[OpenISO] OpenISO.org Core Guidelines

[Std Lib0]

On 9/10/07, Norbert Bollow <nb at bollow.ch> wrote:
>
>
> ## Decision-making Processes and Requirements ##
>
> The fundamental decision-making principle is of course that decisions
> should be made in a fact-based manner.  The fundamental question is:
> What happens when there is no consensus about an appropriate
> fact-based decision?


could you please give an example of a situation like this?
it seems to me that if the process is fact-based, then there should be no
doubt about the correct decision in an open discussion, since it is possible
to verify if arguments are factual or bogus based on publicly available
facts.
if the standard does not allow such argument verification (because of
confidential/undisclosed information, intentional ambiguous claims etc),
then the standard is incorrect/broken and should be modified to be made
verifiable.
It looks like 'public verifiability of claims' is the fundamental property
here
that should be required in every sentence of an open approved standard.

the problem with the decision process of ooxml was that it wasn't fact-based
-- it wasn't even public in some national bodies! -- but mainly
voting-based. if ooxml
had been approved by openiso because of lenient working groups, the
fact-based approach would automatically let it vulnerable to a quick
fact-based problem report and an immediate rating correction.

as a side note, I suggest the creation of a draft standard containing a tree
of
pertinent recurrent issues in submissions, which should always be verified
by any working group against any submitted standard, to verify if the
submission
contains a bare minimum fact-based verifiability. if not, the standard
won't be even considered further.
e.g.
- are all the references public and easily accessible by direct ftp/http
download of a url on the Internet, without having to fill forms or sign
covenants?
- does it propose a complete verifiable mapping between the referenced
functionalities and the standard's functionalities (e.g. mapping between
ooxml-doc formats, between ipv6-ipv4 protocols etc)?
- is it mature (have a suitable freely-licensed available and
Internet-accessible implementation which allows to verify implementability -
so we don't have things like doLineBreakLikeMyUndisclosedMethod)?
- is it patent-free (is verifiably unencumbered)?
- (add your preferred recurrent verifiability issue here)

extra nice requirements (not verifiability-based):
- is it original and do not replicate other standard's functionalities?
(functionality duplication in different standards is bad, they should be
referenced instead)
- ...

In most existing standardization organizations,
> at that stage the decision-making process either breaks down or the
> committee members vote.  Both of these mechanisms are inadequate
> because is the first case a company like Microsoft can prevent any
> decision it doesn't like by means of breaking the consensus, and in
> the second case it can manipulate every vote by means of telling
> enough "gold certified" (economically dependent) partner companies to
> vote, like it happened in the case of the ISO/IEC evaluation of OOXML.
>
> Therefore, OpenISO.org adopts the rule that all disgreements which
> cannot be resolved by means of Working-Group consensus processes must
> be addressed by means of an independent, professional evaluation of
> the relevant facts.


given my observations above, it seems that an independent closed-doors
evaluation is not required in a fact-based decision process, so DCSs are
not necessary.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://OpenISO.org/pipermail/discuss/attachments/20070911/dcb813b4/attachment-0001.htm