Self-interest September 1996

self-interest@lists.selflanguage.org

7 participants
8 discussions

WinSelf ?
by Gilad 22 Sep '96

22 Sep '96

------------------------------------- Name: Gilad Haimov E-mail: reshume(a)netvision.net.il Date: 5/1/80 Time: 9:13:02 PM ------------------------------------- Is there( going to be ) a Window version of Self ? Gilad Haimov

1 0

Win Self ?
by Gilad 22 Sep '96

22 Sep '96

------------------------------------- Name: Gilad Haimov E-mail: reshume(a)netvision.net.il Date: 5/1/80 Time: 9:09:48 PM ------------------------------------- Hellow there. Is there an existing\planned Self version for Windows ? Gilad Haimov

1 0

Self and Smalltalk
by Jecel Assumpcao Jr 18 Sep '96

18 Sep '96

Here are two ideas inspired by Mario's excellent Smalltalk emulation: Wouldn't it be fun if Self+emulator could be X3J20 complient? That is, compatible with the future ANSI Smalltalk standard. That would get some attention. I haven't look into it (I don't have a copy of the draft) and so don't know how hard it would be to do this. I do know that the standard doesn't say anything about GUI programming. Having recently tried various Smalltalks (Smalltalk/X, Dolphin Smalltalk, Smalltalk MT) and looked at the documentation for others I have to say that I much prefer Self's morphic UI. Here is my idea: wouldn't it be neat if someone ported Kansas to Smalltalk (GNU Smalltalk, for example, which doesn't have any built-in GUI)? I know there are several fancy window builders for Smalltalk, but this would be really a great alternative. -- -----=============( Jecel Mattos de Assumpcao Jr )===========----- http://www.lsi.usp.br/~jecel/merlin.html | mailto:jecel@lsi.usp.br

1 0

Re: inline outliners
by randall.smith＠Eng.Sun.COM 10 Sep '96

10 Sep '96

Neat idea! But not so sure about your last case there... what if you have a tightly-packed run of very small morphs (like on a tool bar) and you're trying to look at a bunch of them at once? Even of most of these outliners were collapsed, it could be a pain and visually obscuring, confusing or misleading. We talked about having outliners pop-up adjacent to the morph, yet stuck to the edge of the morph so that the two moved together. But stuck in a "loose" way, so that some simple gesture could undo the connection. Perhaps simply grabbing the outliner would free it from its morph. Some animation or direction indicator might be used to show a disconnected outliner's desire to "return home." If this appeals, it should be easy to do using the same mechanism that is used for the sprouted "pointers" from outliner slots. If you sprout the little button (or pointer thing) at the end of a slot it runs out and sticks to appropriate target outliner, and will even follow it around if the target moves. But if you grab the pointer, you just end up holding the pointer. --Randy > From jecel(a)lsi.usp.br Mon Sep 9 09:43:46 1996 > Sender: root(a)ns2.internetco.net > Date: Mon, 09 Sep 1996 12:56:00 -0300 > From: Jecel Assumpcao Jr <jecel(a)lsi.usp.br> > Organization: University of Sao Paulo - Brazil > X-Mailer: Mozilla 3.0Gold (X11; I; Linux 1.2.13 i486) > MIME-Version: 1.0 > To: self-interest(a)self.sunlabs.com > Subject: inline outliners > Content-Transfer-Encoding: 7bit > > Here is an idea I had some time ago. I thought I had posted it > here already, but looking back though my old mail it seems > that I didn't. Sorry if I am wrong about this... > > It really messes up the "single object identity story" in the > Kansas user interface to have both morphs and outliners for > the same object present on the screen at the same time. > > I can feel a little better about this if I think about the outliner > as the morph for the mirror of the object - then we have two > morphs for two objects. I would prefer, however, to have the > outliner appear with the corresponding morph embedded in it > as its title. Instead of having the outliner say "aCircleMorph" > it would show the circle morph near its top. The outliner > would pop up in such a place that the circle morph would be > at the exact same spot it was before. > > What if the morph has other morphs embeded in it? No problem - > the whole morph-tree would be the "title" for the outliner > and the user would have to understand that the outliner was > for the "lowest" morph of the bunch. > > What if the morph is embeded in some other morph? This is > very tricky to solve, but I would "grow" the outliner under > the given morph but over its sibling and parent morphs. The > outliner would be pseudo-embeded in the morph's parent > visually, but shouldn't have any effect on layout. One > problem with this is that browsing more than a single > submorph would lead to a great visual clutter. If it is > easy to sprout and shrink outliners for these embeded > morphs, however, limiting oneself to looking into one > morph at a time won't be such a big bother. > > -- > -----=============( Jecel Mattos de Assumpcao Jr )===========----- > http://www.lsi.usp.br/~jecel/merlin.html | mailto:jecel@lsi.usp.br > >

2 1

inline outliners
by Jecel Assumpcao Jr 09 Sep '96

09 Sep '96

Here is an idea I had some time ago. I thought I had posted it here already, but looking back though my old mail it seems that I didn't. Sorry if I am wrong about this... It really messes up the "single object identity story" in the Kansas user interface to have both morphs and outliners for the same object present on the screen at the same time. I can feel a little better about this if I think about the outliner as the morph for the mirror of the object - then we have two morphs for two objects. I would prefer, however, to have the outliner appear with the corresponding morph embedded in it as its title. Instead of having the outliner say "aCircleMorph" it would show the circle morph near its top. The outliner would pop up in such a place that the circle morph would be at the exact same spot it was before. What if the morph has other morphs embeded in it? No problem - the whole morph-tree would be the "title" for the outliner and the user would have to understand that the outliner was for the "lowest" morph of the bunch. What if the morph is embeded in some other morph? This is very tricky to solve, but I would "grow" the outliner under the given morph but over its sibling and parent morphs. The outliner would be pseudo-embeded in the morph's parent visually, but shouldn't have any effect on layout. One problem with this is that browsing more than a single submorph would lead to a great visual clutter. If it is easy to sprout and shrink outliners for these embeded morphs, however, limiting oneself to looking into one morph at a time won't be such a big bother. -- -----=============( Jecel Mattos de Assumpcao Jr )===========----- http://www.lsi.usp.br/~jecel/merlin.html | mailto:jecel@lsi.usp.br

1 0

self source?
by Jordan K. Hubbard 06 Sep '96

06 Sep '96

I don't see a "self FAQ" anywhere on self.sunlabs.com, so I'll have the temerity to ask here whether or not a release of sources is ever planned. It would certainly help to make the language more popular if it were ported to architectures other than SPARC, no? Thanks.. Jordan

2 1

Re: syntax ambiguities.
by David.Ungar＠Eng.Sun.COM 02 Sep '96

02 Sep '96

Do you know that there is a Self 4.0? It is available at our WEB site (http://self.sunlabs.com). Maybe the manual is clearer. Anyway here are the specific answers (hope they help): (| mySlot = ( randomNumber ) |) This creates an object with one slot, "mySlot" containing a method that sends randomNumber to self. Informal rule: slot initializers enclosed in parens parse as methods. aVar: ( randomNumber ) sends random Number to self and then sends aVar to self with the result as its argument. Rule: in the body of a method, parens are just for grouping. Finally: foo.bar is always parsed as a resend. If you want a statement separator, you must leave a space (I think). (foo. bar) Rule: always leave a space (or newline, etc) after a period used as a statement separator. Good luck! At 3:16 PM 9/2/96, Pierre De Pascale wrote: >Hello selfer's, > >I have read the self 3.0 reference manual and stopped at the appendix >where the syntax of the self language was presented. In particular the >syntax presented is highly ambiguous, but a small foot-note says that >the ambiguities are resolved using standard precedence techniques. > >But I still got some problems: > >1. when you initialize a slot. How do you determine if it is a method >or a simple value ? for example: > > (| mySlot = ( randomNumber ) |) > >what is the meaning of the mySlot initialization: do I initialize >mySlot with a one-time call to randomNumber (i.e. mySlot is a value >slot) or mySlot returns a new randomNumber everytime (i.e mySlot is a >method) ? > >2. a same kind of ambiguities appears in expressions: what is the >meaning of this expression ? > > aVariable: ( randomNumber ) > >Do I initialize aVariable with one random number, in this case () are >for precedence reason, or with the object "(randomNumber)", in this >case a method object. > >3. how can you detect a resend ? for example: > > myResendSlot.resendMethod > >This could be interpreted as a directed resend or two expressions >consisting of two method sends, one myResendSlot and one >resendMethod. How do you make the difference ? is there a runtime >decision of the meaning ? > >It may be possible that those issues were already discussed in this >mailing list. Since I subscribed only lately I was lazy enough not to >check the past issues. Just tell me if I need to kick myself :^) > >thanks, > >- Pierre. -- Dave

1 0

syntax ambiguities.
by Pierre De Pascale 02 Sep '96

02 Sep '96

Hello selfer's, I have read the self 3.0 reference manual and stopped at the appendix where the syntax of the self language was presented. In particular the syntax presented is highly ambiguous, but a small foot-note says that the ambiguities are resolved using standard precedence techniques. But I still got some problems: 1. when you initialize a slot. How do you determine if it is a method or a simple value ? for example: (| mySlot = ( randomNumber ) |) what is the meaning of the mySlot initialization: do I initialize mySlot with a one-time call to randomNumber (i.e. mySlot is a value slot) or mySlot returns a new randomNumber everytime (i.e mySlot is a method) ? 2. a same kind of ambiguities appears in expressions: what is the meaning of this expression ? aVariable: ( randomNumber ) Do I initialize aVariable with one random number, in this case () are for precedence reason, or with the object "(randomNumber)", in this case a method object. 3. how can you detect a resend ? for example: myResendSlot.resendMethod This could be interpreted as a directed resend or two expressions consisting of two method sends, one myResendSlot and one resendMethod. How do you make the difference ? is there a runtime decision of the meaning ? It may be possible that those issues were already discussed in this mailing list. Since I subscribed only lately I was lazy enough not to check the past issues. Just tell me if I need to kick myself :^) thanks, - Pierre.

2 1

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 September 1996