Squeak-dev September 2000

squeak-dev@lists.squeakfoundation.org

115 participants
411 discussions

Re: [enh] atlanta: a fixed-width font
by Bruce ONeel 30 Sep '00

30 Sep '00

On a mac celeste saved them properly to something which QuickTime's image reader could understand. cheers bruce Bob Arning <arning(a)charm.net> wrote: > On Fri, 29 Sep 2000 12:40:23 -0700 Dan Ingalls <Dan.Ingalls(a)disney.com> wrote: > >>There are also two GIF files attached which > >>show what the font looks like. > > > >If this is true, then it is an extraordinary encoding scheme ;-). > >Would you please confirm? > > Dan, > > They are there, but their content type is application/octet-stream where image/gif might have clued email clients into their potential use. > > Cheers, > Bob

1 0

Re: [updates] 29 more for 2.9alpha
by Duane T Williams 30 Sep '00

30 Sep '00

When I did the update, I got a message that 21 files were loaded, but that 8 files, beginning with 2471 could not be loaded. I'm using Squeak VM 2.8gamma PPC. Duane

4 3

Re: AttachmentMorph design questions
by Ned Konz 30 Sep '00

30 Sep '00

danielv(a)netvision.net.il wrote: > > The motion of the attached Morph would be done via a callback to the morph, > > so it could do something appropriate. The default would be to move the entire > > Morph. > Other possible behaviors - > Nothing (like you connection morphs) > Resizing (like when you pull a resize handle) > Shared resizing (like panes in a system window thats being resized) Yes, I can see this strategy being different for each end of the attachment. Attaching a line to a rectangle, for instance, we probably don't want the rectangle to move when you try to move the line, but we do want the line endpoint to move when the rectangle is moved. > I think i would want to first do all the connections, and then be able > to play with how each morph reacts to each connection. So it should be > easy to see (in edit mode?) what each connection behaves like, and > modify it without reattaching manually. Implementation idea - using a > state/strategy pattern (the attachment morph delegates decisions to a > policy object it holds, which can be replaced at runtime) for the > attachment morph might help doing this. I like this idea. This is yet another case, though, where there are lots of options, which will probably appear on a menu (out of laziness). -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

1 0

Re: AttachmentMorph design questions
by Ned Konz 30 Sep '00

30 Sep '00

Joshua Channing Gargus wrote: > If you have another morph connected to the rectangle and move > it in a big circle around the rectangle, you might want the > connection point to "slide" around the perimeter of the > rectangle (didn't your earlier changeset do this? I don't > have it handy). Yes, I like that behavior sometimes. My ConnectionMorphs do this in a very simple way by picking the nearest point (of 8) around the bounds of the attached Morph. I was thinking that one strategy could be for the AttachmentMorph to have a reference point (perhaps the center of the other attached Morph by default), and inquire of the attached Morph which connection point to use. To get behavior like my existing ConnectionMorphs, the reference point would be the prior line vertex. > A shortcut for opaque morphs would be to > attach to each morph's center point, but this would look bad > for translucent/irregularly-shaped morphs. I can't think of a > general way to do this without requiring too much computation, > though. I couldn't, either (a general way would be to draw the Morph again on a private Canvas and search for its edge along a line to the reference point...). -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

1 0

Re: ConnectorMorph, Mind-Mapping and Concept-Mapping
by Dan Ingalls 30 Sep '00

30 Sep '00

While Morphic may not be designed specifically for embedded semantic information, it's really easy to experiment with this using the setProperty: #propertyName toValue: value valueOfProperty: #propName protocol (Morph>access properties). I would think this would be adequate for playing around. If you get serious, then you'll probably want to develop a whole semantic network formalism, and just use morphic as a viewer for that. - Dan ------------------- Joshua Channing Gargus <schwa(a)cc.gatech.edu> wrote... >I agree that Squeak is the best current platform to implement these >ideas. However, I don't think that it is sufficient as-is. The >main reason is that there is no semantic information associated >with a given morph. Concept maps that you make with simple morphs >are difficult to traverse programmatically. I suppose, for example, >that you could say that the amount of "redness" in a connector >represents the strength of the connection, and traverse the morphs >comprising the concept map based on this type of information. >However, this is much more difficult and unclear than embedding >semantic information in the components of the concept map from the >start. > >On Thu, Sep 21, 2000 at 08:31:59PM +0200, Hannes Hirzel wrote: >> At the beginning of this month Ned Konz sent a change set with a >> connector morph to this list. Together with the recent additions by Dan >> Ingalls (selecting multiple morphs and undo) Squeak is >> now quite a nice mind-mapping tool. Over the more graphically >> polished versions of this kind of software it has the tremendous >> advantage that the map can be traversed programmatically be writing >> scripts. >> > >Are you talking about EToy scripts, or regular Smalltalk scripts? >I think that it would be great to integrate a concept mapping tool >with the EToy scripting system. > >Joshua > >> Hoping that this feedback might be useful for someone >> >> Hannes Hirzel >> > >-- > > -Joshua Gargus (schwa(a)cc.gatech.edu)

2 1

[BUG]TranslucencySliderRe: [updates] 29 more for 2.9alpha
by Karl Ramberg 30 Sep '00

30 Sep '00

Dan Ingalls wrote: > > 2730TranslucencySlider-di -- Dan Ingalls -- 28 September 2000 > Adds a slider effect to the translucency control of the color picker. Fixes a couple of glitches in moving from modal to modeless operation. When alpha is set to transparent( alpha: 0), on reentry to the color picker the color is set to opaque ( alpha:1). Karl

2 1

Re: CORBA for Squeak?
by Johan Fabry 30 Sep '00

30 Sep '00

jchludzinski(a)worldkey.net wrote: > Is there CORBA support for Squeak? Is anyone working on such a > beastie? AFAIK there is no CORBA support for squeak yet. If you are looking for a RPC implementation, take a look at Remote Message Passing, by yours truly. You can find it at http://prog.vub.ac.be/PoolMembers/johanfabry/RMP.zip The readme is also separately available: http://prog.vub.ac.be/PoolMembers/johanfabry/RMP-Readme.txt Here's a quick snippet from the readme: ---snip--- Remote Message Passing (RMP) is a lightweight Remote Procedure Calling (RPC) implementation for Squeak. (2.7 or 2.8) The basic requirement for RMP was SIMPLICITY. By keeping things simple RMP is easy to learn and to use, but it still has the needed functionality for a usable RPC implementation. Quote from Göran Hultgren: "The code looks well written (no sweat grokking it) and it is SIMPLE." Thanks, Göran! ;-) RMP is inspired by CORBA and Java RMI (no need to *completely* reinvent the wheel). However, RMP is a lot simpler to use (no defining of interfaces, generation of stubs & skeletons). ---snip--- -- "You are more than the sum Johan Fabry - Johan.Fabry(a)vub.ac.be of what you consume. Vrije Universiteit Brussel Desire is not an occupation." Programming Technology Lab, Room 10F709 -- KMFDM Pleinlaan 2, 1050 Brussels, Belgium

1 0

Re: [enh] atlanta: a fixed-width font
by Bob Arning 29 Sep '00

29 Sep '00

On Fri, 29 Sep 2000 12:40:23 -0700 Dan Ingalls <Dan.Ingalls(a)disney.com> wrote: >>There are also two GIF files attached which >>show what the font looks like. > >If this is true, then it is an extraordinary encoding scheme ;-). >Would you please confirm? Dan, They are there, but their content type is application/octet-stream where image/gif might have clued email clients into their potential use. Cheers, Bob

2 1

Re: [enh] atlanta: a fixed-width font
by Dan Ingalls 29 Sep '00

29 Sep '00

>There are also two GIF files attached which >show what the font looks like. If this is true, then it is an extraordinary encoding scheme ;-). Would you please confirm? - Dan

3 2

openInWorld vs activate
by David Farber 29 Sep '00

29 Sep '00

Here's a question for all you morphic gurus out there: Why does aMorph openInWorld behave differently from World addMorph: aMorph. aMorph activate. (or, more to the point, why does the former not behave like the latter.) To illustrate further, compare the behavior of these two snippets when execute at the end of the following: | alignMorph win | alignMorph := AlignmentMorph new. alignMorph centering: #center; hResizing: #shrinkWrap; vResizing: #shrinkWrap; orientation: #vertical. alignMorph addMorphBack: (BasicButton new label: 'a button'). alignMorph addMorphBack: (BasicButton new label: 'this is a button'). alignMorph addMorphBack: (BasicButton new label: '''nother button'). alignMorph fullBounds. win := SystemWindow labelled: 'fooMail console'. win bounds: ((alignMorph position - ((0@win labelHeight) + win borderWidth)) corner: alignMorph bottomRight + win borderWidth + (2@2)). win addMorph: alignMorph frame: (0@0 extent: 1@1). "World addMorph: win. win activate" "verses" "win openInWorld" When I use win activate, i get my buttons-inside-an-alignment-morph all wrapped in a system window. When I use openInWorld, the system window reverts to some default size. What's the scoop? thanks, david -- David Farber dfarber(a)numenor.com

1 0

Jump to page: