Blender 3D is a great project in my opinion. You should be able to cut the lines of code to 1/10th of C by using Smalltalk. From 1M that's 100K lines which with a programmer coding an average of 10 debugged lines per day would take 10K man-days or 460 man months. Let's say a class size of 20 could reduce that to 23 months or about 2 years.
You could save more work by careful and elegant redesign of the core 3D objects and messages since elegant simplicity will help new people understand how to contribute faster. Therefore you'll have to teach moving to highly understandable object names instead of the usual amateur tagger philosophy where lots of open source software is developed using names that nobody except the developer understands - until the software gets popular and others have to learn it.
-----------------
Koans for learning Smalltalk on the other hand strikes me as another waste of time and books like the entire "patterns" movement. IBM has or had a great little book for introducing people to Smalltalk using its Visual Age line. Even simpler, stick to Smalltalk itself. Have people start exploring the code for Dictionary until they completely master it - able to code it and use it. One could move from there to other key objects. My understanding of Dictionary once got me a job.
------------------
Language popularity is a matter of advertising. Nobody heard of Java until its creator advertised heavily claiming it was the best language of all, when to people who studied languages it was not as good as Smalltalk. Instead of advertising Smalltalk like that Adel Goldberg implemented a greedy pricing strategy which killed interest in ParcPlace Smalltalk creating many intellectual theives who stole Smalltalk's essentials to put objects in other lesser languages like C++, Java, etc..
The way out of that legacy in my opinion is projects like Blender 3D, maybe OpenOffice, and other killer applications in Squeak (or Cuis) which can potentially be so good people won't have a reason to stay with kludge languages.
Kirk Fraser
Teach this:
"Regardless of how talented the introverted software designer may be, forcing himself to also become involved in the outside work of sales and investor meetings will assure him of a far more rewarding life both psychically and financially." - Rabbi Daniel Lapin
Kirk wrote:
Language popularity is a matter of advertising. Nobody heard of Java until its creator advertised heavily claiming as the best language of all,
Perhaps Smalltalk needs a different approach to advertising. I don't think many Smalltalkers believe that Smalltalk is the best of all. It seems to me that there has always been a widespread and justified belief that it will be a stepping stone to something better, such as Self, Strongtalk, Newspeak, or something yet to be created by researchers such as those at VPRI/STEPS, Potsdam, Berne and several other enlightened places.
In a sense it is a victim of its own simplicity: of the ease of making new class libraries and new compilers.
Java might have been sold as the best, but in reality it only had to be better than C++ to stand a chance of winning.
In any case, Smalltalk has had some high profile advertising: by IBM, then by Alan Kay for Squeak, and more recently by Alan again for Croquet, and Randall for Seaside. I think that there is a law of averages that says that even great products with great advertising don't make it.
Even so, there is now a Smalltalk in the hands of each of two million schoolchildren, through the Squeak Etoys and DrGeo II images on the OLPC. Scratch is even more widespread: to the extent that it is a killer application.
The way out of that legacy in my opinion is projects like Blender 3D, maybe OpenOffice, and other killer applications
I agree: killer applications are better than advertising.
I also think that Python and Ruby's strategies of attracting entrepreneurs and hackers are, in the long run, more rewarding than Java's early successes in corporate sales.
It is not too late: Objective-C finally made it after a 25 year run for the prize. So why not Smalltalk?
Can you develop your killer application proposal, please? Are you suggesting a Squeak contribution to Blender, or a fork?
David
p.s. Kirk also wrote:
Koans for learning Smalltalk on the other hand strikes me as another waste of time and books like the entire "patterns" movement.
Forgive me if my history is wrong, but wasn't the patterns movement a joint venture of the Smalltalk and C++ communities? In any case, "Koans" seems to a product of the agile movement, which began entirely in Smalltalk. I intend to waste my time by reading it.
On Wed, Dec 14, 2011 at 3:10 PM, David Corking lists@dcorking.com wrote:
Kirk wrote:
Language popularity is a matter of advertising. Nobody heard of Java until its creator advertised heavily claiming as the best language of all,
Perhaps Smalltalk needs a different approach to advertising. I don't think many Smalltalkers believe that Smalltalk is the best of all. It seems to me that there has always been a widespread and justified belief that it will be a stepping stone to something better, such as Self, Strongtalk, Newspeak, or something yet to be created by researchers such as those at VPRI/STEPS, Potsdam, Berne and several other enlightened places.
One weakness of the Smalltalk environment is that every development image is in essence a fork. A lot of tools help easing the problems that bring, but it is not smooth sailing.
Another issue is what you mention; the grass is always greener on the other side. So much effort and focus is used on developing the newer and better stuff that older tools get half finished and abandoned. That make the Smalltalk experience a little rocky at times.
Karl
In a sense it is a victim of its own simplicity: of the ease of making new class libraries and new compilers.
Java might have been sold as the best, but in reality it only had to be better than C++ to stand a chance of winning.
In any case, Smalltalk has had some high profile advertising: by IBM, then by Alan Kay for Squeak, and more recently by Alan again for Croquet, and Randall for Seaside. I think that there is a law of averages that says that even great products with great advertising don't make it.
Even so, there is now a Smalltalk in the hands of each of two million schoolchildren, through the Squeak Etoys and DrGeo II images on the OLPC. Scratch is even more widespread: to the extent that it is a killer application.
The way out of that legacy in my opinion is projects like Blender 3D, maybe OpenOffice, and other killer applications
I agree: killer applications are better than advertising.
I also think that Python and Ruby's strategies of attracting entrepreneurs and hackers are, in the long run, more rewarding than Java's early successes in corporate sales.
It is not too late: Objective-C finally made it after a 25 year run for the prize. So why not Smalltalk?
Can you develop your killer application proposal, please? Are you suggesting a Squeak contribution to Blender, or a fork?
David
p.s. Kirk also wrote:
Koans for learning Smalltalk on the other hand strikes me as another
waste
of time and books like the entire "patterns" movement.
Forgive me if my history is wrong, but wasn't the patterns movement a joint venture of the Smalltalk and C++ communities? In any case, "Koans" seems to a product of the agile movement, which began entirely in Smalltalk. I intend to waste my time by reading it. _______________________________________________ Beginners mailing list Beginners@lists.squeakfoundation.org http://lists.squeakfoundation.org/mailman/listinfo/beginners
beginners@lists.squeakfoundation.org
-
David Corking -
karl ramberg -
Overcomer Man