Hi,
Perhaps some of you are interested in this event.
Cheers,
Doru
Begin forwarded message:
> From: Harald Gall <gall(a)ifi.uzh.ch>
> Date: 26 October 2009 22:56:24 GMT+01:00
> To: choose-news(a)iam.unibe.ch
> Cc: Harald Gall <gall(a)ifi.uzh.ch>
> Subject: [scg-staff] [Choose-news] CHOOSE Forum 2009: Human-centric
> software development, Nov 23, Bern
> Reply-To: scg-all(a)iam.unibe.ch, scg-staff(a)iam.unibe.ch
>
> Dear members of CHOOSE,
>
> We are pleased to invite you to the CHOOSE Forum 2009 to be held on
> November 23 at the Allegro Grand Casino Kursaal in Bern.
>
> The full day event consists of five talks and a panel on the topic of
> "Human-centric software development."
>
> The speakers are:
> - Prof. Arie van Deursen, Delft University of Technology
> - Prof. Jim Whitehead, University of California, Santa Cruz
> - Prof. Andrew Ko, University of Washington
> - Peter Stevens, Sierra Charlie Consulting
> - Dr. Tudor Gîrba, Sw-eng. Software Engineering GmbH
>
> The day also features the CHOOSE General Assembly, including reports
> from the President and the Treasurer.
>
> More details together with the registration form can be found on the
> official webpage:
> http://choose.s-i.ch/events/forum2009
>
> We hope to see many of you at the Forum!
>
> Kind regards,
> Harald Gall, Michele Lanza and Tudor Girba
>
>
> --
> CHOOSE News is a *moderated* mailing list for
> announcing non-commercial events related to
> Object Technology in Switzerland.
> To learn more about CHOOSE, see: http://www.choose.s-i.ch/
--
www.tudorgirba.com
"Relationships are of two kinds: those we choose and those that
happen. They both matter."
Hi!
There is a #generated method in some metaclasses of Mondrian. For
example:
MORectangleShape class>generated
^ #(( string '#asString' 'The string to be displayed.' )
( font 'TextStyle defaultFont' 'The font to be used.' )
( color 'Color black' 'The color of the text.' ))
The only sender of #generated is MethodPragmaTest
Is this really used? If not, is there a perpective of using this?
Cheers,
Alexandre
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
Dear All,
I adapted to Glamour what I did for Mondrian
You will find below the report. I tried to find a bottleneck, but
Glamour behaves linearly in all my tries. This is a good news
therefore :-)
The report has been generated by doit-ing:
GlamourousHealth new produceReport
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Report produced on 2009-10-23T16:33:00+00:00
Version a MCVersionInfo(GlamourLoader-tg.34)
------------------
Opening Browser Benchmark:
15 openings => 3622 ms
------------------
------------------
Selecting Item Benchmark:
100 size and selections => 709 ms
200 size and selections => 815 ms
300 size and selections => 895 ms
400 size and selections => 993 ms
500 size and selections => 1094 ms
600 size and selections => 1197 ms
700 size and selections => 1705 ms
800 size and selections => 1431 ms
900 size and selections => 1533 ms
1000 size and selections => 1639 ms
1500 size and selections => 2493 ms
2000 size and selections => 2729 ms
------------------
------------------
Selecting in finder item Benchmark:
1 size and selections => 26 ms
5 size and selections => 148 ms
10 size and selections => 328 ms
15 size and selections => 483 ms
20 size and selections => 636 ms
25 size and selections => 833 ms
30 size and selections => 999 ms
35 size and selections => 1289 ms
40 size and selections => 1483 ms
45 size and selections => 1700 ms
50 size and selections => 1881 ms
------------------
33.02 % of methods are covered
Progress from last time: 0.0 %
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
I also worked on a test coverage tool. Here is what I obtain for
Glamour.
Execute:
GlamourousHealth new viewTestCoverage
to have a mondrian rendering of the coverage. Inner light red boxes
are methods that are not directly or indirectly executed by unit
tests. Inner white boxes are methods executed when running the tests.
Outer boxes are classes. Edges designate superclass links.
Feedback are welcome
Cheers,
Alexandre
NB: You need to install www.squeaksource.com/Spy.html in order to get
the test coverage
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.
Hi,
I am happy to report that I ported SmallDude to Pharo. It is available
via:
Gofer new
squeaksource: 'SmallDude';
addPackage: 'ConfigurationOfSmallDude';
load.
(Smalltalk at: #ConfigurationOfSmallDude) perform: #loadDefault
All tests are green :).
Cheers,
Doru
--
www.tudorgirba.com
"Not knowing how to do something is not an argument for how it cannot
be done."
Hi,
I added a new DiffPresentation in Glamour based on the DiffMorph. You
can see an example of a Glamorous Diff Browser, by executing the
following in the latest Glamour:
GLMSmalltalkExamples new diffBrowser openOn: (
Collection withAllSubclasses asOrderedCollection sortBy: [:a :b | a
name < b name])
Cheers,
Doru
--
www.tudorgirba.com
"We cannot reach the flow of things unless we let go."
Hi,
I created new Metacello Configurations for Moose, Mondrian, Glamour
and DSM. They are in packages named by the scheme: ConfigurationOfXXX
To load it in the latest Pharo dev image, you should run:
Gofer new
squeaksource: 'Moose';
addPackage: 'ConfigurationOfMoose';
load.
(Smalltalk at: #ConfigurationOfMoose) perform: #loadDefault
Cheers,
Doru
--
www.tudorgirba.com
"Be rather willing to give than demanding to get."
Hi!
I have been working on a "health report" of Mondrian. For now, it
essentially do some benchmarks for nodes, edges and provides an
estimation of the test coverage.
This is very early work for now. I am still experimenting. The
measurements below clearly show a problem: adding edges is definitely
too costly. I am currently working on this.
In the future, I would like identify speed bottleneck.
For now, it is only for Mondrian. Will people be interested in getting
similar data for Moose?
Cheers,
Alexandre
Report produced on 2009-10-20T16:00:26+00:00
Version 'Mondrian-Alexandre_Bergel.277'
Benchmark ManyNode (simple rendering of nodes) :
100 nodes => 14 ms
200 nodes => 50 ms
300 nodes => 108 ms
400 nodes => 198 ms
500 nodes => 335 ms
600 nodes => 442 ms
700 nodes => 627 ms
800 nodes => 768 ms
900 nodes => 969 ms
1000 nodes => 1223 ms
1600 nodes => 2907 ms
Benchmark ManyEdges (simple rendering of edges) :
10 edges => 1 ms
20 edges => 6 ms
30 edges => 14 ms
40 edges => 27 ms
50 edges => 52 ms
60 edges => 82 ms
70 edges => 126 ms
80 edges => 178 ms
90 edges => 364 ms
100 edges => 355 ms
200 edges => 4719 ms
300 edges => 36495 ms
Test coverage: 60.03 % of methods are covered
Progress from last time: 0.0 %
--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.