[
Lists Home |
Date Index |
Thread Index
]
On Tuesday 17 December 2002 16:40, Bullard, Claude L (Len) wrote:
> Is there a set of properties that define "open"?
>
> Is there a subset of properties of all office systems
> that will make that cut?
I've been thinking that there are various properties that are 'open'; an open
system need not have all of them, and it's a fuzzy logic problem to say how
many you need to be 'open'.
Properties might include:
1) The data formats / protocols being publicly readable and implementable;
not necessarily all of them but enough to interoperate, eg some
implementations of open systems (even reference implementations) can of
course have private implementation-specific data formats that are used
internally
2) Reference source code for a programmer's library providing basic access to
the data / protocol being available.
3) A fully fledged usable reference implementation of the system being
available; this is a superset of (2).
4) The results of (2) or (3) being free software and usable, as opposed to
just 'source you can play with but not use for commercial purposes' or 'a
lame or crippled implementation that can at best just be a guide to
implementors'.
5) The published data formats / protocols being maintained by a standards
body or reasonably unbiased industry consortium, as opposed to being
published by the original manufacturer with little or no external input
6) The presence of more than one decent commercial implementation that are
actually competing, which tends to imply that if any one of them starts
trying to add proprietary extensions they won't be used because users will
resist the loss of interoperability
...any others? I've put them in an approximate ordering by strength.
--
A city is like a large, complex, rabbit
- ARP
|
