Re: second draft of Charter for Foundation (Needs reorganization)

[jg pa dec com (Jim Gettys)]

Bart,

The biggest single issue I have with the document as it stands is that
too much of the details are hardwired in the (single) document.
Unfortunately, it is a problem that requires more reworking than simple
edit suggestions can make.

As I noted in a previous message, what is really needed are two documents:

	1) the charter document which establishes the foundation, its general
	goals, and as little of the basic process as possible (though
	some of these decisions are truly fundamental, such as whether the
	process	can be modified by the board, or by the normal operating
	processes, or by referendum, or...). In fact, these are the truly
	fundamental discussions about how any organization is run, rather than
	the detailed processes of most of the discussion to date.

	2) the initial set of processes to be followed, which can be modified
	by whatever is decided in 1).

If things aren't organized this way, we may find that modifications to
the processes found unworkable by experience are somewhere to between
difficult and impossible.  For example, the rules on voting (I've already
commented on why I believe voting should be the process of last resort)
may be found to be, in practice, impossible (getting quorum in the small
town I live in can be a challenge for the town meetings).  As I understand things
(not well), changing the basic rules of an organization after incorporation
gets much harder.

All my experience is that fixing meta-process after the fact is REALLY
painful, and sometimes down-right impossible.  In W3C's case, some parts
of its process effectively CAN'T be changed (they are enshrined in > 300
contracts each member has signed).  I observed a number of major, key
players, almost decide not to participate early on as a result of this
situation, and it causes trouble to this day, despite near unanimous
agreement including the consortium staff itself that if W3C was to be
founded again, it would do so under different terms.

This is similar to the structure of most governments: there is the
"constitution", and the "laws".  Amending the constitution is very difficult,
whereas laws come and go (and much of the processes for lawmaking itself
are left in the hands of the legislature itself, in their rules).

					- Jim


> From: jg@pa.dec.com (Jim Gettys)
> Date: Fri, 14 Jul 2000 07:03:01 -0700 (PDT)
> To: bart@eazel.com
> Cc: Nat Friedman <nat@helixcode.com>, Maciej Stachowiak <mjs@eazel.com>,
>         foundation-list@gnome.org
> Subject: Re: Draft of Proposal for the GNOME Foundation.
> -----
> One other important global comment:
>
> What is needed now is the "meta-process": the way that Gnome processes
> for making decisions get made.  This is all that should be enshrined in
> any organizational document.  You know that over the life of Gnome
> (potentially very long), the processes under which Gnome is governed will
> need to change.  Much/Most of the discussion to date has just been discussion
> at the process level, rather than at the meta-process level.
>
> To enshrine the processes themselves in the fundamental organizational
> document is an organizational nightmare, as it can be impossible to
> change them without great hassle: instead, documents should be
> structured:
>         Metaprocess -> initial process
> as a pair.  Only the metaprocess document is part of any legal incorporation,
> though they might often state things like:
>         "There shall be an appeals process."
>
>         "Processes can be modified by the process modification process" or by:
>         "Vote of the board", or what ever.  Note that the difference is key:
>         in one case it is a board decision, in the other, the normal governing
>         processes are used to change how decisions get made.
> The companion document, not part of the incorporation documents directly,
> has the actual initial appeals process/and or process modification process.
>
> There are legal terms for all this kind of stuff, but I'm not a lawyer.
>
> Again, I look to the IETF: its processes can be changed by the normal
> standards process they defined (its processes are RFC's along with all
> other documents)  As in GNU systems, the IETF is fully recursive
> internally... :-).
>
>                                 - Jim
>
> --
> Jim Gettys
> Technology and Corporate Development
> Compaq Computer Corporation
> jg@pa.dec.com


> From: jg@pa.dec.com (Jim Gettys)
> Date: Fri, 14 Jul 2000 07:03:01 -0700 (PDT)
> To: bart@eazel.com
> Cc: Nat Friedman <nat@helixcode.com>, Maciej Stachowiak <mjs@eazel.com>,
>         foundation-list@gnome.org
> Subject: Re: Draft of Proposal for the GNOME Foundation.
> -----
> One other important global comment:
>
> What is needed now is the "meta-process": the way that Gnome processes
> for making decisions get made.  This is all that should be enshrined in
> any organizational document.  You know that over the life of Gnome
> (potentially very long), the processes under which Gnome is governed will
> need to change.  Much/Most of the discussion to date has just been discussion
> at the process level, rather than at the meta-process level.
>
> To enshrine the processes themselves in the fundamental organizational
> document is an organizational nightmare, as it can be impossible to
> change them without great hassle: instead, documents should be
> structured:
>         Metaprocess -> initial process
> as a pair.  Only the metaprocess document is part of any legal incorporation,
> though they might often state things like:
>         "There shall be an appeals process."
>
>         "Processes can be modified by the process modification process" or by:
>         "Vote of the board", or what ever.  Note that the difference is key:
>         in one case it is a board decision, in the other, the normal governing
>         processes are used to change how decisions get made.
> The companion document, not part of the incorporation documents directly,
> has the actual initial appeals process/and or process modification process.
>
> There are legal terms for all this kind of stuff, but I'm not a lawyer.
>
> Again, I look to the IETF: its processes can be changed by the normal
> standards process they defined (its processes are RFC's along with all
> other documents)  As in GNU systems, the IETF is fully recursive
> internally... :-).
>
>                                 - Jim
>
> --
> Jim Gettys
> Technology and Corporate Development
> Compaq Computer Corporation
> jg@pa.dec.com