Lambda the Ultimate

inactiveTopic GaTech SqueakEnd 2002
started 1/19/2002; 2:21:07 AM - last post 1/7/2003; 1:26:23 PM
Ehud Lamm - GaTech SqueakEnd 2002  blueArrow
1/19/2002; 2:21:07 AM (reads: 2313, responses: 12)
GaTech SqueakEnd 2002




May 3-5, 2002 College of Computing Georgia Institute of Technology

The tentative schedule looks interesting.

But using Squeak as a first language? Now that's courage. I was stoned for even proposing such mainstream languages as Haskell and Scheme. I can just hear them: "Squeak? What's Squeak? Oh. It's Smalltalk? Well Java is better isn't it?"

Posted to teaching/learning by Ehud Lamm on 1/19/02; 2:22:22 AM

Dan Moniz - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 6:03:04 AM (reads: 1606, responses: 1)
GA Tech does, however, have a pretty well-entrenched Squeak community. Mark Guzdial teaches there, and at least one course uses Squeak throughout and is a mandatory part of the undergrad CS curriculum, from what I understand.

Ehud Lamm - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 9:19:51 AM (reads: 1669, responses: 0)
What I wrote about Squeak as a first language was a bit tongue in cheek, if anyone failed to notice

I'd love to know how CS students respond to Squeak. Don't they want more marketable skills?

rev - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 11:20:39 AM (reads: 1597, responses: 1)
Browsing some of the class swikis and speaking with Lex Spoon (grad student at GaTech), Squeak is met with a little adversity. I do imagine this is largely because a lot of CS students come into a CS program with knowledge of more "marketable" languages already. My school uses nothing but marketable languages, except for one bleeding class that uses CL. Here, it seems, the focus is a bit much on marketability rather than actualy CS; outputing programmers rather than scientists. It's a shame.

Ehud Lamm - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 11:32:29 AM (reads: 1661, responses: 0)
Squeak is met with a little adversity.

This is a little adversity as opposed to little adversity?

My experience is that students come thinking they know C++ or Java, when all they know is syntax and surface semantics, and are very angry when someone tries to teach them things of substance.

I much prefer those who come admitting they know nothing about computers. They have hope.

Paul Perry - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 8:48:10 PM (reads: 1564, responses: 0)
Adam Bosworth, the former microsoft manager that started the XML group I think summed it up best for me: SOAP is more loosely coupled, course-grained, and asynchronous than CORBA. Check the following article: Adam Bosworth, the former microsoft manager that started the XML group I think summed it up best for me: SOAP is more loosely coupled, course-grained, and asynchronous than CORBA. Check the following article: ???

Chris Rathman - Re: GaTech SqueakEnd 2002  blueArrow
1/19/2002; 8:55:22 PM (reads: 1566, responses: 0)
I guess in my old age, I'm becoming more and more a data analyst, writing dozens of SQL queries and scripts per day - some permanent fixtures, many more disposable. Wish someone would come up with a better alternative than SQL for data analysis, but I haven't seen anything come close. In the meantime, there are two computer languages that I've found to be brimming with energy and ideas - Haskell and Squeak.

When I first made the conversion from C to C++, the ideas of OO programming just didn't click in the brain. Too many years spent hacking at low level assembly and C code resulted in a short circuit preventing me from seeing what it was Bjarne was trying to get at. Sure some of the C++ extensions made life a bit easier, but the idea of OOP never quite took hold. It was not until I set down with Smalltalk and started experimenting with a pure OO language that it made any sense.

I do think that any serious student of C++ should get acquanted with a language where OO is not just one more language feature but rather is integrated to the core - Eiffel & Smalltalk being the best candidates that I can think of. That said, anyone that messes about with Smalltalk will likely become disillusioned with C++ as it's glaring weaknesses in OO programming become all too obvious - things like reflection pale in comparison. But, then again, I'm not one that believes that you necessarily should be in awe of the tools that you use to solve problems. Those who know the limitations are far more likely to see the weakness in their efforts - resulting in more robust software than otherwise.

Smalltalk brought MVC and Graphical User Interfaces to the programming language community. Lots of languages have adopted and adapted this approach. Beyond these efforts, Squeak tries to extend the idea of building user interfaces with Morphic worlds, based upon the experimental work that took place, but never quite realized, with the Self language.

Anyhow, count me as a fan of the Squeak community. The students at Ga. Tech will probably be better C++ and Java programmers as result of their exposure to Squeak, than those that strictly hone in on the trivia associated with the details of those languages. Not sure whether the ambition of Ingals and Kaye will ever pan out, but it's sure nice to see people try to extend the definition of programming languages and environments, not accepting that we've reached nirvana by copying the work of others.

Ehud Lamm - Re: GaTech SqueakEnd 2002  blueArrow
1/20/2002; 8:43:30 AM (reads: 1547, responses: 0)
Chris, I agree with you. I just suggested that the students themselves might have a different opinion

Mark Guzdial - Re: GaTech SqueakEnd 2002  blueArrow
1/21/2002; 7:08:38 AM (reads: 1527, responses: 1)
We have had problems in the past with students complaining about the non-marketability of Squeak. That problem went away a couple of semesters ago when I started directly addressing it in lecture on Day #1 of the course ( I tell students that the point of the course is NOT to teach marketable skills, but to teach them how computer scientists think, esp. with a historical perspective. That's helped alot -- we have had almost no complaints the last two semesters.

The issue of Squeak as a first language is complex, and I do think that Stephen Pope and the folks at UCSB are brave for doing it. Elliot Soloway and I have just written a paper for CACM arguing for a multimedia approach to teaching CS1 that Squeak is perfect for.


P.S. We have space for more workshop/SqueakEnd attendees, if any of you would like to come! New things are getting added to the schedule daily, so keep an eye on it.

Ehud Lamm - Re: GaTech SqueakEnd 2002  blueArrow
1/21/2002; 7:21:32 AM (reads: 1577, responses: 0)
That problem went away a couple of semesters ago when I started directly addressing it in lecture on Day #1 of the course

I've been doing the same thing with our Ada course, but I can't say I got the same result. Students are still very grumpy.

Enough so that it hinders their learning, since many are too busy complaining instead of learning...

Why do you think multimedia is good for CS1? I know it can be motivating (though personally I find algorithms much more interesting than graphics). Doesn't multimedia take too much attention away from core CS1 topics?

(Indeed, what are course CS1 topics? I am still not sure. I have a couple of course approaches I'll just have to try someday, to see which one works best).

Brent Fulgham - Re: GaTech SqueakEnd 2002  blueArrow
1/21/2002; 12:22:22 PM (reads: 1471, responses: 1)
While I was a UCSB graduate, friends of mine at Cal Tech tell me that they have lately had near-riots as first year CS students gripe about having to learn Scheme. The arguments are the same as those you cite in the Squeak example (marketability, lack of familiarity).

I must admit to having walked into my first programming course (FORTRAN of all things!) thinking that my years of PASCAL programming in High School had provided me with just about everything I needed to know. I was certainly wrong, and find that even now as a "Professional" I still find much to learn from Haskell, Scheme, and other so-called "research-oriented" languages.

In fact, an interesting discussion on this very topic was held on the plt-scheme mailing list.

Ehud Lamm - Re: GaTech SqueakEnd 2002  blueArrow
1/21/2002; 12:29:57 PM (reads: 1559, responses: 0)
I like this story. However, it is basically an argument from authority. I prefer to try to convince them. This is impossible to do before they have enough experience, of course.

Catch-22, I guess.

One thing that comes strongly from this story, and I think is true in general, is that the charisma of the instructor plays a large role in introductory courses. This is true, I think, in most fields.

An aside: Someone may have built large projects in C, and still be wrong about choosing a teaching language. Much as I like the conclusion, I know that the ineference is invalid

Toby Reyelts - Re: GaTech SqueakEnd 2002  blueArrow
1/7/2003; 1:26:23 PM (reads: 833, responses: 0)
I was perusing the many months worth of archived postings here and couldn't help but chuckle when I saw this thread on Squeak. I actually took Mark's class and really enjoyed it. (Mark is an excellent professor - which lends some form of truth to Ehud's statement about charisma). I believe my Squeak-based windowing system (Vistas) still lives on at Tech to this very day.

> I'd love to know how CS students respond to Squeak.

Well, I would say that advanced students really enjoy the opportunity to expand their horizons. IMHO, this is one aspect of pedagogy which Tech got right. Each programming course uses the most appropriate language to teach the applicable theory - Java, C, C++, Smalltalk, Lisp, Prolog, Asm, etc... For example, Mark uses Squeak to teach his class on Modeling and Design. This helps students learn to match the right tool to the right thumb.