[ale] Object Model on Linux (fwd)

[charlie]

---------- Forwarded message ----------
Date: Sun, 22 Dec 1996 18:33:51 -0500 (EST)
 From: charlie <charlie at felix.cc.gatech.edu>
To: R I Feigenblatt <docdtv at mail.peachlink.com>
Subject: Re: [ale] Object Model on Linux

On Sat, 21 Dec 1996, R I Feigenblatt wrote:

> I would argue that while the stdio pipe mechanism introduced with Unix
> made sense for the teletype-oriented world, it is too simple to be of
> very broad use today. (Consider, for example, the sort of need served
> by Don Libes' Tcl-based "Expect", even in the limited world of text
> streams.)

I'm not asking or saying imagine a world with STDIO streams in a GUI environ.
But the idea that I'm advocating exits in the command line environment and
NOT at the GUI.  Sharing information on a GUI level is far from acceptable.
In fact I agree stdio text streams is NOT the way either, but with the
advent of CORBA, ILU, OPENDOC, and Java we could share this information to
the point where programmers don't need THOUSANDS of DLLs and don't need
all of the extra STUFF we've accumlated.  What I AM asking is for a workspace
that has taken these ideas of good Object Oriented Design and given us users
a mechanism for working with executables and files as if they were OBJECTS
in (insert your favorite OO language here).  Allowing us to share information,
reuse code, and maninpulate things like they were objects.  OS/2 has tried 
this.  This is what the orginal idea of the thread was.  (I believe *right
brad :) *)  Think about if everyone worked in a Smalltalk environment (now
not looking at performance etc etc ) All of our executables equate to objects,
and the main workspace is our enviroment in which we take these objects and
we work with them from there.  Example:  Say I write an object to handle voice
recoginition then you write and object that does spreadsheets.  Then the user
could take each object executables and work with your spreadsheet object using
voice commands!  So the user would say "Open Spreadsheet blah.wps" and bang
spreadsheet blah.wps opens.  This is a completely different paradigm for
desktop enviroments.  It embodies everything that us programmers having been
doing for years to ease the work load of huge programs and complex high level
ideas.  It's time we pass this onto the user, and provide these mechanisms
to him so that he could "build" what he needs from components.

> I would argue that so-called OLE-automation fits the bill for today's
> more demanding needs, and is the KEY contribution OLE makes to computing.
> (I think Microsoft hurt itself by turning OLE into a grab-bag of disjoint
> software technologies for so long.) Visual Basic controls revolutionized
> Windows programming, whereas initiatives like C++ class libraries often
> served to destroy software engineering efforts. Third-parties, not only
> Microsoft, reaped rich rewards by creating them. And ActiveX Controls
> are just the successors to Visual Basic Controls via OLE Controls.

OLE, Visual Basic, and C++ are the worst examples of these ideas!
The only thing they fit is Micro$oft's wallet.
Take this from a C/C++ fan.  C has it's merrits and C++ does, but as examples
of good Object Orient language it is not.  Visual Basic is a 
Bandwagon/Micro$oft let's make our own standard product.  And ActiveX Ole is
a joke!  

'Oh no!' Bill Gates says, 'The internet and Java are cooler than I 
thought. What should I do if C++/VB is not as popular as I imagined in my 
plan.'
'I could wrap up OLE for you,' software engineer.  

As Steven Jobs so eloquently put it: 'Microsoft,' *sigh*, 'there products are 
just so third rate.  They don't have any culture in them.'

Besides the obvious problems with Micro$oft they have NEVER invented anything 
that has had a major impact in the computer industry.  Now I'm not talking
about media hype or profits.  They've hyped and hyped until they can't hype 
anymore and all they ever do is come up with things that other people have 
already done, and those people did it better to begin with.  The ONLY thing
Microsoft ever had was backward compatibility.  Windoze would have never
caught on unless the run MSDOS stuff wasn't there.  They staked out a market
with very little competition, and egomaniacs like Jobs never really tried
to compete with them.  That's why Scully fired him.
Sure they sell billions and billions of stuff, but so does McDonalds.
So where do you want to eat at?  McDonalds?! Or a home cooked backyard burger.
Look at Apple.  They broke the 64K barrier.  They created one of the first
successful GUI enviroments the MAC, and now they're moving into the next
realm of computing PDAs.  Sure Apple has it's problems too, but they've
created, orchestrated, and stole better ideas than Bill Gates ever has.
Micro$oft doesn't get off it's butt until they can either make it a new
standard, or it's popular enough that people want it. 
Micro$oft is still in a browser war with Netscape because they didn't get
on the internet bandwagon soon enough.  While Apple is off doing it's thing
with wearable and portable computing.  While Micro$oft can't stop drooling
over the internet and it's profit margins.

> 'Hiding messy details from the user" (encapsulation) may hold the
> potential for simplying use, but only if the interfaces are well
> documented and very LOCALIZED, not dependent on hidden variables
> which are set by distant, unrelated activities.

							        Charles Hubbard
	   		 	          Internet: charlie at felix.cc.gatech.edu
 ". . .the pope talks a lot about sex, of which he knows nothing. . ."  
	- Robert Anton Wilson