Version 6 suggestion

Emailing Ian on the myth of self-documenting code, I was inspired with an
idea on how to get closer to it.  OA should rename every class, method, and
variable that isn't
immediately clear to non-Smalltalkers.

For example:
class URLMonLibrary -> WebLibrary
class WebBrowserShell -> WebExplorer

This new clarity should increase Dolphin's user base.

Another suggestion toward self-documenting code is to minimize spaghetti
code.  I don't have an example in Dolphin but I have one in VisualWorks
where the Smalltalk compiler is exposed.  That compiler is an example of
rat's nest programming.  Taking a look at the difference between it and
clear structured code may inspire improvements in some code.

Kirk W. Fraser

The reaction to my inspired suggestion reminds me of people's reactions to
the gospel.

On soc.religion.quaker I recently posted:
"Why not bypass all that Quaker trivia and grow directly toward Jesus
Christ?"

Of three responses only one had a clue, which actually is amazingly good
results out of 6+ billion people in the world.  Eventually the number of
people who get it will grow.

Perhaps I narrowed the focus of my above post too much by reference to OA.
Perhaps every Dolphin user should refactor the entire image to whatever is
clearest to non-Smalltalk programmers.  People could share these updates and
maybe a whole new sub-group would be started, "Clear Smalltalk."

Kirk Fraser
http://prayer_child.tripod.com/

True Christian wrote:

> The reaction to my inspired suggestion reminds me of people's reactions to
> the gospel.

Well, I for one didn't reply because I couldn't think of anything to say that
wasn't rude.

However this time it sounds more as if you are actually proposing a new
language (a new library for Smalltalk is effectively a new language), in which
case why not start implementing it yourself and see if the world joins
in/converts to it ?

    -- chris

On Fri, 10 Dec 2004 09:19:26 -0800, "True Christian"
<[hidden email]> wrote:

Even better, rename the product to Visual S++
;-)

Re: renaming things for the benefit of non-Smalltalkers:

Why?

"Bob Jarvis" <[hidden email]> wrote
> Re: renaming things for the benefit of non-Smalltalkers:
>
> Why?

1) To attract a much larger number of customers
2) To make it much easier to follow for existing but non-expert Smalltalkers
3) To provide a huge example of how to write easier code.
4) To make code written 6 months ago as easy to modify as code written a
week ago.

"Chris Uppal" <[hidden email]> wrote
> However this time it sounds more as if you are actually proposing a new
> language (a new library for Smalltalk is effectively a new language), in
> which
> case why not start implementing it yourself and see if the world joins
> in/converts to it ?
>
Right now I have a student at Stanford hired to write a new language I
designed, which is not another Smalltalk.  When finished, I'll be writing
code in it.  I'll be using this level of clarity.  If my language doesn't
turn out to be 3 times more productive than Smalltalk, I may reconsider your
suggestion.  Thank you.

This certainly sounds good.  I would be interested to see the data that
supports these points.  Have you done surveys or conducted research
that led you to the conclusion that the names of classes, methods, and
variables in Dolphin Smalltalk is standing in the way of broader
acceptance?  I'm also curious what your criteria for making these names
"immediately clear to non-Smalltalkers" is.  And how would you deal
with migration issues?  Today there is a large body of code that uses
the "old" names.  How would you handle the problem of, for example,
loading packages that refer to "old" class names into an image with the
"new" class names?

"Bob Jarvis" <[hidden email]> wrote
> This certainly sounds good.  I would be interested to see the data that
> supports these points.  Have you done surveys or conducted research
> that led you to the conclusion that the names of classes, methods, and
> variables in Dolphin Smalltalk is standing in the way of broader
> acceptance?

If it's not obvious to you, I suggest you donate to fund such a study.

> I'm also curious what your criteria for making these names
> "immediately clear to non-Smalltalkers" is.

Actually it's part of the whole Smalltalk paradigm to use meaningful names
but only people involved long enough to remember the new paradigm
propaganda, in the days of Smalltalk-80 seem to obey it.

> And how would you deal with migration issues?  Today there is a large body
> of code that uses the "old" names.  How would you handle the problem of,
> for example, loading packages that refer to "old" class names into an
> image
> with the "new" class names?

The new package could be sent with a set of refactoring statements that
could be applied to any code -- automation is sometimes more useful than the
manual mouse.  Of course those who are experts in Smalltalk the way it is
wouldn't have any problem absorbing the hot newness.