To view parent comment, click here.
To read all comments associated with this story, please click here.
Excellent point. Packages do represent a finer level of granularity than what would typically be called an application. I don't really know how to take this into account in any reliable way. Many packages that we would certainly consider to be "applications" can serve as a dependencies for other "applications," so the distinction of not having any dependents is not useful. We could count the number of packages that attempt to install menu items in the native DE, but I feel that this list would be significantly too narrow. I guess we could define an application as a set of packages that share a unique project homepage?
This incongruence is a direct consequence of the practical considerations for delivering software on proprietary and open source systems. Applications in the proprietary world represent the largest set of code that provides a set of related features with minimal external dependencies. Packages in the open source model are rather the smallest set of code that provides a set of corequisite features with maximal external dependencies. At least these are the ideals. It's not hard to find OSS packages that could--and should--be split into finer-grained packages, such as the various components of OpenOffice.






Member since:
2006-01-06
Once again, disregarding development tools and library packages, only considering packages marked stable on x86 (which is Vista's most supported platform), and only counting each package once even if multiple versions of that package are marked stable, I count 4734 packages.
Packages != Applications.
Firstly, many of those packages will be things that are part of the base system (ie. Bash, X, text editors, utilities, etc).
Secondly, many apps are made up of more than one package. OpenOffice is about 5 packages, GIMP is 2, etc.