From: Simon Denier <Simon.Denier(a)inria.fr>
Date: 20 novembre 2009 16:45:05 UTC-03:00
To: Simon Denier <Simon.Denier(a)inria.fr>
Subject: Re: [Moose-dev] The case of MondrianPaintings
On 18 nov. 2009, at 12:39, Simon Denier wrote:
> On 17 nov. 2009, at 12:15, Alexandre Bergel wrote:
>> Dear List,
>> MondrianPaintings is a package belonging to the Mondrian
>> repository that contains few classes apparently to compute cycles,
>> render a distribution map, and render some views related to Flair.
> Problem is we never really discussed what this MondrianPaintings is
> about. I understand it as a bunch of applications based on
> Mondrian, more complex than basic stuff, and independent of Moose.
> But it still a bit in the gray.
>> There is a number of problems with this package:
>> - It contains too few tests. Among the 9 test run, 4 raise an
>> error. Apparently CollectionExtensions is required, which belongs
>> to Moose. Mondrian cannot depends on Moose. Maybe
>> CollectionExtensions may be moved into a standalone package for
>> which both Mondrian and Moose can depend on.
>> - MOSmalltalkPaintings>>viewWAPackageHierarchy depends on
>> WAPackage which is not present in Mondrian and in a Dev image
> yes this is viz for seaside packages
>> What should we do with this package? We could move it to Moose,
>> but identifying cycles is an interesting feature that go beyond
>> Moose. As far as a I can see, making CollectionExtensions an
>> independent package looks the best compromise.
> Actually this package does not contain any cycle identification
> algo (there are in MooseAlgos and DSM for now), it's just stand-
> alone visualizations in which you can put data.
Fixed. I actually move the cycle tables and layer tables classes to
MooseAlgos. Please test.
> I agree that CollectionExtensions should have its own project.
>> Alexandre Bergel http://www.bergel.eu
>> Moose-dev mailing list