Squeak-dev May 2002

squeak-dev@lists.squeakfoundation.org

160 participants
453 discussions

Re: Morphic, Dynapad, and Squeak performance
by reic0024＠d.umn.edu 24 May '02

24 May '02

Tommy Thorn <tt1729(a)yahoo.com> wrote: > 1. the use of something like Morphic as opposed > to, say, MVC, is critical for enabling small > apps to be designed on the road Morphic can be a little easier to write for than MVC, which can cause this effect of which you speak. However, if I did choose to move Dynapad "officially" to MVC, it would only be temporary- a new widget set would be created. Something on top of MVC or BitBlt, I cannot say. But if I did go down this road, it would be a long time until anything useful was born out of it, so it's likely not going to be the direction for anything other than the back-burner. > 2. the portability of Squeak is outstanding. In > all likelyhood you/we would move Dynapad with > out as you upgrade your PDA to the next > (necessary more powerful) unit. The choice > between MVC and Morphic will almost certainly > look different at that time and I believe > that designing for the present might hurt > in the future. See above. An extended MVC/new widget set would have three main goals: 1. Have adequate performance on a PDA: I use a 206 MHz SA device, but I can only hope that it'd work OK on something slower- 100 MHz perhaps. 2. Be straightforward to program for: Probably a level up on the abstraction food-chain, but I'm interested in something for creating UIs programatically that's as simple as Tk. I'm working on this for Morphic as well, but it would exist for MVC+ if I choose that route. 3. Have the widgets that make sense on a PDA, and have the ability to create new ones and compose existing ones into a new widget. All this, but the first, is pretty easily achievable in Morphic on Squeak 3.0. > 3. [...] > I stand in awe at how well designed Squeak is > and how works, but never the less, it seems > to me that the performance could be improved > quite a lot if it had to. Well, please do have a whack at it. I, and many others, would love to see what you come up with. > On embedded devices we need better performance > for Morphic to be usable. And one solution to that problem is moving to an older version of Squeak, and still using Morphic. I'm creating 2.6 and 2.8 images with the same setup as the Dynapad R0.01 image, to see what performance is like. CharRecog is a lot faster in 2.6 with the usual text morphs than it is in 3.0- almost as fast as it is in MVC. On VM optimization and native code generation: it's not my thing. I'm a biologist (or will be someday :P ), and have little interest and even less ability in such things. Naturally, I'd love the end-result, but I've no knowledge in the area. If you have the time, please look into it, again, many would be greatful. Very greatful, I'm sure. > I'm convinced that it's technically feasible > to make Squeak on StrongArm fast enough to make > Morphic usable. I'm just not sure if we have > enough collective spare cycles to make it happen. Indeed, I don't doubt that it's possible. However, I'm interested in getting a system that works, something that I can carry around with me and take notes in a way that suits me. For the time being, I think using an older version of Squeak (with Morphic, mind you!) may be the best way to achieve that. Regards, Aaron

1 0

Re: [BUG?] ProtoObject class superclass = nil
by Dan Rozenfarb 24 May '02

24 May '02

Sorry, my mistake. I felt sure because I have tried on 3 different images (with different versions), but probably there is some change set that I filed in all of them. It is strange because I haven't been playing aroud with such issues but, evidence suggests the contrary. Next time I will triple check :-( ProtoObject class superclass returns Class class, as it should be. Dan Rozenfarb --- Dan Rozenfarb <drozenfa(a)yahoo.com.ar> wrote: > Hi all, > I was preparing the class about metaclasses for our > students, which tend to be scared about such issues, > when I got scared myself. > Much to my surprise, ProtoObject class superclass > returns nil. I was expecting it to return Class, > but > nope. Am I wrong? > For further checking, I tried Class inspect and > obtained an inspector over Class class, which is > fine. > Then I looked at subclasses of it, and it DOES > contain ProtoObject class among them. > Is this a bug? > > I have tried it on 2.8, several 3.2 versions and in > 3.0 (which is the one that we recommend to our > students). > > Regards, > Dan Rozenfarb > > __________________________________________________ > Do You Yahoo!? > LAUNCH - Your Yahoo! Music Experience > http://launch.yahoo.com > __________________________________________________ Do You Yahoo!? LAUNCH - Your Yahoo! Music Experience http://launch.yahoo.com

1 0

Re: morphic slowdown
by reic0024＠d.umn.edu 24 May '02

24 May '02

Avi Bryant <avi(a)beta4.com> wrote: > I assume this slowdown is a known thing for those who were around pre-3.0. > What was the change from 2.8 to 2.9? I wondered this too, during the change between 2.8 and 2.9a. It was a pretty signifigant speed difference. I remember asking on the list about it, but I can't remember any difinitive response. Aaron

1 0

Re: [BUG?] ProtoObject class superclass = nil
by Blaine Buxton 24 May '02

24 May '02

Hey, I thought this was interesting. I looked at the class definition of ProtoObject and got the following: ProtoObject subclass: #ProtoObject instanceVariableNames: '' classVariableNames: '' poolDictionaries: '' category: 'Kernel-Objects'. ProtoObject superclass: nil Strange, but reasonable. I guess we need an instance to subclass from initially, but later we set the superclass to nil. I got the following when I inspected things in the workspace: ProtoObject class superclass (returned Class class) But, I got the following when I did this: ClassDescription class superclass (returned Metaclass for Behavior class) Now, it seems I can use any class for the above and I get the same type of object (a Metaclass instance for a class). So, now, my question is why does ProtoObject's metaclass (ProtoObject class) return an instance of Class and NOT nil or metaclass instance used for nil. I understand that there could be a difference in instance and class inheritance heirarchies right? But, for all practical purposes, they are the same. So, if this is the case, why wouldn't ProtoObject class (metaclass of ProtoObject) return the Metaclass instance for Class (if it's inherited from Class) or nil (sicne its instances inherit from nil and thus, the heirarchies are the same). Let me know if you find anything...=) >From: Dan Rozenfarb <drozenfa(a)yahoo.com.ar> >Reply-To: squeak-dev(a)lists.squeakfoundation.org >To: squeak-dev(a)lists.squeakfoundation.org >Subject: [BUG?] ProtoObject class superclass = nil >Date: Thu, 23 May 2002 07:32:55 -0700 (PDT) > >Hi all, >I was preparing the class about metaclasses for our >students, which tend to be scared about such issues, >when I got scared myself. >Much to my surprise, ProtoObject class superclass >returns nil. I was expecting it to return Class, but >nope. Am I wrong? >For further checking, I tried Class inspect and >obtained an inspector over Class class, which is fine. > Then I looked at subclasses of it, and it DOES >contain ProtoObject class among them. >Is this a bug? > >I have tried it on 2.8, several 3.2 versions and in >3.0 (which is the one that we recommend to our >students). > >Regards, >Dan Rozenfarb > >__________________________________________________ >Do You Yahoo!? >LAUNCH - Your Yahoo! Music Experience >http://launch.yahoo.com _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx

2 1

[ENH][Modules] DynamicWorldMenu
by danielv＠netvision.net.il 24 May '02

24 May '02

from preamble: "Change Set: DynamicWorldMenu Date: 24 May 2002 Author: Daniel Vainsencher For 3.3a Prerequisites: the unloadable celeste patches. Make the list of applications in the World menu a registry. Adapt Celeste and IRCChat to use it. Makes IRCChat unloadable, and returns Celeste to the world menu."!

1 0

Re: Best Old Version?
by reic0024＠d.umn.edu 24 May '02

24 May '02

"Les Tyrrell" <tyrrell(a)canis.uiuc.edu> wrote: > Don't know how devoid of features you'd like it to be, but Paul Fernhout has a > modified Squeak 2.2 system that he has stripped down to about 200 classes for > his Embedded Squeak release: That's a little too devoid of features for me. :) My question was worded pretty poorly. I'm mostly curious what version, out of 2.4-2.8 is the most stable, and what possible caveats are for that version. Stripping images isn't a problem for me. Thanks anyway Les! (I've not seen Paul on the list in forever!) Aaron

1 0

Re: A new user question
by goran.hultgren＠bluefish.se 24 May '02

24 May '02

Another good thing to check out are the two first issues of SqueakNews that are available for download: http://www.squeaknews.com/download/index.html Of course I urge anyone to buy them also! :-) There are quite a few articles about teaching children in SqueakNews (though I can't remember exactly what was on the first two). I have all issues sofar and there are lots of good stuff there. regards, Göran

1 0

Re: Behaviors vs Modules
by Anthony Hannan 24 May '02

24 May '02

Jecel Assumpcao Jr <jecel(a)merlintec.com> wrote: > On Friday 22 February 2002 23:38, Anthony Hannan wrote: > > - A Smalltalk expression now knows, via the selectors it calls, which > > protocols and corresponding default behaviors it depends on. So > > starting from a Smalltalk expression we can trace selectors and > > methods recursively and load all default behaviors that will be > > needed. > > Two slight problems: > > 1) #perform: perform: will have to take a selector object that already knows its protocol instead of just a symbol. The selector's protocol will be loaded at runtime when the perform is triggered. > 2) if #foo is used in only one protocol in your system, you don't > have to say which one that is, right? Now imagine that you send your > module to me and I load it into my system which also has a #foo in only > one protocol, but it happens to be a different protocol from yours! The compiled method will have the selector object "foo" in its literals which knows its protocol, not the symbol #foo. So if you send the compiled method to another image (via image segments) then the protocol will go with it. If you just send the source code, your right, "foo" will be looked up in the new environment. But this is the same problem we have with globals now. How do I know "String" in my image means the same as "String" in your image? Cheers, Anthony

3 2

BBC Micros and cooling (was RE: gettimeofday() revisited)
by Peter Crowther 24 May '02

24 May '02

> From: Ian Piumarta [mailto:ian.piumarta@inria.fr] [All relevant content snipped] > PS: Anyone who has encountered an early-revision Acorn "Beeb" > motherboard > might fondly remember having to balance bags of ice cubes > on top of > the video chip during the summer months to achieve a stable > display. ;) I still have one (of my two BBC B's) that has problems if run with the lid on. At least Beebs were quiet. Anyone had any experience of the products at http://www.quietpc.com/ for silencing the jet-engine roar of a modern P4 with a few disk drives? - Peter

4 3

[BUG] Can't write files with Squeak 3.2.7Beta5 on MacOS X 10.1.2
by Andrew P. Black 24 May '02

24 May '02

In an attempt to resolve some other mysterious problems, last night I downloaded a fresh Squeak3.2gamma-4857.image and, while I was there, the Squeak 3.2.7Beta5 VM (this is all on MacOS X 10.1.2). I read some file Ins, saved the image, but could not do a file out of my modified chgangeSet! A little poking about showed that StandardFileStream new primOpen: 'foobar.cs' writable: true returns nil, which gets propagated all the way up to ChangeSet>>fileOut, where nil fails to understand #header. Yes, the FilePlugin is loaded, no the disk is not full. What is weird is that at some earlier point I was successfully able to save the image under a new name, and even after this failure could checkpoint to that image file. Using the checkpointed image, I restarted with the Squeak 3.0Alpha21MTCarbon VM, and was able to do the fileout without any problems. 3.0Alpha21MTCarbon uses FilePlugin 30 May 2001 (i), whereas Squeak 3.2.7Beta5 uses FilePlugin 9 May 2002 (i). So this looks to me like a VM problem. Has anyone else seen this? Andrew

2 3

← Newer
1
...
6
7
8
9
10
11
12
...
46
Older →

Jump to page: