[erlang-questions] Parameterized module initialization
Thu Jun 28 01:30:37 CEST 2012
Definitely, things need to be tested before being added to the language. Many ideas which sound great when you describe them are terrible when you actually use them, or in the best case need some tuning before being added.
I think the only thing we really miss is process for testing new things before deciding whether they go or stay. Something along the lines of deprecating things before removing them. Couldn't we have an "experimental" tag so the compiler/system verbosely flags things as experimental until a decision is made. In that way they are there, you can test them but you know not to depend on them as they may go away. The say nothing is experimental for more than one release.
----- Original Message -----
> On 27/06/2012, at 5:33 AM, Andrew Berman wrote:
> > I'd also like to know why there are "hidden" features of Erlang in
> > the first place.
> Because it isn't dead. New ideas get tried, and sane designers don't
> take it
> for granted that their precious new idea is going to be right first
> So initial implementations get short-cuts so that people can get some
> experience with a new feature before designing and documenting the
> Real Thing.
> Strictly speaking, in this case it isn't the *feature* (parameterised
> that is hidden but the *implementation* of that feature.
> erlang-questions mailing list
More information about the erlang-questions