I agree that packages should be promoted to real objects vs. the result of clever (ab)use of category names. That said, I frequently use a general-to-specific naming scheme for related classes on the grounds that it seems to help me when I get an alphabetical view of names. With that same logic, I suppose one could argue for just about any of the options described here :) Sorry. In addition to reified packages, I would like to float the idea of allowing multiple method categories to apply to any given method. That has been very useful in Dolphin. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab@anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029

...

...

luc.fabresse@gmail.com 9/11/2008 3:09 AM >>> I understand the problem and I did not explain well but the solution of the Graphics package seems great : Graphics-Text Graphics-Files

and GraphicsTests-Text GraphicsTests-Files Other packages like Graphics : CollectionTests-*, KernelTests-*, ... Packages with other naming conventions that do not seem great : File-Tests, Tests-*, Monticello-Tests Sure, a system that does not rely on string maching would be better. #Luc 2008/9/11 Stéphane Ducasse <stephane.ducasse@inria.fr>

On Sep 11, 2008, at 9:35 AM, Luc Fabresse wrote:

the convention <category>Tests[<-subcategory>] seems great IMO.

...

e.g. Graphics-Text -> GraphicsTests-Text

With this naming pattern it is possible to have 2 monticello packages : <category> and <category>Tests And it is possible to not load Tests packages in a deployed image.

Yes it does not work if Graphics is composed of several subpackages

With this I cannot have

Collections-OrderedCollection and Collections-OrderedCollection-Tests

separated from

Collections-Set and Collections-Set-Tests

For me I want to have packages that I can load independently and the tests should be attached to the package (to be eventually not load with it).

...

I really hate the fact that we feel everywhere category and string matching.

Stef

...

Luc

2008/9/11 Alexandre Bergel <alexandre@bergel.eu> Having a package name without any - will prevent Monticello from being confuse on some point. Stef and I already experienced that...

I was never happy in inserting a - in a package name.

Alexandre

On 11 Sep 2008, at 07:23, Damien Cassou wrote:

On Wed, Sep 10, 2008 at 9:32 PM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote: Hi

I noticed that we have CollectionsTests-Ordered kind of package name and I would prefer to get Collections-Ordered-Tests What do you think. I bit of regularity would help.

Collections-Test-* categories belong to the Collections-Test package on monticello. If you rename it to Collections-Ordered-Tests, you will put everything in the Collections package (because there is no Collections-Ordered monticello package).

-- Damien Cassou Peter von der Ahé: «I'm beginning to see why Gilad wished us good luck». (http://blogs.sun.com/ahe/entry/override_snafu) _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

-- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project