Borland and me: We both want less code and more pictures!

[Mike Mitchell]

I just read the story on Borland's Galileo project
(http://zdnet.com.com/2100-1104-954958.html), and one sentence shines
out:

"Shelton said Borland plans to offer features not currently available
in Visual Studio.Net, such as the ability to "model," or build
graphical representations of software, so programmers don't have to
write all the code by hand."

Well, amen to that, I say! The way we develop applications is still in
the dark ages, with folks (sometimes clever folks) writing thousands
upon thousands of lines of code - total drudge work. Make application
delevelopment rapid, make it pictorial, reduce code to a minimum, and
open up the entire process to end users. That's the way to Software
Shangri-la.

MM

[Patrick Troughton]

Have you seen SoftWIRE? It lets you create apps....without writing any code.
You drag and drop icons around like it was a flow chart.....

http://www.fawcette.com/products/vsn...tems/softwire/

http://www.softwiretechnology.com/dot-net/

/Pat

kylix_is@yahoo.co.uk (Mike Mitchell) wrote:
>I just read the story on Borland's Galileo project
>(http://zdnet.com.com/2100-1104-954958.html), and one sentence shines
>out:
>
>"Shelton said Borland plans to offer features not currently available
>in Visual Studio.Net, such as the ability to "model," or build
>graphical representations of software, so programmers don't have to
>write all the code by hand."
>
>Well, amen to that, I say! The way we develop applications is still in
>the dark ages, with folks (sometimes clever folks) writing thousands
>upon thousands of lines of code - total drudge work. Make application
>delevelopment rapid, make it pictorial, reduce code to a minimum, and
>open up the entire process to end users. That's the way to Software
>Shangri-la.
>
>MM

[Robert Lantry]

It's called programming. Look into it sometime.



"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
news:3d68ac6c.3551620@news.devx.com...
> I just read the story on Borland's Galileo project
> (http://zdnet.com.com/2100-1104-954958.html), and one sentence shines
> out:
>
> "Shelton said Borland plans to offer features not currently available
> in Visual Studio.Net, such as the ability to "model," or build
> graphical representations of software, so programmers don't have to
> write all the code by hand."
>
> Well, amen to that, I say! The way we develop applications is still in
> the dark ages, with folks (sometimes clever folks) writing thousands
> upon thousands of lines of code - total drudge work. Make application
> delevelopment rapid, make it pictorial, reduce code to a minimum, and
> open up the entire process to end users. That's the way to Software
> Shangri-la.
>
> MM

[Miha Markic]

Yeah, you might try with Photoshop or PhotoPaint.

Miha


"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
news:3d68ac6c.3551620@news.devx.com...
> Well, amen to that, I say! The way we develop applications is still in
> the dark ages, with folks (sometimes clever folks) writing thousands
> upon thousands of lines of code - total drudge work. Make application
> delevelopment rapid, make it pictorial, reduce code to a minimum, and
> open up the entire process to end users. That's the way to Software
> Shangri-la.
>
> MM

[Mike Mitchell]

On Mon, 26 Aug 2002 09:57:08 +0200, "Miha Markic"
<miham@pleasenospam.spin.si> wrote:

>Yeah, you might try with Photoshop or PhotoPaint.
>
>Miha
>
>
>"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
>news:3d68ac6c.3551620@news.devx.com...
>> Well, amen to that, I say! The way we develop applications is still in
>> the dark ages, with folks (sometimes clever folks) writing thousands
>> upon thousands of lines of code - total drudge work. Make application
>> delevelopment rapid, make it pictorial, reduce code to a minimum, and
>> open up the entire process to end users. That's the way to Software
>> Shangri-la.

Oh, ye of little faith! Here's an interesting quote from The Decline
and Fall of the American Programmer:

"...here is a good test to determine whether this [see below - MM] is
just philosphical bullshit or a /real/ change: tell your programmers
that they can no longer make any changes to source code (just like you
told them 10 to 20 years ago that they could no longer patch binary
object code); /any/ changes must be accomplished by changing the
specification or the design model and regenerating the code."

The "this" referred to in the above quote was the "post-CASE
environment" as outlined by Carma McClure. The P-CE, as opposed to the
"Pre-CASE" environment, was one of "Emphasis on analysis, design"
(P-C: "Emphasis on coding, test"); "Automated code generation" (P-C:
"Manual coding"); "Maintaining specs and design" (P-C: "Maintaining
code"). Other elements of a post-CASE environment were also listed.

Now all of this was under consideration well over ten YEARS ago! No
wonder that software production is in such a state now if everyone is
STILL thinking in terms of writing reams and reams of code as the
be-all-and-end-all aproach to application development! You obviously
*like* writing code! Your aim *is* to write code, not develop
applications! The application just happens to be a fortunate
by-product of code-writing. Just suppose for a moment that no code
changes could be made, what loud shrieks of protest we would hear
then!

MM

[Michael D. Kersey]

Mike Mitchell wrote:
> >"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
<snipped>
> You obviously
> *like* writing code! Your aim *is* to write code, not develop
> applications! The application just happens to be a fortunate
> by-product of code-writing. Just suppose for a moment that no code
> changes could be made, what loud shrieks of protest we would hear
> then!
> MM

"fortunate"? Maybe for the developer, probably not for the user!-))
BTW Paul Graham author of "On Lisp and "ANSI Common Lisp" has some
interesting comments on OOP and large IT shops at
http://www.paulgraham.com/noop.html :
pg> 2.Object-oriented programming is popular in big
pg> companies, because it suits the way they write
pg> software. At big companies, software tends to be
pg> written by large (and frequently changing) teams of
pg> mediocre programmers. Object-oriented programming
pg> imposes a discipline on these programmers that
pg> prevents any one of them from doing too much
pg> damage. The price is that the resulting code is bloated
pg> with protocols and full of duplication. This is not too
pg> high a price for big companies, because their software
pg> is probably going to be bloated and full of duplication
pg> anyway.
pg>
pg> 3.Object-oriented programming generates a lot of what
pg> looks like work. Back in the days of fanfold, there
pg> was a type of programmer who would only put five
pg> or ten lines of code on a page, preceded by twenty
pg> lines of elaborately formatted comments.
pg> Object-oriented programming is like crack for these
pg> people: it lets you incorporate all this scaffolding right
pg> into your source code. Something that a Lisp hacker
pg> might handle by pushing a symbol onto a list becomes
pg> a whole file of classes and methods. So it is a good
pg> tool if you want to convince yourself, or someone
pg> else, that you are doing a lot of work.

and on another vein (one that you (MM) have tapped into on these
newsgroups) Graham states:
pg> There is a kind of mania for object-oriented programming at
pg> the moment, but some of the smartest programmers I know
pg> are some of the least excited about it...

[Daniel Pratt]

"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
news:3d68ac6c.3551620@news.devx.com...
> I just read the story on Borland's Galileo project
> (http://zdnet.com.com/2100-1104-954958.html), and one sentence shines
> out:
>
> "Shelton said Borland plans to offer features not currently available
> in Visual Studio.Net, such as the ability to "model," or build
> graphical representations of software, so programmers don't have to
> write all the code by hand."
>
> Well, amen to that, I say! The way we develop applications is still in
> the dark ages, with folks (sometimes clever folks) writing thousands
> upon thousands of lines of code - total drudge work. Make application
> delevelopment rapid, make it pictorial, reduce code to a minimum, and
> open up the entire process to end users. That's the way to Software
> Shangri-la.

First, I'm perplexed by Shelton's comments because VS .NET Enterprise
Architect includes Visio for Enterprise Architects, which does exactly as
Shelton describes.

Second, graphical development tools have been available for some
time now, and though they have not achieved wide acceptance yet, it seems
likely that they have a good future once they get beyond their current
limitations. However, (picture me shouting this) they have *no* bearing on
the merits of .NET vs. any other development platform currently in
existence, which, I'm pretty sure, is the general topic of discussion in
this newsgroup.

You know I'd love to hear from some folks at SoftWire because I suspect
that, due to the openness and flexibilty of the VS .NET environment and, in
particular, of the .NET platform itself, the implementation of SoftWire for
..NET was much easier than for prior platforms.

Regards,
Dan

[Patrick Meader]

Hi Daniel:

>> You know I'd love to hear from some folks at SoftWire because I
suspect that, due to the openness and flexibilty of the VS .NET environment
and, in particular, of the .NET platform itself, the implementation of
SoftWire for .NET was much easier than for prior platforms.<<

I'll drop a gent I know there a line, see if he'll address that point.

BTW, Mike: As others have stated, SoftWire might be your cup of tea. Unlike
Delphi or Java, it takes the notion of graphical programming to heights only
hinted at in VB proper. The tool functions as an add-in to VB, and there are
versions for both VB6 and VS.NET.

Patrick

[Keith]

Well, I'm probably not the one you were looking for, but I guess I'll
de-lurk for a moment. VS.Net was easier for us to integrate into for a
couple of reasons. 1) MS provided support for integration as a product
feature, rather than as and add-on to an existing product. This allowed the
integration to be deeper, so we can become more a part of VS than just an
addition. (2) When you went under the hood, you can see a bit of the history
of VB6. It did not present a consistent interface, and things were not
always as they appeared. For example, the controls in VB6 appeared to be COM
objects, but they actually have a wrapper for extended properties. These
objects don't really act quite like they should, and don't expose interfaces
that are supposed to be there. I believe this just reflects that COM was
added into VB afterwards, and they kept their control code rather than
totally rewriting it. There were a number of areas where VB6 caused us to
scratch our heads, and some, like control arrays, that we just couldn't
support.

In general, .Net has been a much more stable, efficient environment for us,
and I would suspect that MSs experience with VB addins and designers helped
in their making it so. BTW, the subject of this thread was not lost on me.
Any guesses as to how similar Delphi and VS are in terms of integration,
given some of the common personnel? Just wondering Back to lurking.
Keith

"Patrick Meader" <pmeader@fawcette.com> wrote in message
news:3d6a74f0$1@10.1.10.29...
> Hi Daniel:
>
> >> You know I'd love to hear from some folks at SoftWire because I
> suspect that, due to the openness and flexibilty of the VS .NET
environment
> and, in particular, of the .NET platform itself, the implementation of
> SoftWire for .NET was much easier than for prior platforms.<<
>
> I'll drop a gent I know there a line, see if he'll address that point.
>
> BTW, Mike: As others have stated, SoftWire might be your cup of tea.
Unlike
> Delphi or Java, it takes the notion of graphical programming to heights
only
> hinted at in VB proper. The tool functions as an add-in to VB, and there
are
> versions for both VB6 and VS.NET.
>
> Patrick
>
>
>
>

[Mike Mitchell]

On Mon, 26 Aug 2002 11:59:33 -0700, "Patrick Meader"
<pmeader@fawcette.com> wrote:

>BTW, Mike: As others have stated, SoftWire might be your cup of tea. Unlike
>Delphi or Java, it takes the notion of graphical programming to heights only
>hinted at in VB proper. The tool functions as an add-in to VB, and there are
>versions for both VB6 and VS.NET.

Yep, I'm aware of SoftWire and I think it's a pretty innovative tool.

MM

[David Bayley]

Michael D. Kersey wrote:

> BTW Paul Graham author of "On Lisp and "ANSI Common Lisp" has some
> interesting comments on OOP and large IT shops at
> http://www.paulgraham.com/noop.html :

Sheesh, you do like the lunatic fringe don't you. ;-) Functional
languages like Lisp have pretty much failed outside of academia, and
Paul Graham's opinions as a Lisp advocate should be read in that light.

> pg> 2.Object-oriented programming is popular in big
> pg> companies, because it suits the way they write
> pg> software. At big companies, software tends to be
> pg> written by large (and frequently changing) teams of
> pg> mediocre programmers. [...]

Many big companies are weighed down by legacy code, and don't have the
luxury of re-architecturing them in OO or anything else. Much of the
work is integration and maintenance. Smaller companies were just as
much a part of the mass adoption of OO, if not more so, as big companies
IME.

Graham's assertion otherwise is just a strawman, that enables him to
argue against "large (and frequently changing) teams of mediocre
programmers". Such gross generalizations don't bode well for any
rational arguments he might have.

> pg> [...] Object-oriented programming
> pg> imposes a discipline on these programmers that
> pg> prevents any one of them from doing too much
> pg> damage. The price is that the resulting code is bloated
> pg> with protocols and full of duplication. This is not too
> pg> high a price for big companies, because their software
> pg> is probably going to be bloated and full of duplication
> pg> anyway.

I can sort of agree with this. The discipline of protocols is an
advantage that OOP provides. This works well with big companies, but
Graham misses the even bigger benefits that it provides to even bigger
_communities_ of programmers.

The availability of re-usable, protocol rich, components across company
boundaries, is something that has driven the popularity of OOP
languages. Even the anti-OO Mike Mitchell agrees with this point, in
that he advocates the easy re-use of other people's OO components (so
long as he doesn't have to write them himself!).

> pg> 3.Object-oriented programming generates a lot of what
> pg> looks like work. Back in the days of fanfold, there
> pg> was a type of programmer who would only put five
> pg> or ten lines of code on a page, preceded by twenty
> pg> lines of elaborately formatted comments.
> pg> Object-oriented programming is like crack for these
> pg> people: it lets you incorporate all this scaffolding right
> pg> into your source code. Something that a Lisp hacker
> pg> might handle by pushing a symbol onto a list becomes
> pg> a whole file of classes and methods. So it is a good
> pg> tool if you want to convince yourself, or someone
> pg> else, that you are doing a lot of work.

The "scaffolding" is merely a standarisation of much of the goodness in
well modularised code. There is nothing "elaborate" about it.

[ I also can't help but highlight the phrase "OOP is like crack for
these people". Are Graham's arguments so weak that he has to resort
such hyperbole? It strikes me as bitter and twisted, but then I'm a
crack addict! ]

Where I disagree with "Lisp hackers", is their obsession with
conciseness (that and all those brackets!). They hate any sort of
formalisation or abstraction being built into the language, and want to
be free to hack those abstractions as they see fit. The touble is that
such freedom doesn't scale well to large scale programming communities,
and it doesn't make their code anymore readable (it just means they have
less characters to type).

Graham makes a big point about how modern Lisp enables you to build the
language abstractions, as well as the problem domain abstractions. But
this is a very autistic approach to development IMO, because so much
about good development is communication. Communication with end-users,
as much as communication with other people that use your code. The
scaffolding of OOP is part of that communication.

> and on another vein (one that you (MM) have tapped into on these
> newsgroups) Graham states:
> pg> There is a kind of mania for object-oriented programming at
> pg> the moment, but some of the smartest programmers I know
> pg> are some of the least excited about it...

I would not disagree that some of the smartest programmers are autistic.
It is just sad but true, that smart languages for smart people will
remain largely isolated to the academic community. I do disagree that
the ability to communicate a components function to other programmers is
the realm of "mediocre programmers".

The irony of this discussion, is that it is the very scaffolding of OOP
that Paul Graham and Mike Mitchell dislike so much, that is providing
the impetus for coding classes using visual modelling tools, as an
alternative to textual code.

Executable UML and MDA standards movements, and tool makers such as IBM,
TogetherSoft, Rational, BoldSoft, SoftWIRE etc. are all built on
translating those visual models to OO textual code. Is there anyone
building visual modelling tools around Functional Languages like Lisp?

--
David

P.S. I find the association of MM with Lisp extremely perverse! <g>

[Mike Mitchell]

On Tue, 27 Aug 2002 01:32:51 +0100, "David Bayley"
<dbayley@spamless.aebacus.com> wrote:

>The availability of re-usable, protocol rich, components across company
>boundaries, is something that has driven the popularity of OOP
>languages. Even the anti-OO Mike Mitchell agrees with this point, in
>that he advocates the easy re-use of other people's OO components (so
>long as he doesn't have to write them himself!).

But it's not the OOP aspect of those components that interests me,
it's the components themselves. Were VBX components written according
to OOD principles?

I lift the hood on my car and study the alternator. It makes juice for
the battery. How it works internally interests me not one iota. If it
breaks I can fit a new part for 50 bucks. End of story.

MM

[Tom Shelton]

"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
news:3d6b3623.686803@news.devx.com...
> On Tue, 27 Aug 2002 01:32:51 +0100, "David Bayley"
> <dbayley@spamless.aebacus.com> wrote:
>
> >The availability of re-usable, protocol rich, components across company
> >boundaries, is something that has driven the popularity of OOP
> >languages. Even the anti-OO Mike Mitchell agrees with this point, in
> >that he advocates the easy re-use of other people's OO components (so
> >long as he doesn't have to write them himself!).
>
> But it's not the OOP aspect of those components that interests me,
> it's the components themselves. Were VBX components written according
> to OOD principles?
>
> I lift the hood on my car and study the alternator. It makes juice for
> the battery. How it works internally interests me not one iota. If it
> breaks I can fit a new part for 50 bucks. End of story.

And there Mike is the beauty of OOP. It is so wonderful you finally got it!
Welcome to the fold!

Tom Shelton

[Mike Mitchell]

On Tue, 27 Aug 2002 11:15:03 -0600, "Tom Shelton" <toms@dakcs.com>
wrote:

>And there Mike is the beauty of OOP. It is so wonderful you finally got it!
>Welcome to the fold!

Excuse I, Mr Clever ****! Show me where I got down and dirty with
classes! Show me where I had to plan for my inheritance! Show me where
I first designed my object hierarchy! Those components could be
written in hand-coded assembler for all I know (or care). I just use
the results, and that's all I want to know (or do).

MM

[Tom Shelton]

"Mike Mitchell" <kylix_is@yahoo.co.uk> wrote in message
news:3d6bea9f.1221944@news.devx.com...
> On Tue, 27 Aug 2002 11:15:03 -0600, "Tom Shelton" <toms@dakcs.com>
> wrote:
>
> >And there Mike is the beauty of OOP. It is so wonderful you finally got
it!
> >Welcome to the fold!
>
> Excuse I, Mr Clever ****! Show me where I got down and dirty with
> classes! Show me where I had to plan for my inheritance! Show me where
> I first designed my object hierarchy! Those components could be
> written in hand-coded assembler for all I know (or care). I just use
> the results, and that's all I want to know (or do).

Again, you get it! That is the whole point of OOP - code reuse. The user
of your library cares nothing about implementation - only the interface and
the results. It is modular programming taken to the next step.

HTH,
Tom Shelton