Squeak-dev January 1999

squeak-dev@lists.squeakfoundation.org

143 participants
293 discussions

Squeak on ComputerWorld
by Mark Wai 27 Jan '99

27 Jan '99

There is a tiny little bit coverage (and even some photos of Engelbart and Kay) about Squeak in Computerworld January 18, 1999. Vol 33. No.3 (under the Technology section) To quote some lines from there: "Thirty years later, nothing has evolved" -- Hal Bennett "The place where you die (referring to Squeak I believe)" -- Alan Kay "The symposium was more of a glorious glance backward than an outline of the future" -- the reporter Enjoy -- Mark Wai Wator InnoVision

2 1

Re: ThingLab
by Dan Ingalls 26 Jan '99

26 Jan '99

>I may have missed this someplace, but in the method > > ThingLabInstance makeMethodFor: message > >there are constructs like: > > recursive ifTrue: [#copy.instance:] ifFalse: [#set.instance:] > >What are these blocks? Les - For the benefit of others, this expression occurred in the following construct... copyNode _ MessageNode new receiver: defNode selector: (recursive ifTrue: [#copy.instance:] ifFalse: [#set.instance:]) arguments: (Array with: (encoder encodeVariable: 'self')) precedence: #copy.instance: precedence from: encoder. so copyNode is being set to a MessageNode whose selector is one of two Symbols. To understand where this leads, you have ot jump to TL's def for doesNotUnderstand:, which begins with... "added ThingLab code follows: " (self tryCompilingThingLabCodeFor: aMessage selector) ifTrue: [^self perform: aMessage selector withArguments: aMessage arguments]. This in turn will compile the necessary sharing or copying method on demand, but I haven't tracked it down any farther than that. Some of it may not even work yet, as I don't think any of the examples I tried triggered this code. - Dan

1 0

Morphic/Scripting Qs: Text areas and colors
by Mark Guzdial 26 Jan '99

26 Jan '99

I have a couple of questions about Morphic and the scripting system as I play further with SqueakPages: - Is there a scrollable text area that I can drag onto a SqueakPages book and have it work? All the ones that I can find (e.g., PluggableTextMorph) don't accept text on their own. They expect the model to handle that. Before I go building one, is there a scrollable simple-model text area already defined somewhere that I haven't found? - How does one figure out the color of something in order to tell a Player to do something when it is over that that color? Is there a way of getting the color info or picking up the color for use in a scripting tool? Thanks! Mark -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 (404) 894-5618 : Fax (404) 894-0673 : guzdial(a)cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html

2 1

Re: Problems generating 2.3 VM on Linux
by Peter William Lount 26 Jan '99

26 Jan '99

Hi, SystemDictionary instances in the 2.3 image do understand the method "snapshot:andQuit:". Check which image your using. Maybe the error is a result of the earlier error of $DISPLAY? What are some examples of correct values for $DISPLAY (a unix shell variable I presume)? Peter Matthias Ernst and Joern Eyrich wrote: > > out of memory > > > > 1080610912 Object>doesNotUnderstand: > > 1080610736 Object>doesNotUnderstand: > > <snip ~4000 lines> > > > 1079827008 Object>doesNotUnderstand: > > 1079826832 Object>doesNotUnderstand: You get this (from here downwards) if you haven't got your $DISPLAY set right. From here upwards there obviously is a bug. Doesn't SystemDictionary understand snapshot:andQuit: anymore ? It did in 2.2 > > 1079786892 SystemDictionary>snapshot:andQuit: > > 1079786784 ScreenController>quit > > 1079786424 ScreenController>yellowButtonActivity > > 1079786156 MouseMenuController>controlActivity > > 1079786080 Controller>controlLoop > > 1079786004 Controller>startUp > > 1079786232 [] in ControlManager>activeController: > > 1079786308 [] in BlockContext>newProcess > Matthias

2 1

ANNC: Linux Sound 6.2
by Lex Spoon 26 Jan '99

26 Jan '99

I stuck a newer version of my Linux sound support at: http://www.cc.gatech.edu/~lex/squeak/sound It's also available from the DownloadingForUnix page off the Swiki. This version has a few improvements: - startPlayingImmediately: actually starts playing pretty quickly - no mods to the image are needed It still requires pthreads. After some discussion on this list, it's clear that pthreads aren't strictly a requirement, because you can use processes instead. In the most direct process-based implementation, one would simply fork() a mini-sound server when Squeak starts up, andcommunicate with it using pipes. The only real disadvantage of this scheme is that there is some overhead in copying sound samples across pipes, when they could just be written straight to the kernel. This overhead is probably not significant, though, because of all the copies and accesses a sound buffer goes through anyway before it actually gets submitted for playing. I think a process-based implementation would be better than the threads-based one, but I am unlikely to work on such a version myself. Finally, there are a couple of problems the current version has, that I don't think can be fixed by just changing the sound driver: 1) Buffer sizes. There is no way to increase the low-level buffer size, and there is no way to tell Squeak that the actual buffer size being used isn't as big as it asked for. If you see a lot of "dumping 22000 frames" messages, then you either need to decrease the buffer size Squeak wants (see SoundPlayer's class initialization method), or increase the buffer size Linux is using (recompile your kernel....). 2) Simultaneous playing and recording I can't get to work. If I set CanRecordWhilePlaying to true, then startRecording() never actually gets called by the SoundRecorder! Does full-duplex audio work on *any* version of Squeak? Enjoy :) Let me know if you try it, whether it works or fails, and I'll start making a list on the web page of configurations that work/don't work. In particular, I've never heard from anyone who uses Linux on a big-endian processor. Lex

1 0

Problems generating 2.3 VM on Linux
by Joern Eyrich 26 Jan '99

26 Jan '99

I have successfuly built and run the 2.1 and 2.3b VMs with the files from ftp://alix.inria.fr/pub/squeak/unix/. But when I tried to make a VM with the new 2.3 interp.c, it built fine, but on startup only produced the following on the terminal: > > out of memory > > 1080610912 Object>doesNotUnderstand: > 1080610736 Object>doesNotUnderstand: <snip ~4000 lines> > 1079827008 Object>doesNotUnderstand: > 1079826832 Object>doesNotUnderstand: > 1079786892 SystemDictionary>snapshot:andQuit: > 1079786784 ScreenController>quit > 1079786424 ScreenController>yellowButtonActivity > 1079786156 MouseMenuController>controlActivity > 1079786080 Controller>controlLoop > 1079786004 Controller>startUp > 1079786232 [] in ControlManager>activeController: > 1079786308 [] in BlockContext>newProcess I basically took the 2.3b files from Ian and just substituted the interp.c I had generated from the 2.3 image running on a 2.2 VM. What am I doing wrong here? Joern

2 1

Re: Order of precedence
by Bijan Parsia 26 Jan '99

26 Jan '99

On Tue, 26 Jan 1999, Matthias Ernst wrote: > glenn wrote: > > [snip] > > > > Sure, those adults or children who have been initially taught a particular > > precedence order in certain programming languages seem to come with a > > perferred "use", but they already have the concept of precedence in > > computer languages. > > This is not a matter of which programming language one learns but the > fact that > they all imitate the mathematical notation of multiplication and > addition. And > everyone learns that in maths, '*' (which actually is a \cdot in maths) > has a > higher precedence than '+'. Or adjacency, or 'x' ('times' sign, which is what I used in elementary school; got very muddled when I encountered cross-product), or... [snip] > > d) cater to adults and children who have preconceived rules set by their > maths teacher. I had a dicussion with someone who thought that Smalltalk's operator precedence was wierd until I pointed out that *read aloud* left to right makes perfect sense: 2*3+7 as "Two times three plus seven." (Of course, pauses and inflections can alter precedence, the way parentheses do.) This is particularly true for very long expressions, e.g.: 2*3+4-8/2*5+9+0-57... In so far as Smalltalk should be read aloud (and I think it should) that seems to be some reason to favor the left to right only precedence rule. Cheers, Bijan Parsia

1 0

Re: Order of precedence
by glenn 26 Jan '99

26 Jan '99

I'm an adult and I don't think that 3+4*5=23. I don't think necessarily that 3+4*5=60 either. I don't think adults use any particular precedence order. Sure, those adults or children who have been initially taught a particular precedence order in certain programming languages seem to come with a perferred "use", but they already have the concept of precedence in computer languages. In fact, event they have learned the concept of left-to-right (or right-to-left or some such) precendence order to handle expressions such as 3-4-2. So all they need to deal with is going from two rules to one. Smalltalk does not have a lack of precedence order for infix operators, it just has a singe order, left-to-right. So it seems to be the options are a) disallow multiple-infix operators b) set a single precedence rule, such as left-to-right c) cater to adults and children already computer-skilled who have preconceived notions set by some other computer languages which have multiple rules, one operator-based and the other left-to-right (or right-to-left or some such). I still think Smalltalk's choice of b) is best and works for children and adults. glenn At 02:53 PM 1/25/99 -0500, Ian Bicking wrote: >I was thinking a bit on something talked about a while before -- that Squeak >should support the child up to the adult. I think Squeak is doing a good >job of supporting the adult, but what about the child? > >One of the problems I see of using Squeak with children is the lack of order >of precedence. The problem is that introducing Squeak to a child would be a >bit dangerous in this regard. Order of precedence is *not* intuitive -- >quite the contrary, Smalltalk *is* intuitive in this regard. In this way, >Squeak would reinforce exactly the wrong notions for the child. > >Modern math curriculums seem to be quite careful about avoiding strings of >operations without using explicit order of precedence (lots of parenthesis) >until the later grades. Unfortunately (or fortunately) a computer language >can't do that because a child can explore the language and discover >precedence (or lack thereof) on their own. > >Of course, the parser could be adapted fairly easy to create a special view >of Squeak where precedence existed. But I think it would be nice if >children really could use exactly the same language and syntax that adults >use. > >So, can we (should we) make Squeak believe 3+4*5=23 ? > > > -- Ian Bicking > > >

1 0

Q: DataStream for in ram serialization, or?
by "Gran" Hultgren 26 Jan '99

26 Jan '99

Hi everybody! I am really getting into Squeak and thought I should try to whip something interesting together and why not some sort of "Squeak RMI"? The usual technique by subclassing nil for proxies etc. seems to be the way to go and then looking for serializing code I found DataStream which I thought could be easily used together with some Socket stuff. But that class seems to be tailored for file writing and reading, or am I wrong? Is DataStream (or descendant) the way to go? Or perhaps it should be tweaked a bit so that it can be used together with Sockets etc? Any thoughts? I have not dug around too much so perhaps I am missing something. regards, Gran === Gran Hultgren | goran.hultgren(a)inspire.se | OO Consulting Inspire | Hem: +46 8 54022950 | http://www.inspire.se Karlavgen 18 | Mobil: +46 708 813936 | icq#:6136722 114 31 Stockholm | "All my opinions are MINE, MINE, MINE!!!" _________________________________________________________ DO YOU YAHOO!? Get your free @yahoo.com address at http://mail.yahoo.com

1 0

Re: Mouse events
by Colin Sarsfield 26 Jan '99

26 Jan '99

johnm(a)wdi.disney.com wrote: > Carl's right on all counts. Making the VM report UI input events is easy to do and... > it is on our list for the next release!! :-) :-) Is the next release going to include timer events? Will polling be eliminated? Or is this longer term?

1 0

← Newer
1
...
4
5
6
7
8
9
10
...
30
Older →

Jump to page: