Squeak-dev November 1998

squeak-dev@lists.squeakfoundation.org

128 participants
265 discussions

Re: Hyperbolic Tree Class Browser
by Russell Allen 23 Nov '98

23 Nov '98

At 04:22 PM 21/11/98 -0600, Jerry Bell wrote: >Has anyone looked at hyperbolic trees? I find them hard to explain, but >very easy to grasp once you play with them. Try: > >http://www.inxight.com/Content/7.html > >The Library of Congress demo is especially nice. > >I think a hyperbolic tree would make a great UI for a class browser, maybe >as a replacement for the left-most pane. It would definately help a new >user like myself get a feel for which classes belong where in the hierarchy. > >Unfortunately I don't know if we could put one in Squeak without permission >from Inxight. I know they hold license to their implementation of >hyperbolic trees, but I don't know if they somehow own the idea also. I've >never understood that sort of thing very well. > There are two types of intellectual property which apply to computer programs. The first is copyright law, and the second is patent law. Patents apply to algorythms - ie the .gif format isn't free because of a patent held on the compression algorythm. Similarily, certain crytography algorythms are patented under US law (but not in other countries). Where patents are involved, you need legal advice before writing a program which does the same thing as the patented one. It may not be possible. Copyright, however, unlike patents, is a right to prevent copying (lawyers get paid a lot of money for this :-). Copyright attaches to the actual code which performs the task, not to the task itself. This is normally expressed as saying that copyright protects the expression and not the idea. However, determining what the idea is and what the expression of that idea is can be difficult. People who recreate versions of software which is "compatible" with other software, or which has the same functionality rely a lot on "clean room" techniques - so that they can sucessfully argue that they were only copying the idea and not the expression. This boils down to a few rules, for instance: 1. Never even look at the object or source code of the product you are imitating, let alone copy any of it! 2. Create a user interface which is as unlike the original as possible, or where the similarities are due to generic features of all modern software (ie a File|Open menu). 3. Use different names for all of the features. (This is to avoid allegations that you are attempting to confuse people into thinking that your version is the original.) The line is tricky, however, to draw. There is a line of US cases which gave protection to the "look and feel" of computer programs; remember Apple suing Microsoft, Lotus suing Borland etc? Not all of the cases were sucessful, but some such as Lotus v. Paperback Software were. Indeed Lotus was successful in suing Borland for copying its menu tree, but luckily this was reversed on appeal. The golden rule is, then, copy as little as possible, make your recreation as different as possible, so that you only copy the idea (and change the idea to something better if you can!). PS - this isn't legal advice, or a comment on the legal situation with Inxite/Hyperbolic Trees. It's just a description of what I believe to be the status of the law with regard to software. Have fun! Russell ------------------------------------------------------------ Russell Allen russell.allen(a)firebirdmedia.com Random Quote: "They must often change who would be constant in happiness or wisdom." Confucius, Analects. ------------------------------------------------------------

1 0

RE: Hyperbolic Tree Class Browser
by Alejandro F. Reimondo 23 Nov '98

23 Nov '98

I've tried to enter in the Louvre demo, it takes 10 minutes to load the Java applet and my IExprorer crashed closing all the opened browsers windows. Are there any demonstration not written in Java ? thanks. Ale. ---------- De: Jerry Bell[SMTP:jdbell@fareselaw.com] Enviado: Sábado 21 de Noviembre de 1998 19:22 Para: squeak(a)cs.uiuc.edu Asunto: Hyperbolic Tree Class Browser Has anyone looked at hyperbolic trees? I find them hard to explain, but very easy to grasp once you play with them. Try: http://www.inxight.com/Content/7.html The Library of Congress demo is especially nice. I think a hyperbolic tree would make a great UI for a class browser, maybe as a replacement for the left-most pane. It would definately help a new user like myself get a feel for which classes belong where in the hierarchy. Unfortunately I don't know if we could put one in Squeak without permission from Inxight. I know they hold license to their implementation of hyperbolic trees, but I don't know if they somehow own the idea also. I've never understood that sort of thing very well. They are a Xerox company, by the way. It's another neat idea from PARC. Jerry Bell jdbell(a)fareselaw.com -----Original Message----- From: Stefan Matthias Aust [mailto:sma@kiel.netsurf.de] Sent: Saturday, November 21, 1998 12:42 PM To: squeak(a)cs.uiuc.edu Subject: Re: C code for Block context >How does the Squeak Smalltalk to C translation handle Smalltalk code that >contains Blcok context? It was my impression that it is difficult to >program Block context in C. Actually, it cannot translate real blocks. The translator only translates a subset of Smalltalk, just enough to compile the interpreter. It does the usual optimization for ifTrue: and whileTrue: blocks, but can't handle full block closures. bye -- Stefan Matthias Aust // Are you ready to discover the twilight zone?

4 3

Re: Dynamic Primitives: Preliminary code available
by Hans-Martin Mosner 22 Nov '98

22 Nov '98

Hans-Martin Mosner wrote: > > Hello, > at http://www.heeg.de/~hmm/squeak/Translator-981121/ I've put the > current native code generator source. .... and forgot to include the README file. It's now there, and I've appended it for quick reading, too. Hans-Martin Content-Type: multipart/appledouble; boundary="----------ad2CDD5EE5298E"; x-mac-type="54455854"; x-mac-creator="522A6368"; name="README" Content-Transfer-Encoding: 7bit Content-Description: BBEdit Lite 4.0 Document Content-Disposition: inline; filename="README" Attachment converted: Anon:README (TEXT/R*ch) (000037A3)

1 0

Re: Hyperbolic Tree Class Browser
by Jim Benson 22 Nov '98

22 Nov '98

Jerry, >Has anyone looked at hyperbolic trees? I find them hard to explain, but >very easy to grasp once you play with them. I looked over that pretty thoroughly. There is an article explaining the underlying principle in SIGCHI '95 called 'A Focus+Context Technique Based on Hyperbolic Geometry for Visualizing Large Hierarchies' by Lamping,Rao and Pirolli pp. 401. There are some interesting applications, one example is a web site navigation tool. Basically the hyperbolic tree is built from all of the navigable nodes at a web site, and then the user can navigate the web site based on this tree, rather than following hyperlinks embedded in each page. This is an innovative solution to the problem of traversing large hierarchies using limited screen space. Inxight basically wants US $10,000 for a development version if you want to develop with it for a product. Oh, and they have nice 4 color brochures too !!! Personally, I think it looks pretty cool, but I have not figured out how useful it is. The applications that I have seen so far provide a different way to look at a namespace, but I do not think it provides any extra useful information. For example, in the Louvre demo you could see how all of the nodes of the tree were related, but those relations did not give you any more insight into what each node of the tree contains. The node contents tend to be what I am interested in, not the structure of the tree, and it is not quite clear how I search thru one of these trees. I think perspective walls and document lenses are more useful for looking thru large amounts of data in this context. jb

1 0

C code for Block context
by Richard L. Peskin 21 Nov '98

21 Nov '98

How does the Squeak Smalltalk to C translation handle Smalltalk code that contains Blcok context? It was my impression that it is difficult to program Block context in C. --dick peskin ================================= R. L. Peskin, Rutgers Univ. ; <peskin(a)caip.rutgers.edu>; <http://www.caip.rutgers.edu/~peskin> VT Phone (802) 824-4558 NJ Phone (732) 445-4208 "The corporate culture is concerned less with Occam's razor than his aftershave lotion."

2 1

Re: How To Re-evaluate A Block?
by Craig Latta 21 Nov '98

21 Nov '98

> What is the Squeak code I need to do before I send value to the protected > block so that it can be re-evaluated? That's BlockContext>>fixTemps, I believe. -C -- Craig Latta composer and computer scientist craig.latta(a)netjam.org www.netjam.org latta(a)interval.com Smalltalkers do: [:it | All with: Class, (And love: it)]

2 2

Re: How To Re-evaluate A Block?
by Andreas Raab 20 Nov '98

20 Nov '98

Hi, > What is the Squeak code I need to do before I send value to the protected > block so that it can be re-evaluated? It's not so easy since you're in a handler response. What you need is to make sure that the retry response cleans up the caller stack up to the #value message sent to the block so that the block is not recursively executed. The way I did this in my little exception exercise was to use something like: ExceptionHandler>>executeBlock: doBlock marker: marker retryBlock := [^marker]. ^doBlock value. retryBlock is the block that's been executed on Exception>>retry and the #on:do: method was then actually done by the handler as some sort of [result := self executeBlock: doBlock marker: marker. result == marker] whileTrue. The non-local return cleans up the stack here so you don't send #value recursively to the block. It's a little tricky ;-) Andreas -- Linear algebra is your friend - Trigonometry is your enemy. +===== Andreas Raab ============= (raab(a)isg.cs.uni-magdeburg.de) =====+ I Department of Simulation and Graphics Phone: +49 391 671 8065 I I University of Magdeburg, Germany Fax: +49 391 671 1164 I +=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+

1 0

Squeak Performance
by Jay O'Connor 20 Nov '98

20 Nov '98

All, Hi. I'm using VisualWorks NC on Linux and have Squeak as well. I was going to be using VWNC for a semi-commercial usage but I may not be able to afford it commercially. I'm using it for a Web Server (using a Web Server I built in VW similar in concept to the Squeak PWS and an HTML framework I wrote). I'm curious if Squeak has general performance good enough to investigate using it as a web server. Thanks and take care, -- Jay O'Connor joconnor(a)roadrunner.com http://www.roadrunner.com/~joconnor "God himself plays the bass strings first when He tunes the soul"

3 3

/. mentions squeak
by Ryan Davis 20 Nov '98

20 Nov '98

http://www.slashdot.org/ just had a mention of us (about a mention of us). Check it out: Jim Thompson writes "Today's edition of Project Heresy at CNET News covers a very interesting new (?) project called Squeak by, of all people, The Disney Co. This project, targetted at embedded devices, has a Smalltalk heritage and is based, to some extent, on code from Apple. Apple required that Disney not commercialize their product, so Disney has released it as Open Source. Squeak is not itself an OS, but the Heretics speculate that Squeak will soon be married to a scaled-down linux kernel an ported to all sorts of handheld devices. Squeak sounds reasonably cool in that it's based on Smalltalk-80 and has what sounds like a built-in "skins" feature for its UI. Do any Slashdotters have more info on Squeak, on where to go for more information? " Ryan Davis -=- Zen Spider Software -=- mailto:zss@ZenSpider.com -=- http://www.ZenSpider.com/ -=- I know that you believe you understand what you think I said but, I'm not sure you realize that what you heard is not what I meant.

4 3

Re: Squeak running on Japanese PDA
by ohshima＠is.titech.ac.jp 20 Nov '98

20 Nov '98

Bruce O'Neel wrote: > It now has a place of honor on my wall. BTW, I noticed the 26 > benchFib line in there. If you don't mind, what does > [26 benchFib] timeToRun > return? If the inlining is on (big and (supposed to be) fast interpreter), the result is around 7000 on faster versions of Zaurus. If the inlining is off, the result will be three times bigger than that. (I can't tell the actual result now.) Note that if the executable is bigger than 128k, ZaurusOS starts paging for binary segment, and the non-inlined version of VM is just smaller than 128k. So, the difference of the results of micro benchmark doesn't reflect the performance of 'real' applications. They runs much faster on non-inlined VM. OHSHIMA Yoshiki Dept. of Mathematical and Computing Sciences Tokyo Institute of Technology

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
27
Older →

Jump to page: