. .


Smalltalk Criticism

June 29, 2011 9:55:28.274

This critique seems pretty bitter (and at points inaccurate as well) - but you can't only point to the good stuff, either. See if you think Richard Kulisz has a point.

posted by James Robertson


Re: Smalltalk Criticism

[Randal L. Schwartz] June 29, 2011 11:37:13.333

Wow. I wonder what Smalltalk did to him for him to be so angry at it.

Re: Smalltalk Criticism

[ Terry] June 29, 2011 12:46:37.378

Comment by Terry

If you read his other blog posts you will see that he is just as angry about other topics as well. He must be a rather unhappy person.

Re: Smalltalk Criticism

[anonymous] June 29, 2011 12:59:11.748

But of course, he gains gravitas by dropping f-bombs like candy. Only the truly serious and intelligent do that.

Re: Smalltalk Criticism

[Richard Kulisz] June 29, 2011 14:27:36.849

Ahh, meta-criticism. Nothing spells 'diversion' quite like it. What next, are people going to complain about the typography of my posts? Not enough green in them? Too much blue?

What did Smalltalk do to me? It's very simple: Java (Smalltalk minus minus) triumphed while Self got canned. And then Klein got canned. Smalltalk destroyed all of its superior successors.

Unix did this too. And you might wish to claim that all successful systems do this except LISP DID NOT!

LISP didn't kill off its progeny. Why did Smalltalk have to kill all its progeny like some kind of bloodthirsty psychopath when LISP managed to refrain?

By the way James, saying "it [seems] at points inaccurate" isn't a criticism. It's a perception. Maybe you would like to try your hand at arguing using logic and facts? It's fun!

Re: Smalltalk Criticism

[anonymous] June 30, 2011 4:59:40.124

I think he makes a couple of very valid points, in the sense that Smalltalk hasn't evolved at all (conceptually speaking) since the 80s, which contradicts one of its original objectives.

Projects like Self take purity to the extreme and are really coherent in all of their rules throughout the whole implementation. While Smalltalk does have very little exceptions and is pretty much consistent, it is true that its aim was to be _perfect_, which is what Richard, IMO, is pointing out. Java can't fail because it's aim was never to be a perfect system, but a popular, money-making one.

Even though I agree with many of Richard's points, I'm still a convinced Smalltalker and an irreducible GNU-Linux user. They're not perfect, sure, but I know they're among the best ones out there.

Re: Smalltalk Criticism

[James Robertson] June 30, 2011 8:03:45.181

Ok, Number 6 - #initialize. Not an issue in VisualWorks or Pharo any longer

Number 7 - just wrong. The way #grow typically works is this:

| newSelf |

newSelf := self copyEmpty: newCapacity.

self associationsDo: [:each | newSelf noCheckAdd: each].

self become: newSelf

#copyEmpty just creates a new collection of the same size, and then #become: swaps it with the old collection. What you are thinking of is #copy/postCopy

Most of your other complaints are oddly petty. #at:put: should return self, but then again, so should #add:. It's historical legacy at this point, not some overarching flaw. It also doesn't come up that often.

Re: Smalltalk Criticism

[FDominicus] June 30, 2011 9:31:16.666

I could not care less. I only see one valid argument that is that the Squeak UI is terrible. That's for sure. And his opinion about the state of software development are just "ignorant" There's so much software available this days. And it serves many men so well and just saying that admiring the old fellow (e.g in Genera) is so stupid that it hurts.

As if we (the software developers these days) are so much superior....

 Share Tweet This