[erlang-questions] packages (was: newbie: why c.erl is special?)
Fri Mar 21 11:55:04 CET 2008
> If this is what everyone seems to want, why hasn't the package
> notation caught on?
The package notation is not what everyone wants.
> It seems to me (after reading the list archive) that everyone
> wants this, only some people want even more.
No, some people (myself included) have repeatedly argued that
packages are fundamentally wrong. We want something completely
different, not just better packages. Even having nothing would
be better than having packages.
In addition to recent posts from Richard O'Keefe and others, I
would like to point out two other fundamental flaws with
1) Packages don't actually prevent name clashes: they only reduce
the probability of a name clash. Accepting this, as a language
feature, would be like accepting that Pid ! Msg sends Msg to Pid,
but only most of the time.
2) Packages reduce referential transparency, and by way of
consequence, make it harder to move code around or experiment
with it in the shell.
I elaborated on this and other flaws here:
In addition to the Eiffel-inspired solution described there,
Richard O'Keefe outlined an even more elegant solution here:
More information about the erlang-questions