Squeak-dev August 1998

squeak-dev@lists.squeakfoundation.org

99 participants
188 discussions

Re: Polymorphism without protocol dilution
by m3rabb＠stono.com 31 Aug '98

31 Aug '98

>I'd like to make a third kind of suggestion for your list. Don't mix blocks >and other values. Make the protocol different. For example: > > at: key ifAbsentPut: anObject > self at: key ifAbsent: [ self at: key put: anObject ] > > > at: key ifAbsentPutValue: aBlock > self at: key ifAbsent: [ self at: key put: aBlock value ] > > >Another effect of this suggestion is that it still allows putting blocks >into dictionaries as values, while suggestions 1 and 2 do not. Good point. I was so focused on other aspects of protocol dilution, didn't consider the other form or dilution I was introducing. at: key ifAbsentPutValueOf: aBlock self at: key ifAbsent: [ self at: key put: aBlock value ] Perhaps this would be an even better name. Though this wasn't the best example to illustrate it, sometime you still want an variable to hold either anObject or aBlock that it can evaluate in the same way. If you want to prevent dilution, you shouldn't use #value. However in this case IMHO I think that the name should be #asEvaluated. BlockContext>>#asEvaluated ^self value Object>>#asEvaluated ^self (I still wish the convention for Blocks wasn't #value, and was instead something like #evaluate.) --Maurice --------------------------------------------------------------------------- Maurice Rabb 773.281.6003 Stono Technologies, LLC Chicago, USA

2 1

Re: [Squeak] been dazed by SortedCollection
by m3rabb＠stono.com 31 Aug '98

31 Aug '98

At 10:28 AM 8/31/98, Patrick Logan wrote: > The problem with the new versions is that we are forgeting special > the SortCollection protocol. > > #addLast: adds an element without resorting (sneaky) > >Why is #addLast allowed to be sneaky but #at:put: is disallowed >altogether? > SortedCollections are never guaranteed to have any element at a particular location, just in order. If you insert an element at a location, the ideal SortCollection will immediately resort itself to stay correct. The element will more than likely not still be at the location you put it in, making #at:put: meaningless. By this reasoning _all_ the inherited adding methods are also meaningless except #add:. To this end I have added: SortCollection>>#addWithoutSorting: anObject ^super add: anObject and changed #copyFrom:to: to use #addWithoutSorting: instead of inappropriately using #addLast:. --Maurice --------------------------------------------------------------------------- Maurice Rabb 773.281.6003 Stono Technologies, LLC Chicago, USA

1 0

Re: [Squeak] been dazed by SortedCollection
by m3rabb＠stono.com 31 Aug '98

31 Aug '98

At 3:55 AM 8/31/98, NISHIHARA Satoshi wrote: >and get result... >1) 29672 >2) 29595 >3) 14330 (... super copyFrom: startIndex to: endIndex. ...) The problem with the new versions is that we are forgeting special the SortCollection protocol. #add: adds an element an resorts the list immediately #addAll: adds a collection and then resorts #addLast: adds an element without resorting (sneaky) #resort forces a resort Try using: copyFrom: startIndex to: endIndex "Answer a copy of the receiver that contains elements from position startIndex to endIndex, keeping the sortBlock from the receiver" | newCollection | newCollection _ self species new: (endIndex < startIndex ifTrue: [0] ifFalse: [(endIndex + 1 - startIndex)]). newCollection sortBlock: self sortBlock. startIndex to: endIndex do: [:index | newCollection addLast: (self at: index)]. ^newCollection reSort > | a b | > Transcript cr; show: (Time millisecondsToRun: > [a := SortedCollection new. > a sortBlock: [:element1 :element2 | element1 > element2]. > 1 to: 10000 do: [:i | a add: i]. > 100 timesRepeat: [b := a copyFrom: 1 to: 500]. > ]) printString. > ^b Likewise you should rewrite your test code to make use of #addAll: as well. a := SortedCollection sortBlock: [:element1 :element2 | element1 > element2]. a addAll: (1 to: 10000). Time millisecondsToRun:[100 timesRepeat: [b := a copyFrom: 1 to: 500]] --Maurice --------------------------------------------------------------------------- Maurice Rabb 773.281.6003 Stono Technologies, LLC Chicago, USA

2 1

#at:ifAbsentPutValue[Of]:
by James Foster 31 Aug '98

31 Aug '98

Would it be too wordy to add a preposition? Is it more readable? Is the increased readability worth the extra typing? self at: aKey ifAbsentPutValue: aBlock self at: aKey ifAbsentPutValueOf: aBlock James Foster Fargo, ND -----Original Message----- From: David N. Smith <dnsmith(a)watson.ibm.com> Date: Friday, August 28, 1998 12:10 PM Subject: Re: Polymorphism without protocol dilution ><SNIP> >Don't mix blocks >and other values. Make the protocol different. For example: > > at: key ifAbsentPut: anObject > self at: key ifAbsent: [ self at: key put: anObject ] > > > at: key ifAbsentPutValue: aBlock > self at: key ifAbsent: [ self at: key put: aBlock value ] > > >Another effect of this suggestion is that it still allows putting blocks >into dictionaries as values, while suggestions 1 and 2 do not. > > >Dave >_______________________________ >David N. Smith >IBM T J Watson Research Center >Hawthorne, NY >_______________________________ >Any opinions or recommendations >herein are those of the author >and not of his employer. >

1 0

[Squeak] been dazed by Resizing subview (MVC)
by nishis＠urban.ne.jp 30 Aug '98

30 Aug '98

Sorry for MVC Question. Please evaluate below and resize subview. | topView subView1 subView2 | topView := (StandardSystemView new) model: nil; label: 'resize test'; minimumSize: 64 @ 128. subView1 := View new. subView1 backgroundColor: Color red. topView addSubView: subView1. subView2 := View new. subView2 backgroundColor: Color green. topView addSubView: subView2 toRightOf: subView1. subView1 borderWidth: 1. subView2 borderWidth: 1. topView borderWidth: 1. topView controller open Picking up center line and drag left and right many times, subViewes are shrinking.... or evaluate below | topView subView1 | topView := (StandardSystemView new) model: nil; label: 'resize test'; minimumSize: 64 @ 128. subView1 := View new. subView1 backgroundColor: Color red. topView addSubView: subView1. subView1 borderWidth: 1. topView borderWidth: 1. topView controller open and resizing topview many times, subview is shrinking... It didn't happen on Version 1.31. How can I avoid it (with using MVC) ? regards ------------------------------------------- 西原聡士 (NISHIHARA Satoshi) e-mail: nishis(a)urban.ne.jp : tcc00164(a)nifty.ne.jp URL: http://www.urban.ne.jp/home/nishis/ -------------------------------------------

1 0

language: toward Obliq
by dufrp＠oricom.ca 30 Aug '98

30 Aug '98

Although I do like to read this mailing list I do not use a lot Squeak mainly because I have a small&slow computer which is using mostly DOS as an operating system. Which means I have to borrow my sister big computer to use Squeak (a 486DX4-100, HD 512Mb). I'd like to find a modern combo language+OS that doesn't take much to be run and for now it seems it's going to be Oberon, but I use to change of language or programming language almost every week. :-) First I'd like to tell that the BNF grammar of Squeak that was posted here recently did looked pretty usefull for someone wishing to learn Squeak, without any needs to add more semantical information. Such a BNF grammar is a great reference when learning a new language. I like the ability to explore Squeak with the browsers. I did also like to see the multiple message dispatching of Smalltalk. But like I said before I don't use it much, and the only program I wrote in Squeak was a small 'Guess my number' game that was writing on Transcript and reading numbers from FillInTheBlanks. I was trying to make all this in one window, when I was beginning to explore the MVC concept and I did get lost there and was not trying much harder at Squeak ever since. Although most others GUI must be using controllers I never have seen them before using Squeak. I suppose others language and GUI are simply hiding them. Since it seems that Squeak is, with the 'language and polymorphism' thread, thinking about how to make itself better, this was making me to think about a message I read here about the future of Squeak that could become be more expressive by using new ideas proposed by researcher Luca Cardelli. Although that Cardelli is using a mathematical formalism that I don't really understand, I do have my way of understanding his ideas and I would summarize them in two ideas. First an objects is a bunch of fields, and this fields can all be changed. A field can be any value, including, and that what is new, a method value. So an object can modify his methods or add new ones. Second idea: an object can be cloned. So rather than using classes the way we do with Smalltalk, we can create a function that clone a base object and return it. This seems to give much more expressive power, and with it the possibility to make code that is much more harder to read and understand. Anyway my feeling is that it is worth it. Cardelli have made a language (in 1995?) that is using these concepts and add the idea that you can also make an assignment not only to local fields (normal values or methods values) but also to fields of objects on remote computers, making this way a distributed language. This language is called Obliq. It is distributed with an other language in which it was implemented: Modula-3. Modula-3 take about 1Gb of hard disk space and that much much more than I have. So I never saw Modula-3 neither than I saw Obliq. I'd like to ask if there is someone working at making Squeak using these ideas. This could well be the case because I read about Cardelli here some months ago. Cardelli have also wrote a not so funny paper about user interface. It seems also that Obliq is using a powerful user interface looking like Visual Basic. Maybe Squeak could be one day using such an interface that would maybe be simpler than current MVC model? Since that the original thread was about what's polymmorphism I'd like to say that Cardelli have also wrote one or two papers describing what it is. >From memory, I would say that the web page of Cardelli is: www.luca.demon.co.uk (but my browser seems to make a bad request.) Maybe the easy way is to follow the links from Yahoo on Obliq. I am not saying Squeak should follow this path. I guess Obliq have a mystical look to me because I have not really seen it yet. Somehow I guess I am trying to have others use it so that they can tell me if it is as good as I amagine it. :-)

2 1

Re: Polymorphism and language
by Adam Bridge 30 Aug '98

30 Aug '98

Alan, Teach me to be an English major back in the 60s instead of delving into the computer realm. It took me until the late 70s to get THERE. APL was always a mystery to me: it's coding into symbols, while conceptually rich, was too high a price of admission for me, perhaps because the same learning disability that makes foreign languages so difficult for me was invoked in the same way. Thank you for your remarks and insights. I enjoy Squeak and this news group because it always seems to operate where my mind ought to be. Which makes it both rare and beyond price. Adam On 8/29/98 at 4:37 PM, alank(a)wdi.disney.com (Alan Kay) wrote: > This was a dream of the sixties, and the very first working version of > Smalltalk (72) manifested this dream. It was set up so that each object > could easily implement its own grammar for receiving messages. This allowed > many "small languages" to be developed and mixed. The expressivity was very > high. Unfortunately, we forgot that languages are also for communication -- > for being read, not just for being written -- and what we quickly wound up > with was a Tower of Babel -- as least as far as people who were trying to > learn from our code were concerned. For Smalltalk-76 we decided that the > language should be always be readable, and that the extensibility that was > really needed was in expressing meanings.

3 2

Re: [Squeak] been dazed by SortedCollection
by Bob Arning 30 Aug '98

30 Aug '98

On Sun, 30 Aug 1998 13:28:27 +0900 Satoshi NISHIHARA <nishis(a)urban.ne.jp> wrote: >I've been dazed and confused ((C)Led Zeppelin) by SortedCollection. > > | collection | > collection := SortedCollection new. > collection sortBlock: [:element1 :element2 | element1 > element2]. > 1 to: 10 do: [:i | collection add: i]. > Transcript cr; show: (collection copyFrom: 1 to: 5) printString > >Got SortedCollection (6 7 8 9 10 ), although expected SortedCollection >(10 9 8 7 6 ). >I think that SortedCollection (6 7 8 9 10 ) has sortBlock: [:x :y | x <= >y] (default) *not* [:x :y | x > y] (required). Is it right? You are right about the new SortedCollection having the default sortBlock. OrderedCollection>>copyFrom:to: creates the new collection with self species new: endIndex + 1 - startIndex This just gives you the same kind (with 2 exceptions) of collection you had before. It does NOT recognize what sortBlock the original had. To get your code to work, you can: 1. Explicitly give the new collection a sortBlock as in Transcript cr; show: ((collection copyFrom: 1 to: 5) sortBlock: [:element1 :element2 | element1 > element2]) printString 2. Or, if you are not interested in the result actually being a SortedCollection, Transcript cr; show: (collection asArray copyFrom: 1 to: 5) printString 3. Or you could reimplement #copyFrom:to: in SortedCollection to give the new collection the same sortBlock as the old one had: copyFrom: startIndex to: endIndex "Answer a copy of the receiver that contains elements from position startIndex to endIndex, keeping the sortBlock from the receiver" | targetCollection | (targetCollection _ self species new: (endIndex + 1 - startIndex max: 0)) sortBlock: self sortBlock. endIndex < startIndex ifTrue: [^targetCollection]. startIndex to: endIndex do: [:index | targetCollection add: (self at: index)]. ^ targetCollection Cheers, Bob

1 0

[Squeak] been dazed by SortedCollection
by Satoshi NISHIHARA 30 Aug '98

30 Aug '98

Hello Squeakers! I've been dazed and confused ((C)Led Zeppelin) by SortedCollection. | collection | collection := SortedCollection new. collection sortBlock: [:element1 :element2 | element1 > element2]. 1 to: 10 do: [:i | collection add: i]. Transcript cr; show: (collection copyFrom: 1 to: 5) printString Got SortedCollection (6 7 8 9 10 ), although expected SortedCollection (10 9 8 7 6 ). I think that SortedCollection (6 7 8 9 10 ) has sortBlock: [:x :y | x <= y] (default) *not* [:x :y | x > y] (required). Is it right? Here is my first aid. 'From Squeak 2.0 of May 22, 1998 on 29 August 1998 at 3:43:36 am'! !SortedCollection methodsFor: 'copying' stamp: 'nishis 8/29/1998 03:43'! copyFrom: startIndex to: endIndex "by nishis, 1998/08/29 03:42" | targetCollection | targetCollection := super copyFrom: startIndex to: endIndex. targetCollection sortBlock: sortBlock. ^targetCollection! ! regards -- -------------------------------------------- 西原聡士 (NISHIHARA Satoshi) e-mail: mailto:nishis@urban.ne.jp mailto:tcc00164@nifty.ne.jp URL: http://www.urban.ne.jp/home/nishis/ --------------------------------------------

2 1

Re: Polymorphism without protocol dilution
by Bob Shafer 29 Aug '98

29 Aug '98

I'd like to suggest that folks interested in this have a look at the languages J and K, both of which are discussed on comp.lang.apl. J is designed by Ken Iverson (with help from Ken Hui and others). J is, it can therefore be argued, the latest version of APL. It is distinguished from APL by having an ascii-based representation, and a firmer and more extensive theoretical foundation, including a form of functional programming (called "tacit form") that seems very similar in spirit to Backus' FP. It also has recently been extended with OO capabilities. The system (including free trial versions) and more info are available at http://www.jsoftware.com. The K language is also an APL-like language, similar in many respects to J, but it has grown within a business environment and has some very unique qualities. There is, as of very recently, a downloadable trial version of it, as well, at http://www.kx.com. It seems less mathematically oriented than J, but much more oriented towards business systems, and interprocess communication and rapid prototyping. There are ideas in the K system that totally blew me away (not so much the APL-like aspects, but the (apparently) total integration of language objects with user interface objects -- extremely neat stuff!). Both systems, if you download them, come with extensive documentation. The K documents seem to be especially well written and accessible. There's a treasure trove of good, well thought out ideas in these systems, IMHO. On Fri, 28 Aug 1998, Alan Kay wrote: > Maurice and Travis -- > > Dan Ingalls and I (seperately) were very taken with parts of APL in our > youth, and. over the years, have been quite excited at the possibilities > that polymorphism brings to this kind of generalization of collections. It > is now time to take another pass at the algebraic roots of > Smalltalk/Squeak. Suggestions please -- the best kind are not so much > add-ons, but comprehensive schemes that are both powerful, readable, and > help users think ...

2 2

Jump to page: