Nevin, If one wants to dominate the Smalltalk market, then write a Smalltalk, and a good one. Croquet, IMHO, is missing the mark by being a 3D system first, second, and maybe third, and further into MHO, missing the opportunity to provide a retained mode 3D graphic engine. Note that I am not saying there is anything wrong with the product as delivered, but it could be a great deal more useful to a wide range of users if a few layers were identified and exposed/marketed as useful systems. I also think Tweak is a bad design idea; never do to the language what you can do with/in the language. I see it as a logical extension of inheritance vs. composition; the former is easily over-used when the latter applies, and I see Tweak as a similar trap. To date, I think Pharo is the most likely Squeak fork to result in a good Smalltalk. That is because the process is open, includes just enough iron fist control to make decisions as needed, and because making a good Smalltalk _that people can use as they see fit_ is the stated objective. Bill -----Original Message----- From: pharo-project-bounces@lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Stéphane Ducasse Sent: Monday, July 06, 2009 4:32 AM To: nevin@smalltalkpro.com Cc: Pharo Development Subject: Re: [Pharo-project] Squeak Forks Popularity & Trends Estimate Hi nevin our goal is not predominancy :) It was never. I would be happy that squeak offers a really good multimedia platform. Our goal is to deliver good open source smalltalk. Our goal is to get agile, rethink the system, use good software engineering practices (systematic smallLnt rules), tests.... We want to propose/accept/evaluate new solutions: first class package, first class instance variables, better refactoring support, may be modules, security, new compiler.... Smalltalk deserves a clean and powerful implementation and we will work on that hard. Stef PS: you can forward this answer to squeak-dev if you feel the need. On Jul 6, 2009, at 3:17 AM, Nevin Pratt wrote:

There's no right or wrong answer to the following three questions. And, every response is meaningful, because it helps reveal the general sense of direction that the community is feeling.

*********************** Question #1: Of the various Squeak forks (such as Croquet, Pharo, etc.), what is your feeling of their relative popularities at this point in time, and how does that compare with the current popularity of the base Squeak distribution?

Question #2: This is related to the above-- what is your feeling of the relative popularity *trend* of each of the forks, and how does that compare with your feeling of the *trend* of the future popularity of the base Squeak distribution?

Question #3: What is your justification for your answers to the above two questions? ************************

I'll begin with my own answers...

...I'm beginning to feel like the Pharo fork currently now has maybe 50% or more of the popularity of the base Squeak distro, and I'm beginning to feel like Pharo is headed to become the dominant distribution in the future, surpassing the original Squeak distro. And I base that feeling on several observations:

1. The email traffic comparison on the "Pharo-project" email list compared to the email traffic on "The general-purpose Squeak developers list", and the observed trend in such traffic.

2. The number of developers posting on the Pharo-project email list.

3. The specific developers posting on the Pharo-project email list (known Squeakers who have contributed heavily to Squeak in the past).

4. The fact that Seaside development has moved to Pharo.

At one time I wondered if Croquet would become the dominant Squeak fork, for the simple reason that it had a business organization "officially" supporting it, similar to what Squeak had in the Disney days. But now I think I would bet my money on Pharo.

What do you think? And why?

Also, remember, there's certainly nothing wrong with several "dominant" forks existing. But of those, one of them will still be the most popular overall. And I'm curious about thoughts of which one that is, and which one that will become, and why you think so.

Nevin

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On Jul 6, 2009, at 9:33 PM, Hernan Wilkinson wrote:

Hi Stef, have you thought about have not only first class instance variables but first class variables? (no matter the scope). yes this was included. Now we want to really experiment and provide a real and working solution. I just mentioned that because Smalltalk deserves a better instantiation. :)

I think that would be interesting... observers on variables would be pretty easy to implement, type inference (or type recollection)... not sure about the trade-off cost/convenience,

We did a simple implementation with marcus and the implementation we came up with did not have **any** runtime penalties. Now the classBuilder is so brittle that to make sure that we would have a robust implementation we would have to rewrite it and this is a real nightmare... look at the comment. for now we just created an extra instance variable that serves to allocate our instances :)

but looks interesting... accessing local variables could be made sending messages to thisContext... anyway, just an idea.

so far having plain well working first class instance/class variable would be a really good achiemenet with a good mop. Would be great. Imagine people could implement tweak like field without requiring copying the compiler and adding XML syntax :)

On Mon, Jul 6, 2009 at 6:31 AM, Stéphane Ducasse <stephane.ducasse@inria.fr

...

wrote: Hi nevin

our goal is not predominancy :) It was never. I would be happy that squeak offers a really good multimedia platform. Our goal is to deliver good open source smalltalk. Our goal is to get agile, rethink the system, use good software engineering practices (systematic smallLnt rules), tests.... We want to propose/accept/evaluate new solutions: first class package, first class instance variables, better refactoring support, may be modules, security, new compiler....

Smalltalk deserves a clean and powerful implementation and we will work on that hard.

Stef

PS: you can forward this answer to squeak-dev if you feel the need.

On Jul 6, 2009, at 3:17 AM, Nevin Pratt wrote:

...

There's no right or wrong answer to the following three questions. And, every response is meaningful, because it helps reveal the general sense of direction that the community is feeling.

*********************** Question #1: Of the various Squeak forks (such as Croquet, Pharo, etc.), what is your feeling of their relative popularities at this point in time, and how does that compare with the current popularity of the base Squeak distribution?

Question #2: This is related to the above-- what is your feeling of the relative popularity *trend* of each of the forks, and how does that compare with your feeling of the *trend* of the future popularity of the base Squeak distribution?

Question #3: What is your justification for your answers to the above two questions? ************************

I'll begin with my own answers...

...I'm beginning to feel like the Pharo fork currently now has maybe 50% or more of the popularity of the base Squeak distro, and I'm beginning to feel like Pharo is headed to become the dominant distribution in the future, surpassing the original Squeak distro. And I base that feeling on several observations:

1. The email traffic comparison on the "Pharo-project" email list compared to the email traffic on "The general-purpose Squeak developers list", and the observed trend in such traffic.

2. The number of developers posting on the Pharo-project email list.

3. The specific developers posting on the Pharo-project email list (known Squeakers who have contributed heavily to Squeak in the past).

4. The fact that Seaside development has moved to Pharo.

At one time I wondered if Croquet would become the dominant Squeak fork, for the simple reason that it had a business organization "officially" supporting it, similar to what Squeak had in the Disney days. But now I think I would bet my money on Pharo.

What do you think? And why?

Also, remember, there's certainly nothing wrong with several "dominant" forks existing. But of those, one of them will still be the most popular overall. And I'm curious about thoughts of which one that is, and which one that will become, and why you think so.

Nevin

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

I would insert the phrase "for you" several times into your post. That is, is Python a better language (for me) at this time. Or, was Python an easier language (for me) to learn than Smalltalk. If you come late enough to the party (and the party is large enough), you can Google for all your answers. Smalltalk's popularity spiked in the early 90s (until Java popped its bubble). You couldn't Google for Smalltalk answers then largely due to the lack of Google ;-). I learned Smalltalk just after its peak (or maybe at its peak, hard to say). I'd seen Java first and even learned how to build simple apps. By then, I already knew C, C++, and Pascal. I cut my teeth on Commodores (3584 BASIC BYTES FREE!), Apples, and Ataris. Smalltalk was easy for me, but there were many resources available to me: First, (and probably most important) a paying job writing Smalltalk. 60+ hours a week of paid practice doesn't hurt. Access to commercial Smalltalks (VisualAge and VisualWorks) and an employer to pay the $2K-$5K developer license fees. Time working with 50+ Smalltalkers all working on the same project. An active and growing Smalltalk Users Group in my city. An active subscription to the Smalltalk Report magazine (with great articles by guys like Kent Beck). Access to all of the back issues of the Smalltalk Report. Many printed Smalltalk manuals from IBM and ParcPlace. Dozens of IBM Redbooks showing how to do just about anything you would need in the enterprise. I think Andy Bower had it right - "Smalltalk is dangerous. It is a drug. My advice to you would be don't try it; it could ruin your life." It certainly ruined me for most other languages. I've learned Java, Ruby, Python, and Perl in that order since Smalltalk. Java, Ruby, and Perl were for work and Python was a curiosity. I don't use any of them when I have a choice because I'm so much more productive (and happy) in Smalltalk. I love Pharo because it is working to make a better Smalltalk. I'm lucky today because I get to choose my tools at work which means I get to use Pharo! On Thu, Jul 9, 2009 at 11:29 AM, Robert Peters<thewinterlion@gmail.com> wrote:

Hello everyone,

1. Whichever fork is most professional will win.  I believe this was one of the goals of the Pharo project, so I'd have to say Pharo will eventually lead.

2. Smalltalk in all its forms will fail unless it can grab developer interest from outside the current community.  Pharo may be better than Squeak, but if everyone moves from Squeak to Pharo, your audience is still too small to be significant.

I'm a web-application developer (Django, GAE) and self-taught Python coder. I learned Python because it was easy, and there is a wealth of information available on the internet.  If I have any problem in Python, I simply type 'python issue' into Google and within a few clicks I'm usually looking at the answer.  When I tried to learn Smalltalk, it was intensely frustrating. There were no on-line resources, no tutorials, no explanations.

I think the real question is, "Is Smalltalk a better language than Python?" By better, I mean, "Can I get done more coding tasks faster, in a more sustainable way?"

Right now, the answer to this question is a resounding 'No'.  But that answer might change with a slick Smalltalk package, and an abundance of developer resources.  As it is, Smalltalk is a language for eccentrics.  But it doesn't need to be that way.

All of this to say that I really like the beautiful simplicity of Smalltalk's main paradigm - Message Passing.

It's simple.  It's elegant.  But learning it is a stone bitch, and once you've learned it, it has only limited value (compared, for example, to the wealth of things one can do in Python).  I wish this would change. -- View this message in context: http://n2.nabble.com/Squeak-Forks-Popularity---Trends-Estimate-tp3211130p323... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Yes, you're quite right, David. But my point is that my particular situation is representative of everyone who is not currently in the Smalltalk community. I posted because I want the Smalltalk community to grow, but it's peak has passed. There are no SUGs, there is no Smalltalk magazine, there are no web resources. And the Smalltalk community must grow if it is to be taken seriously. I cannot go back in time and learn it in the 'peak'. I must learn it now, with the resources that are available now. These resources are lacking. If a person was equally proficient in Smalltalk and Python, which would they produce better code faster in? I suspect the answer may be Smalltalk, because of its one, beautiful, paradigm. But I, and millions like me, will never know. I'm hoping Pharo will be the first steps in changing that. Kind regards, Robert David Mitchell-6 wrote:

I would insert the phrase "for you" several times into your post. That is, is Python a better language (for me) at this time. Or, was Python an easier language (for me) to learn than Smalltalk.

If you come late enough to the party (and the party is large enough), you can Google for all your answers. Smalltalk's popularity spiked in the early 90s (until Java popped its bubble). You couldn't Google for Smalltalk answers then largely due to the lack of Google ;-).

-- View this message in context: http://n2.nabble.com/Squeak-Forks-Popularity---Trends-Estimate-tp3211130p323... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Dear David and Stephane, Thank you for your replies. David, you make excellent points. I did not know about VW being free and will try it out. And you're certainly right that however much Smalltalk web resources are smaller than Python's, it is certainly a much better environment for learning Smalltalk now than it was 20 years ago. And Stephane, I say 'your' and not 'our' because if my wife found out I got involved in yet another cause on the web, she'd kill me! And also, I can't code in Smalltalk, so I'm very much a spectator. But I did not intend for my original post to be negative. I absolutely believe Pharo is 100% on the right track, and a critical component in a potential Smalltalk resurgence. So I'll keep waiting, and when 1.0 comes out, I'll give it a try again. Kind regards, Robert David Mitchell-6 wrote:

Sorry, I think my intent got lost in my joy for nostalgia.

There was a big Smalltalk party but it ended. Then it was quiet for a while (but the party was underground).

Now, the party is getting started again.

You may just be a little early coming to the second party.

Things you have today that I didn't have in 1994: You can download (Pharo, GNU Smalltalk, Squeak and use it however you like). There are two free, open source, and maintained cross dialect web frameworks: Seaside and AidaWeb (we wrote our own and it paled in comparison) You can use Gemstone for free for small sites. You can use Magma as a free OODB. You can use VW for non-commercial (including to learn Smalltalk). The Squeak list exists. (All I had was comp.lang.smalltalk). Many good Smalltalk blogs. Almost all of the old Smalltalk books are available free as PDF. Many of the new Smalltalk books are available free and from Lulu. Most of the classics can still be bought used on Amazon. James Roberson's smalltalk podcasts. James Robertson's nearly daily Smalltalk screencasts. Randal Schwartz promoting Smalltalk squeak irc channel

On Thu, Jul 9, 2009 at 1:46 PM, Robert Peters<thewinterlion@gmail.com> wrote:

...

Yes, you're quite right, David.  But my point is that my particular situation is representative of everyone who is not currently in the Smalltalk community.  I posted because I want the Smalltalk community to grow, but it's peak has passed.  There are no SUGs, there is no Smalltalk magazine, there are no web resources.  And the Smalltalk community must grow if it is to be taken seriously.

I cannot go back in time and learn it in the 'peak'.  I must learn it now, with the resources that are available now.  These resources are lacking.

If a person was equally proficient in Smalltalk and Python, which would they produce better code faster in?

I suspect the answer may be Smalltalk, because of its one, beautiful, paradigm.  But I, and millions like me, will never know.

I'm hoping Pharo will be the first steps in changing that.

Kind regards, Robert

David Mitchell-6 wrote:

...

I would insert the phrase "for you" several times into your post. That is, is Python a better language (for me) at this time. Or, was Python an easier language (for me) to learn than Smalltalk.

If you come late enough to the party (and the party is large enough), you can Google for all your answers. Smalltalk's popularity spiked in the early 90s (until Java popped its bubble). You couldn't Google for Smalltalk answers then largely due to the lack of Google ;-).

-- View this message in context: http://n2.nabble.com/Squeak-Forks-Popularity---Trends-Estimate-tp3211130p323... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

-- View this message in context: http://n2.nabble.com/Squeak-Forks-Popularity---Trends-Estimate-tp3211130p323... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

On Jul 9, 2009, at 6:29 PM, Robert Peters wrote:

Hello everyone,

1. Whichever fork is most professional will win. I believe this was one of the goals of the Pharo project, so I'd have to say Pharo will eventually lead.

this is not our goal. Our goal is to deliver and create a biotope for other people to create their world and wealth.

2. Smalltalk in all its forms will fail unless it can grab developer interest from outside the current community. Pharo may be better than Squeak, but if everyone moves from Squeak to Pharo, your audience is still too small to be significant.

why your and not our?

I'm a web-application developer (Django, GAE) and self-taught Python coder. I learned Python because it was easy, and there is a wealth of information available on the internet. If I have any problem in Python, I simply type 'python issue' into Google and within a few clicks I'm usually looking at the answer. When I tried to learn Smalltalk, it was intensely frustrating. There were no on-line resources, no tutorials, no explanations.

This is true but we are working on fixing that; squeak by example and other.... they are free books

I think the real question is, "Is Smalltalk a better language than Python?" By better, I mean, "Can I get done more coding tasks faster, in a more sustainable way?"

Right now, the answer to this question is a resounding 'No'. But that answer might change with a slick Smalltalk package, and an abundance of developer resources. As it is, Smalltalk is a language for eccentrics. But it doesn't need to be that way.

All of this to say that I really like the beautiful simplicity of Smalltalk's main paradigm - Message Passing.

It's simple. It's elegant. But learning it is a stone bitch, and once you've learned it, it has only limited value (compared, for example, to the wealth of things one can do in Python). I wish this would change.

I do not think so. Message passing is simple and you can do a lot with it :) But give a try. I mean a real one :) Stef

-- View this message in context: http://n2.nabble.com/Squeak-Forks-Popularity---Trends-Estimate-tp3211130p323... Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project