headers
Ian Piumarta | 25 Mar 1998 11:53
Picon
Picon

Re: Is Squeak ready for BeOS for Intel

> I recall that at one point in time there were a couple of individuals who
> were working on the port. Is it available?

I started last year, but that particular process got moved from my
runnable queue to the blocked list due to an avalanche of higher priority
things that arrive ceaselessly.  The tricky part is reconciling Squeak's
assumptions abouth the platform's graphics model with the MVC-like triad
that is imposed on applications by the BeOS application framework.

I forget the name of the other gentleman who was working on it, but
his progress seemed promising for a while...

A real "quick and dirty" solution would be to use the X-server-in-a-
window that someone did for BeOS, but I haven't investigated this.  I
might take a look at the possibility of this approach, and if it requires
only a few hours of work then I'll try to get something going.

Ian
Permalink | Reply | 0 comments |
headers
Luis Quijano | 23 Mar 1998 02:16

Re: Self and pure object thinking

Hi,

An interesting but little "marketed" notation appears in the book:
     Title: Object-Oriented Systems Analysis
     Authors: David W. Embley, Barry D. Kurtz, Scott N. Woodfield
     ISBN: 0-13-629973-3
     Publisher: Yourdon Press
     Year: 1992

Regards,
Luis Quijano
quijano <at> panasoft.com
p.s.  Check-out the Smalltalk Links at http://www.panasoft.com/stlinks/
-=-=-
Permalink | Reply | 0 comments |
headers
sqrmax | 29 Mar 1998 03:43
Picon

1.31

Hi.

Yes! The Squeak 1.31 VM doesn't take up cpu when it's doing "nothing"! 
Great.

Andres.
Permalink | Reply | 0 comments |
headers
Georg Gollmann | 13 Mar 1998 18:36
Picon

Re: YAW

>I have been thinking about  that page locking mechanism.
>What happens if someone edits a page for more than 10 minutes,
>then sends it in?  is it rejected or not?
>
>What if Andy edits a page, typing in his thoughts etc for
>20 minutes... then Barb edits the same page, typing in her
>thoughts... then Andy submits his edited page... then Barb
>submits her edited page.
>
>what if Barb submits hers before Andy submits his?

Some points:
- Timeout is only activated on demand (when someone else accesses the page)
- While an edit is in progress others don´t even get update forms, update
requests are rejected.
- Editing is a multi step process: Check the page out, edit it, edit it
some more..., commit or abort. So just send the intermediate stages to the
server every few minutes (good for checking the formatting anyway), only
after a commit others will see the new version.

BTW, the YAW code has only seen light testing, so expect it to contain bugs !

Georg

----
Dipl.Ing. Georg Gollmann                   TU-Wien, EDV-Zentrum

phon:(+43-1) 58801 - 5848
fax: (+43-1) 587 42 11
mail:gollmann <at> edvz.tuwien.ac.at
(Continue reading)

Permalink | Reply | 0 comments |
headers
Mike Klein | 3 Mar 1998 05:40

Re: On a more...(Inboard Squeak sources)

1) I would never give up the security of the change log.  I want the sources
   internally in addition.  (You people saving hard-disk power can disable
   if you want...)

2) Having to distribute a new changes file with each new revision kind of
   defeats the purpose of having an unchanging sources file.  Besides,
   The changes file is almost twice the size of the sources.  Does the
   changes file contain the source of every released method (including
   past released versions)?

3) I know that it's easy to change.  But if I do it in my own little world,
   all of those people who are getting their first take on Smalltalk from
   Squeak do not benefit.  (I am an experienced Smalltalker, and my first
   take on Squeak was: "Sheesh... I can't seem to get the line end convention
   to show all of the methods properly.... what a hassle").

4) Having things internal has the nice feature of not needing weirdo platform-
   specific configuration.  If the windows come up, "everything" works.

5) A quick dig around in Squeak  (especially internalizeSources, etc.)
   reveals a rather annoying problem:  CompiledMethod is a byte-type object,
   and therefore cannot store a pointer to a source object.  It can only store
   some number that the source code system must then translate to the source.
   So much for non-string based source code...

   -- Mike Klein

> Folks -
> 
> It's not hard to bring the Squeak source code inboard -- it actually all
(Continue reading)

Permalink | Reply | 1 comment |
headers
Ken Dickey | 3 Mar 1998 00:07
Picon
Favicon

Sorry

Apologies to all, wrong return address.

-Ken
Permalink | Reply | 0 comments |
headers
Tim Rowledge | 18 Mar 1998 18:25

Re: Portable Cheese Protocol vs. Implementation

I may well ahve mentioned this paper before, but take look at:
'Creating Host Compliance in a Portable Framework: A Study in the use of Existing Design Patterns'
P.M.Yelland, pp18-30, OOPSLA'96 Conference Proceedings, ACM SIGPLAN Notices Vøl.31, Number 10, ISBN 0-201-92109-X
which is a pretty good paper on how this was tackled by the VanGogh project at ParcPlace. Lots of useful ideas
and possibly some warnings on what not to do.
Basically describes a three layer structure:-
  portable window code (protocol and implemetation)
  bridge code (portale api, platform specific implementation)
  system call code (platform specific all the way)
and the ways of swapping layers in and out, where state has to go etc.

Of course, if you can find another paper on the same area, then it
becomes research rather than any sort of plaigarism :-)

tim
--

-- 
Useful random insult:- Donated her body to scientists...  Before she
was done using it.
Tim Rowledge:  rowledge <at> interval.com (w)  +1 (650) 856-7230 (w)
 tim <at> sumeru.stanford.edu (h)  <http://sumeru.stanford.edu/tim>
Permalink | Reply | 0 comments |
headers
CYNTHIA NELSON | 2 Mar 1998 03:36

On a more mundane level Was: "Open Implementations

At 03:52 PM 3/1/98 -0800, Alan C. Kay wrote:
>At 3:36 PM -0800 3/1/98, Bruce Cohen wrote:

Hi, 

I have a more mundane question than the those
that normally are posted to this list...however
I would like to know WHAT I am supposed to do 
about the fact that ver 1.31 seems to capriciously
change my carefully chosen temporary variable names
to t1, t2,.... This is a royal waste of my time to
have to go back an change this. I have tried everything
in the "documentation" to no avail. I even created a
changes folder before I saved as something else...
nothing seems to work. Also I have had a problem
saving workspaces to an external text editor sometimes
it works and sometimes it does not.

I am used to C++ (OK throw MUD now) where everything may 
be very verbose and dangerous for the inexperienced but 
at least it is fast and reliable...maybe that explains 
part of my frustration. Now you guys are most likely going
to say it is a bug in the C compiler used to create Squeak 
that is to blame. I do not care just answer the question 
I would appreciate it.

Cindy Nelson

PS. I would like to try the sound capabilities since I am a
professional musician but first the "nuts and bolts' have to
(Continue reading)

Permalink | Reply | 3 comments |
headers
Maloney | 23 Mar 1998 22:52
Picon

Re: Just some thoughts

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
(Continue reading)

Permalink | Reply | 2 comments |
headers
Ken Dickey | 2 Mar 1998 23:23
Picon
Favicon

Dinner Plans

Mike,

Los Gatos? Where? When?  Which Meta-MOP? 8^)

Cheers,
-Ken
Permalink | Reply | 0 comments |

Gmane