20 May
2005
20 May
'05
11:19 a.m.
On Fri, May 20, 2005 at 01:31:16 +0100, Simon Jenkins wrote:
On Fri, 2005-05-20 at 11:03 +0100, Dave Griffiths
wrote:
A heirarchical ID system - a bit like OIDs in SNMP - might be worth
considering. The IDs form a tree rather than a flat space, and once you
own an ID node in the tree (like in SNMP when you are allocated an
"enterprise" OID) then you own every node beneath that node in the tree.
Or, we could use a system thats been demonstrated to work really well...
http://www.w3.org/Addressing/
- Steve
On Thu,
May 19, 2005 at 03:19:50 +1000, Dave Robillard wrote:
In the commercial world (of one application I develop for anyway) a number
of unique plugin ids are allocated per developer licence of a piece of
host software. There are also some that are published but never allocated,
and kept free for other development purposes.
In practise, everyone randomly uses id's that don't cause problems - or
uses the ones from some company they worked for a few years ago (this is
mostly for internal use)
It's chaos in other words, so it's probably not the model to follow :)
(I still think the central repository is a good
idea anyway, FWIW)
So do I, but its a hell of a lot of effort, and its unlikly to be kept
up to date.