[Go-essp-tech] Access control for data with different QC Level

Wed Jul 21 02:28:16 MDT 2010

Hello All,

A couple of thoughts on enforcement and openness (supporting Karl's
approach).

Firstly, on openness, we don't want people saying "the fact that only
10% of the data is made available to the public completely invalidates
the review process". It is the appearance of holding back data that
matters, not the efforts exerted to make the important data as widely
available as possible.  

Secondly, on enforcement: basically, we can't. If we make access
conditional on approval from modelling groups it will be essentially
random and based on who you know -- some groups will approve everyone
automatically, others will spend months trying to decide what the
process should be. One approach would be to have self-assessed
categories: "Expert User" and "Assured Data User", where the names
indicate things which would be spelt out in more detail in the terms of
use: the "Expert User" has access to data which should be used with
caution, the "Assured Data User" has access to data which has been
through an independent quality control process. 

I think it would put modelling groups in a difficult position if there
is any indication that their permission is needed prior to publication
of analyses based on the data. If there is any change to Karl's wording,
I would limit it to something indicating that "Expert Users" should
notify the modelling centres prior to publication of any analysis of
their data.

Cheers,
Martin

> -----Original Message-----
> From: go-essp-tech-bounces at ucar.edu [mailto:go-essp-tech-
> bounces at ucar.edu] On Behalf Of Karl Taylor
> Sent: 20 July 2010 18:30
> To: go-essp-tech at ucar.edu
> Subject: Re: [Go-essp-tech] Access control for data with different QC
> Level
> 
> Dear all,
> 
> I should have said "The other 1% of the *research articles* will use
> data that is based on output that has not be replicated or assigned
> DOI's.  This data is still of substantial scientific interest ..."
> The non-replicated data volume will be similar or larger than the
> replicated data volume.
> 
> Karl
> 
> On 7/20/10 10:24 AM, Karl Taylor wrote:
> > Hi all,
> >
> > We clearly need to get clarification/guidance from the modeling
> groups.
> > I must say, however, that I'm guessing that the DOI assignment will
> > cover data published  in 99% of the research articles.  The other 1%
> of
> > the data is still of considerable research value (just as all the
> CMIP3
> > data was valuable, even without DOI's).   This "other" data should
> *not*
> > be excluded from research (and journal publication).  Given that
> there
> > seem to be limits to what can be accommodated in the "replicated"
> subset
> > (which is what is destined for DOI assignment), I see no reason to
> > modify what I said earlier other than:
> >
> > We should add a line to the terms of use that when data is used in a
> > research publication, the user must follow the rules for citation
and
> > acknowledgment that will be found on our website.  One of these
rules
> > will be that if a DOI has been assigned to the data, then it must be
> cited.
> >
> > We will have to get the modeling groups to agree to this, but, like
> > Balaji, I think they will be in favor.
> >
> > Best regards,
> > Karl
> >
> > On 7/20/10 10:00 AM, V. Balaji wrote:
> >
> >> Karl, I must disagree. I think the principle that the modeling
> centers
> >> deserve acknowledgment from users of the data in the form of a
> citation
> >> has to be a bedrock requirement, and I thought the WGCM leadership
> >> understood and agreed.
> >>
> >> I agree that the WGCM may not have a position on QC and versioning
> per
> >> se, but they are necessary steps for a robust citation system, as
> the
> >> Michael et al document shows. The WGCM must take a stand on
> citations.
> >> If they do the rest follows.
> >>
> >> I apologize if I am being hasty and misunderstood your remark, as
> >> I'm rushing off to the next damn thing... (It's one of those "one
> >> damn thing after another" days...)
> >>
> >> Thanks,
> >>
> >> Karl Taylor writes:
> >>
> >>
> >>
> >>> Hi all,
> >>>
> >>> Hold the presses please!
> >>>
> >>> My understanding is that WGCM expects *all* output registered with
> ESG to be
> >>> made available to anyone who signs the terms of use, as soon as
the
> output is
> >>> available (with perhaps a requirement that model documentation
also
> be in
> >>> place).  There are no requirements for QC or replication placed on
> the
> >>> output.  [There aren't even any expectations that "versioning" be
> >>> implemented, although I think they will be pleased if it is.]
This
> is a
> >>> simple requirement, which should be easy for us to meet.
> >>>
> >>> It would, of course, be helpful for us to include information to
> users
> >>> concerning what QC checks have been performed, but that is not
part
> of the
> >>> requirement, as I understand it.
> >>>
> >>> I'll try to respond in more detail later today.
> >>>
> >>> Best regards,
> >>> Karl
> >>>
> >>> On 7/20/10 9:05 AM, Bryan Lawrence wrote:
> >>>
> >>>
> >>>> Hi Balaji
> >>>>
> >>>> On Tuesday 20 July 2010 16:39:26 V. Balaji wrote:
> >>>>
> >>>>
> >>>>
> >>>>> Bryan, overall I agree with you both about the content of the
> tokens
> >>>>> and the machinery for delivering and using tokens.
> >>>>>
> >>>>> Small clarification, when you say
> >>>>>
> >>>>> Bryan Lawrence writes:
> >>>>>
> >>>>>
> >>>>>
> >>>>>> However, as it stands, we can't give a DOI to output which is
> not
> >>>>>> replicated, but people will need to use it. I *do* think it's
ok
> to
> >>>>>> restrict this to modellers (despite Martin's point about what
> PCMDI
> >>>>>> are advertising). I think most of the non-modelling community
> will
> >>>>>> be happy with the replicated data ...
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>> Is "modellers" supposed to mean IPCC-WG1? I think so.
> >>>>>
> >>>>>
> >>>>>
> >>>> Yes!  Sorry, lazy language.
> >>>>
> >>>> Bryan
> >>>>
> >>>>
> >>>>
> >>
> >>
> > _______________________________________________
> > GO-ESSP-TECH mailing list
> > GO-ESSP-TECH at ucar.edu
> > http://*mailman.ucar.edu/mailman/listinfo/go-essp-tech
> >
> >
> _______________________________________________
> GO-ESSP-TECH mailing list
> GO-ESSP-TECH at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/go-essp-tech
-- 
Scanned by iCritical.