W3C home > Mailing lists > Public > public-dxwg-wg@w3.org > September 2017

Re: Stating requirements

From: Rob Atkinson <rob@metalinkage.com.au>
Date: Thu, 14 Sep 2017 00:39:19 +0000
Message-ID: <CACfF9Lx6xf6OdPk3NCsfe9LMgnW_brVAAw5c9ichn3cbF_3zdA@mail.gmail.com>
To: Jaroslav Pullmann <jaroslav.pullmann@fit.fraunhofer.de>, Rob Atkinson <rob@metalinkage.com.au>
Cc: kcoyle@kcoyle.net, public-dxwg-wg@w3.org
Thanks Jaro

agree with concerns re "Finer grained descriptions of datasets and
distributions" but feel that "semantics" is even more vague - its all
semantics...

maybe "describing content" - it is all about describing actual content (of
both datasets and distributions) - not its provenance or related resources.

other groupings perhaps:
"provenance"
"references"
"related data"
"catalog control"

Rob


On Thu, 14 Sep 2017 at 09:11 Jaroslav Pullmann <
jaroslav.pullmann@fit.fraunhofer.de> wrote:

>
>  Hello Karen, hello Rob,
>
>   I suggest to treat these grouping-tags like any other. Later on we might
> think of
>  a better organization of the tag switches. From the previous postings I
> gathered the tags :
>
>     - "Profiles": a tag "profile" is already present (identifies the AP
> deliverable and "anything profile-related")
>
>     - "Versions", to be added
>
>  I have some problems with "Finer grained descriptions of datasets and
> distributions"
>  since this is really vague, and 6.12 is hard to grasp/concretize. Maybe a
> sign of weak
>  requirement quality. There is a separate tag for "semantics", both seems
> ok for 6.10, 6.11?
>
>  There was the tag "referencing" intended for identification and citation
> purposes, does it fit?
>
> > Whatever happens, its going to be hard to come up with disjoint
> categories and groupings
>
>  The tags allow for multi-aspect groupings and do not force us to find a
> single hierarchy.
>  I assume that some requirements naturally belong to multiple categories.
>
>  Currently the requirements share tagging of their UCs. In order to make
> the filtering more precise I'd
>  suggest to leave the tagging of UCs quite general and add more specific,
> distinguishing tags to the
>  requirements themselves?
>
> > move requirements up to top of document
> > have the list of UC collapsed by default in the overview
>
>   fine! Let's experiment, both alternatives make sense to me
>
>   Best regards
>     Jaroslav
>
> On Wednesday, September 13, 2017 23:39 CEST, Rob Atkinson <
> rob@metalinkage.com.au> wrote:
>
> > Jaro - do we just copy the tag mechanisms and make up new tags for
> > requirements groups, re-use existing tags or create a new scoped set of
> > tags for requirements and a separate file control.  It would also be nice
> > to have the list of UC collapsed by default in the overview - or move
> > requirements up to top of document?  IMHO the simple clean requirements
> > will be the point of entry, and where we need to work, and the UC almost
> an
> > appendix with further evidence as required.,
> >
> > Rob
> >
> > On Thu, 14 Sep 2017 at 00:18 Karen Coyle <kcoyle@kcoyle.net> wrote:
> >
> > > Thanks, Rob. This gives us four groups that we can discuss.
> > > Can we give the requirements in these groups a unique tag that we can
> > > filter on? That way we all get the same view while discussing them.
> > >
> > > kc
> > >
> > > On 9/12/17 3:08 AM, Rob Atkinson wrote:
> > > > Hi Karen et al,
> > > >
> > > > The first batch of requirements have been edited heavily, the second
> 50%
> > > > less so - was trying to get feedback on progress so far because its a
> > > > fair bit of work to try to get the requirements both simple and self
> > > > explanatory, and not be lots of repetitions of very similar things.
> > > >
> > > > I have just committed a few edits I had started cleaning up such
> things.
> > > >
> > > > For the record, the groupings to review are:
> > > >
> > > > 6.1 -> 6.4  "Profiles"
> > > > 6.5 -> 6.9 "Versions"  (add 6.43 here)
> > > > 6.10 -> 6.13 "Finer grained descriptions of datasets and
> distributions"
> > > >
> > > > I would like to focus on identification and citation matters as a
> group
> > > > - it kind of overlaps Version perhaps
> > > > 6.17 6.22 6.23 6.36?
> > > >
> > > > Are spatial and temporal extent part of "fine grained semantics" - in
> > > > the same way that classifications schemes used in attributes are
> part of
> > > > the fine grained data structure - is this just an expression of the
> > > > range of a property of the data?
> > > >
> > > > Whatever happens, its going to be hard to come up with disjoint
> > > > categories and groupings - and I dont want to spend my life
> justifying
> > > > my arbitrary decisions, so I've done a few and lets see whether an
> > > > agreed structure falls out looking at the rest please.
> > > >
> > > > Rob
> > > >
> > > > On Tue, 12 Sep 2017 at 12:06 Karen Coyle <kcoyle@kcoyle.net
> > > > <mailto:kcoyle@kcoyle.net>> wrote:
> > > >
> > > >     I'm sorry I missed the last meeting, so I might be repeating
>
> > > something
> > > >     that was already said, but... I think it would be helpful if the
> > > >     requirement "headings" were stated as requirements. That way we
> could
> > > >     look at the list of requirements and it would make sense. As an
> > > example,
> > > >     we have:
> > > >
> > > >     ----
> > > >     6.17 Cite datasets
> > > >
> > > >     Provide a way to specify information required for data citation
> > > (e.g.,
> > > >     dataset authors, title, publication year, publisher, persistent
> > > >     identifier)
> > > >     ----
> > > >
> > > >     I would modify this to be something like:
> > > >
> > > >     ----
> > > >     6.17 Provide full citation information for datasets
> > > >
> > > >     Currently missing from DCAT are:
> > > >      - full range of identifiers,
> > > >      - dates,
> > > >      - contributors and
> > > >      - resources supported by [DataCite]
> > > >     ----
> > > >
> > > >     (I copied from the use case - that list of missing may not be
> > > correct.
> > > >     This is just an example.)
> > > >
> > > >     Some requirements are already worded this way, like:
> > > >
> > > >     6.3 Create a way to list the profiles implemented by a dataset
> or a
> > > >     specific distribution
> > > >
> > > >     If this makes sense, I may be able to make a number of
> suggestions
> > > >     before the next meeting.
> > > >     --
> > > >     Karen Coyle
> > > >     kcoyle@kcoyle.net <mailto:kcoyle@kcoyle.net> http://kcoyle.net
> > > >     m: 1-510-435-8234 (Signal)
> > > >     skype: kcoylenet/+1-510-984-3600 <+1%20510-984-3600>
> <+1%20510-984-3600>
> > > <tel:+1%20510-984-3600>
> > > >
> > >
> > > --
> > > Karen Coyle
> > > kcoyle@kcoyle.net http://kcoyle.net
> > > m: 1-510-435-8234 (Signal)
> > > skype: kcoylenet/+1-510-984-3600 <+1%20510-984-3600>
> <+1%20510-984-3600>
> > >
> > >
>
>
>
> --
> Jaroslav Pullmann
> Fraunhofer Institute for Applied Information Technology FIT
> User-Centered Ubiquitous Computing
> Schloss Birlinghoven | D-53757 Sankt Augustin | Germany
> Phone: +49-2241-143620 <+49%202241%20143620> | Fax: +49-2241-142146
> <+49%202241%20142146>
>
>
>
Received on Thursday, 14 September 2017 00:40:11 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 October 2019 00:15:38 UTC