[ge-talk] A layout manager

[Christof Lutteroth]

Fredrik Holmqvist wrote:
> Looks like *Ingo *Weinhold has been working on something similar. Dunno 
> if he is on the GlassElevator mailing list so he may have missed this 
> discussion:
> http://cia.vc/stats/project/OpenBeOS/.message/642ad5
This commit caught my eye, too.
It seems as if Ingo was using QOCA (which includes the Cassowary solver) 
and is in the process of replacing it with his own code. As far as I can 
see, the Haiku layout system does not support linear constraints but 
uses a simpler approach.

Some news about ALM:
on the homepage
http://www.cs.auckland.ac.nz/~lutteroth/projects/alm/
there is now a zip archive with .net binaries for performance testing. 
The binaries implement the examples from the examples page
http://www.cs.auckland.ac.nz/~lutteroth/projects/alm/examples/index.html
It would be really good to know how the current implementation performs 
on older hardware, so if someone with such hardware could run the 
examples and send me the performance.txt file that is created in the 
working directory, that would be nice. You only need to make sure that 
the .net 2.0 framework is installed:
http://www.microsoft.com/downloads/details.aspx?FamilyID=0856EACB-4362-4B0D-8EDD-AAB15C5E04F5&displaylang=en
After starting an example, you need to resize the GUI a bit. The time 
for each layout calculation is recorded (with the initial calculations 
naturally taking longer than the subsequent ones).

In other news, here at my university a summer scholarship was granted 
for extending ALM and porting it to Java and... Haiku.
That is, we will most likely have a student working on it in the months 
before Christmas (summer in New Zealand).

Best regards,
Christof