Smalltalk portability folks), and the initial
releases (1.0 and
2.0) were appallingly slow even by Smalltalk standards at the time.

So to answer your assertion, IBM build VisualAge in Smalltalk because it
was originally
intended to be a Smalltalk development environment.  It just so happens
they did it
before Java even existed (it was still Oak at the time).

Similarly, Visual Studio is most likely written in C++ because that was
the original
target for the environment.  Then Microsoft reused the code to develop
VisualJ++.

JBuilder was written with Delphi (Pascal) because Inprise had already
developed a
substantial IDE in Delphi.  FYI, has been substantial talk of Inprise
porting JBuilder
to eventually be 100% pure Java.

  loverde.vcf
< 1K Download


Wed, 18 Jun 1902 08:00:00 GMT  
 Red Herring (was: Bitten by dynamic typing...)

Quote:

> Microsoft, at least in 1995, did not use MFC internally because it was too
> buggy and had too many memory leaks.  This is according to MS project
> leaders.

I'd heard more recently that they were, but I know that
it has had its share of problems. We ran a memory utility
on our software in 1996 and all of the leaks and memory
corruption were in the MFC and ODBC drivers, not one of
them in our code. At that time, the ODBC drivers were
particularly bad. So not only was our code better
designed, it was written better, too. :-)

Quote:
> Dylan compilers are written in Dylan.  In Dylan, the development
> environment is separate from the program being developed, but the two can
> 'talk' to eachother. Harlequin Dylan has a 'listener' where the programmer
> can iteractively write and execute code.  The now-defunct Apple Dylan
> technology release did too.  Gwydion Dylan will hopefully someday have a
> listener.

See, you're tempting me towards Dylan increasingly. I
read the COM article too. We're in the middle of a
"great language boom" of some sort. Dylan, Smalltalk,
Python, Eiffel, etc, are all too much fun it seems.

Someone make me a 36-hour day, please.

--
--------- There is no craftite {*filter*} -------------
Tim Ottinger         Object Mentor      OO Training and

-------------------------------------------------------
We can interpret a bad temper as a sign of inferiority.
                                        -- Alfred Adler



Wed, 18 Jun 1902 08:00:00 GMT  
 
 [ 7 post ] 

 Relevant Pages 

1. no noalias not negligible: more red herrings, conclusion unchanged

2. Bitten by dynamic typing (was Re: Smalltalk vs Java)

3. Bitten by dynamic typing (was Re: Smalltalk vs Java)

4. IDEs written in target language (was Re: Bitten by dynamic typing )

5. Design by Interface in Squeak [LONG] (was Re: Bitten By Dynamic Typing)

6. Types, types: static vs. dynamic -- classification

7. Types, types: static vs. dynamic --

8. Types, types: static vs. dynamic -- classification

9. ML as Scheme: Dynamic typing in a strongly-typed language

10. Types, types: static vs. dynamic -- classification

11. Question about dynamic types of record types

12. Red-faced professor gets bitten in search for portability

 

 
Powered by phpBB® Forum Software