Squeak-dev March 1998

squeak-dev@lists.squeakfoundation.org

67 participants
130 discussions

Squeak and object persistence
by Carl E Gundel 25 Mar '98

25 Mar '98

Squeak is meant to be a laboratory for exploring the pink and blue directions of objects and Smalltalk, correct? What kinds of blue ideas are bumping around in people's heads about object persistence? I'm currently involved in development using TopLink as a persistence manager, and it's dissatisfying to design object oriented systems, and then have to throw away what's right with what performs well when mapped to a relational database. So... I'm wondering what the next big thing is in object persistence. Carl ------------------------------------------------------------------ Carl Gundel carlg(a)world.std.com Shoptalk Systems 508-872-5315 author of Liberty BASIC, a 1996 PC Magazine Awards Finalist! http://world.std.com/~carlg/basic.html ------------------------------------------------------------------

2 1

Re: Just some thoughts
by Maloney 24 Mar '98

24 Mar '98

Andreas Raab wrote: >How can we bring things in the "main distribution" then? As far as I >understand the problem is not that of "being in the main distribution" but >rather of "letting people know that this particular application exists". >How many people out there _do_ actually know the contents of the >app/goodies directory at the ftp servers? > >Letting people know that the application is only a (ftp-)fingertip away, >however, is a completely different issue. I could think of a scheme of >"remotely loadable applications" in which the actual image only contains >the names and the (ftp or http) locations of these apps. Then, while in >the image you can have a look at these apps. If you want to try one, you >just click on "download" and the app is being fetched from the next ftp >server around the corner. Or it can fetch a description first, or tell you >what other apps are required first if it has been tested with your current >image/vm version or whatever. > >Ok, now back to the coordination issue. Given that we would use this >sort of remotely loadable apps, all we need for coordination is actually >dropping a note to Squeak Central saying: "Well here is this app called >MyGreatSqueakApplication and its on ftp://ftp.nowhere.domain/pub/myApp" >This doesn't sound like big deal to neither the author or Squeak Central >and could probably easily managed. Good analysis! I like your idea of filing in applications directly from an FTP or HTTP server at the click of a button. However, I'd suggest using the Squeak Wiki to advertise applications and their URL's rather than Squeak Central. That way, application information can be updated by the application maintainer directly as often as necessary. If this information were maintained in the image by Squeak Central (say in the System Workspace or a similar window), then it would only be updated once per release cycle. To support this mode of use, I've created a new area on the Squeak Wiki (http://c2.com:8080) entitled "SqueakApplications", and added an example entry for Georg's Web Server. (Georg, I hope you don't mind. Feel free to edit it!) I also added a page entitled "ApplicationedWanted", similar to the PortWishList page. This may help coordinate sets of people interested in porting some application (such as T-Gen) to Squeak. Wiki's are easy, wiki's are slick, Check out the Squeak wiki, It really is quick! -- John

3 2

Re: headless vm changes suggestions
by Eliot & Linda 24 Mar '98

24 Mar '98

Andreas Raab wrote: [snip] > ... In particular I hate the strange Win95 behavior in > some aspects of win32 (if you need proof then grep the VM sources for "due > to a Win95 bug...") and I am not sure to what extent threading is really > working on '95. In providing a threaded interconnect for VisualWorks I've not found any differences in functionality between '95 and NT. I have found that NT is more efficient by about a factor of 2, but '95's threads seem very solid. _______________,,,^..^,,,_______________ Eliot

1 0

Re: headless vm changes suggestions
by Craig Latta 24 Mar '98

24 Mar '98

Hi John-- > Just to clarify: this issue is Win32-specific. The way Win32 signals > changes of network socket status (data arrivied, other end closed, etc.) > is through events that are placed in the window's event list. > On other platforms, such notifications are handled in other ways, > such as by using an asynchronous callback. Did my previous message on this subject get through? Win32s supports threads. For each socket, one could have a thread for reading/connecting, and a thread for writing, and use blocking procedure calls in each. The procedure calls themselves conform to the BSD sockets API (like the other platforms), and the thread overhead is small, whereas the win32s async stuff is unique and platform-dependent. Why not do this? thanks, -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)]

3 2

Comparing Fractions
by sqrmax＠cvtci.com.ar 23 Mar '98

23 Mar '98

Hi. Fraction implements #<, but not #>. This results in super > and so... for instance, what should take 15ms took about 2 minutes! Not funny at all! Check implementing #> like, for instance, this: #> aNumber ^(self < aNumber or: [self = aNumber]) not The methods on #< and #= will take care of adapting and so. Andres.

1 0

Re: headless vm changes suggestions
by Andreas Raab 23 Mar '98

23 Mar '98

John, > I'm not entirely clear what the benefits of true "headlessness" > are on the Windows platform. It seems to me that one can always > make the Squeak window quite small (to use less Display memory), > then minimize it so it is out of the way. That way, if there are > any errors, one can re-size the window and debug the problem. Not if you're going to install it on any jointly used machines. Think of a Swiki web server running as a service under NT. > Headless-ness seems to make a bit more sense on a Unix box. > But there, again, one could open Squeak's display as a remote > X window on some other machine, just as a debugging console. Yeah, this would be cool. Install a Swiki demon, telnet on a special port and run your debugging session remote without even letting others know that something happens right now. WHOW! 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 >=============+

2 1

Re: headless vm changes suggestions
by Ian Piumarta 23 Mar '98

23 Mar '98

> Yeah, this would be cool. Install a Swiki demon, telnet on a special port > and run your debugging session remote without even letting others know > that something happens right now. WHOW! Would VNC be useful in this context? http://www.orl.co.uk/vnc Ian

1 0

FileDictionary and CachedSwikiAction
by Mark Guzdial 23 Mar '98

23 Mar '98

FileDictionary is the super-simple database for Squeak that I mentioned earlier in the list. It seems to be relatively stable on multiple platforms, so let's call it beta-ware. It's available (with some docs) off of http://www.cc.gatech.edu/fac/mark.guzdial/squeak/ I've also updated the PWS pages with some new PWS/Swiki facilities, including a CachedSwikiAction which seems to be improving speed for a large Swiki, the one for my Squeak-using class (http://pbl.cc.gatech.edu:8080/cs2390) with well over 300 pages. Mark -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 (404) 894-5618 : Fax (404) 894-0673 : guzdial(a)cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html

1 0

Re: Just some thoughts
by Maloney 23 Mar '98

23 Mar '98

Re: >> S Q U E A K N E E D S C O O R D I N A T I O N ! > >What do you think about managing this with a Wiki/Swiki? Some of the >software engineering folk here have been talking about using a Wiki for >development efforts. Ward has used a Wiki to document Wiki (as a form of >literate programming) and for releasing a version of the code at >http://c2.com/cgi-bin/wikibase. The Wiki structure really lends itself to >System Documentation and Specs Maintenance. We could use Ward's Wiki >(http://c2.com:8080/) or one of my Swikis' (say >http://pbl.cc.gatech.edu:8080/myswiki), or even a special-purpose (s)wiki >(maybe named squeakCentral? :-). For uploading images or other binaries, >we could extend Swiki with a file upload/download facility. We've started >working along these lines here (for up/download of image files). Squeak >1.31 has FTP client sockets, which is a nice start. Joerg: I agree with Mark; using a Wiki is the ideal solution to this problem. And it so happens that Ward already has a lovely Wiki devoted to Squeak at: http://c2.com:8080 If you don't see a section there that satisifies your requirements, just create one! For example, you could create a "Applications Wanted" section and add an entry for T-Gen describing what it does and where to find existing source code and documentation for it. It takes less time than sending an email. -- John

1 0

Cheese thoughts
by Andreas Raab 23 Mar '98

23 Mar '98

Boris, Sorry that I've been so quiet in the discussion about Cheese. In general, I'd _love_ to have some native window support if the platform supports it. It makes one feel like being at home ;-) I have had a look at a previous version of Cheese and I like the general approach of it. There are, however, still some open questions I would like to raise. Such as for instance the question, how events should be handled. To give an example, your OS/2 implementation defines an external callback function for the window events. While this is a suitable approach, it hides some significant functionality from the ST side. There may be events one wishes to process directly (i.e. blocking resizes or other stuff) and not to leave to the VM. An example for this is, how I intercept the usual ALT-F4 window closing mechanism in the Windows VM of Squeak. There is probably no easy way of dealing with it, but here is a suggestion: One could try to "configure" the VM (i.e. the external callback) with actions to be taken, such as: * do the default action (this is triggered if the window has not registered interest in an event) * ignore it (don't do anything upon receiving this event) * do the default action and inform the window that this has happened (could be default for actions like window resize etc.) * ignore it and inform the window of this event (could be default for closing a window, when the contents of the window needs to be checked for modifications) Another question is that of drawing graphics. While the modern GUIs provide us with some high-level widgets for things like scroll bars, lists or even whole text editors, in a whole lot of applications this is just not enough. Given that Cheese is not only for "08-15" applications [1] we need some sort of graphics output device. I could think of using class Canvas as a base for these operations. It's already in the image and its functionality seems rather clear and well structured to me. Even more, by using Canvas as the base, we would immediately have the ability to run any morphic world in any OS window. Finally, I would love to see a pure emulation layer in Cheese which is entirely based on the current Squeak desktop approach. Since Squeak can be used on machines with virtually no OS there should also be some way of using Cheese on it. Because then, we could easily use an approach like "develop on your desktop machine - run it on any handheld device". And I would love to see Squeak/Cheese on any toaster in this universe ;-) Just some thoughts, Andreas [1] "08-15" - the standard rifle of the germans in world war I. Has become a general synonym for not doing anything besides the main direction (or, with Alan's words, entirely pink stuff). -- 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 >=============+

2 1

← Newer
1
2
3
4
5
6
7
...
13
Older →