Self-interest February 1996

self-interest@lists.selflanguage.org

7 participants
7 discussions

self-VM documentation
by we37394＠is1.vub.ac.be 01 Mar '96

01 Mar '96

Hy, I'm doing research for my thesis on various virtual machines. I'm looking for documentation on the seff-VM. I 've already downloaded some papers from the Self-homepages. Can somebody help me with other references to papers or technical documents. I'm particulary interested in the op-codes and a description of the virtual machine. Thanks. Dimitri Van de Putte, a new Self-fan. Dimitri Van de Putte. we37394(a)vub.ac.be, Free University of Brussels.

2 1

Re: Latin-1 Characters
by David.Ungar＠Eng.Sun.COM 22 Feb '96

22 Feb '96

Sounds great! At some point we should "buy back" your changes, if that is OK with you. At that point, we'll have to use the other module approach. Till then, the way you are doing it is the best! At 6:53 PM 1/23/96, Thierry Goubier wrote: >David Ungar writes: > > How did it turn out? > > > > At 10:55 AM 12/1/95, Thierry Goubier wrote: > > >That probably sound stupid, but why is there a limit on printables > > >characters that forbid the user to enter a caracter with an accent (or > > >any composed character) ? If you change this limit to 255, all > > >accentuated characters work correctly. Or am I missing something ? > > > > > >The slot is lastPrintableCharacter in traits string, and it's standard > > >value is 127. I'll try to check if I break something by changing the > > >value. > > > > > >Thierry. > >I've modified all the slots that seems to be related, like isVowel, >isLetter, isUpperCase, etc... Nothing seems to be broken yet. It only >allow for accents in comments and strings, not in identifiers (slots >names). > >By the way, I've created a new module for this. I've put in it all >modified methods, like isVowel, instead of their original module. Is >this the right way of using modules? > > > > -- Dave > > > > > >Thierry. >___________________Thierry.Goubier(a)enst-bretagne.fr__________________ > <Am I dreaming the dream, or is the dream dreaming me ?> > [Corwainer Smith] >http://www-info.enst-bretagne.fr/~goubier/ -- Dave

1 0

concurrency in self
by Albertina Lourenci 09 Feb '96

09 Feb '96

Dear SELF group: I am interested in knowing how you deal with concurrency in SELF? And where can I find a "book" about the language? It seems I have already arranged an "advisor" for me ...guess where? In Brazil| This is too much good to be true| I am a green architect, I am so proud of betting in the "purest OO paradigm" and have been courageously against bad use of OO programming. Until now I have been interested in BETA. Since Ole Lehrman Madsen is part of BETA and SELF group I do think he could compare both languages and tell the differences between the two. Cheers Albertina

1 0

Solaris For Intel
by CharlesMonteiro 09 Feb '96

09 Feb '96

Hi. I am a professional Smalltalk developer who is very interested in learning more about SELF. I however, do not have a Unix box to play with but have heard that Sun has come out with Solaris for the Intel platform and therefore I was hoping and wandering if some very gifted individual at Sun is thinking about porting SELF to this platform. Thanks. Charles.

1 0

Re: Multiscale Kansas?
by Jecel Assumpcao Jr 09 Feb '96

09 Feb '96

Maloney, John wrote: > Morphic (the graphical object framework underlying Kansas) was originally > designed to support arbitrary scaling, and the scaling code worked at one time, > although to my knowledge it has not recently been tested. I'll take a look. Did the mysterious 3D GUI use this code too or was it a completely different beast? > However, we later > started using pixmap images in the system (e.g., in the web browser) but > haven't yet written code to scale these images. If you limit yourself to power of two scales, then you can store every bitmap in all possible resolutions (like mip maps). I do it for anti-aliasing and its not hard (don't ask about how long it takes me to calculate these maps, though). > You'd want to do something > intelligent about "greeking" text below a certain size, and you'd probably want > to put in some sort of detail suppression to simplify the drawing of composite > morphs when zoomed way out. The Pad++ folks also discovered that you need to > turn off mouse sensitivity as the screen size of objects shrink or you find > that you're clicking on sub-pixel buttons without realizing it! Right - every drawing operation and event dispatching must take the different scales into account. It is almost as hard as doing 3D, though not quite (no "angled" objects and you get to stick to power of two). > I'd love to see this happen. I think it would require 4-6 weeks of solid work. That is exactly what I was asking - is it worth it? It might be neat to be able to zoom into the little "icon" in an object's slot and have it turn out to be another full object! Of course, the "portals" idea ends up allowing an object to show up in more than one place on the screen (which doesn't seem to be popular among Self fans), though it only really *exists* in one place at a time. > [lots of great sort comments] Thanks! This was great - it is probably exactly what my friend wanted to know. I'll just pass it along to him. Since I always just use bubble sort I never had to worry about equal elements getting out of order (just kidding!). -- Jecel

1 0

Re: Multiscale Kansas?
by J.MALONEY＠AppleLink.Apple.COM 08 Feb '96

08 Feb '96

Hi, Jecel, Morphic (the graphical object framework underlying Kansas) was originally designed to support arbitrary scaling, and the scaling code worked at one time, although to my knowledge it has not recently been tested. However, we later started using pixmap images in the system (e.g., in the web browser) but haven't yet written code to scale these images. You'd want to do something intelligent about "greeking" text below a certain size, and you'd probably want to put in some sort of detail suppression to simplify the drawing of composite morphs when zoomed way out. The Pad++ folks also discovered that you need to turn off mouse sensitivity as the screen size of objects shrink or you find that you're clicking on sub-pixel buttons without realizing it! I'd love to see this happen. I think it would require 4-6 weeks of solid work. Mergesort is used to sort indexable collections because it preserves the order of elements that are "equal" according to the search comparison function. (That is, it is a stable sort.) Thus, you could sort a list of files first by name, then by size, and the second sort would preserve the name ordering for for files of the same size. Thanks to Dave for pointing out the utility of stable sorts. Heapsort is used to sort collections that aren't initially sorted, like sets, since there is no ordering to preserve. It is the default sort algorithm for traits collection. I believe these are both O(n log n) sorts. Quicksort was tried but mergesort outperformed it for in some cases where the collection was already nearly sorted or where all elements were the same. This was true even for various pivot-selection algorithms. -- John

1 0

Multiscale Kansas?
by Jecel Assumpcao Jr 07 Feb '96

07 Feb '96

Its great that the list is back online - I have been saving a number of questions to ask here! I think it is really great that the Kansas user interface has been positioned as a remote education project! What I wanted to ask is if a multiscale Kansas would be a good idea (multiscale interfaces are also called "zooming interfaces" or "fractal interfaces" as in Ken Perlin's PAD - http://www.cs.unm.edu/pad++/begin.html). I looked at these types of interfaces a few years ago and was very impressed, but decided to continue working with 3D instead. But it does seem a good match for the kind of things that Kansas tries to do. Another question if from a friend of mine: why was Mergesort chosen for collections in Self? I think Heapsort is also used sometimes, right? Cheers, - Jecel

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

Self-interest February 1996