Support for non-unique process labels?

Michael P. empro2@REDACTED
Wed May 19 16:24:30 CEST 2021


On Fri, 07 May 2021 10:42:36 +0200
Nicolas Martyanoff <khaelin@REDACTED> wrote:

> > * register processes using any term and not just atoms

> I always wondered why there was such a limitation. Supervisors can use
> any term as child identifiers.

Is a registered name not like an office ID?
When the worker in there dies, they do not take
the office with them, nor does a new worker
bring a new office.
Is it not a certain position in ones application
structure, like an office in a building,
not the worker there?

What use is a globally registered name (atom or not),
that no-one knows of, apart from who- or whatever
created and registered that process?

Is this not about approaching s/th like an object request
brokery thingy, rather than about displaying,
say ?MODULE:?FUNCTION of a process,
in Observer and error messages?

~M

--

Curiosity killed the cat --
by simply using up its time.







More information about the erlang-questions mailing list