I've already written about it in one of the previous talks - I think that
we can have a good approximation of categories from the "Community Over
Code".
Except few language-specific categories there (Groovy which is clearly cut
by "language" not category) pretty much tracks there are pretty
As for languages, some DOAPs currently contain categories that don't
agree with the current valid list.
The outliers are:
data-management-platform
distributed-sql-database
education
ftp
geospatial
graphics
hadoop
IDE
identity-management
identity-provisioning
integration
Kerberos
observability
