this is putting up some really good examples of where we have to be
expeedient to get the job done. "First release" is fine as it never
changes. "Most recent" opens up some issues, but we should be able to
cope with that at update time with some appropriate automation.
I was thinking of the use of dates too. Time since last release would
have to be done outside the ontology, but querying the ontology. Of
course, it will only tell us time since last recorded update in SWO.....
At 17:06 10/06/2011, Andy Brown wrote:
>Very interesting. Looking at the competency questions that came up at the
>workshop, I don't think we need to worry too much. No-one actually asked 'is
>this mature software?', so we can provide concrete information and allow them
>to compare. Support was a big topic, and answerable with pointers, and
>questions about how actively software is being developed can be inferred (to
>an extent) if we include release date information for versions. I'll put
>together a list of questions/software parameters that are straightforward and
>concrete and implement them.
>On Fri, 10 Jun 2011 13:30:10 +0100
>James Malone <[log in to unmask]> wrote:
> > Hi Andy,
> > Maturity is evil. Not least because there are several formal models
> > (standards even Robert) that exist which have various pointers to
> > processes that must be followed and repeated to be considered mature.
> > And of which some are easily quantifiable and some not so, such as this
> > from level 4 of CMM: " In particular, management can identify ways to
> > adjust and adapt the process to particular projects without measurable
> > losses of quality or deviations from specifications."
> > So I think you are right to focus on the stuff we can quantify and
> > represent. I wonder if there is a way of formulating already the query
> > for first and last versions of software, something like:
> > first version of software is software which has a version which has no
> > preceeeding versions
> > similarly, last version of software is software which has a version
> > which has no following versions.
> > This may be possible thought "not"s are hard.
> > I suppose you could potentially link the version individual with some
> > property of "first release" so that we always know which version is
> > first. With latest this could also be done but there is a small
> > maintenance issue which could potentially be done by the inferencing I'm
> > suggesting above. Just had a quick play with some restrictions but it
> > won't play yet, for example this does get you everything that is not
> > followed by version Microsoft 2003 which you might expect:
> > software
> > and ('has version' some
> > ('version name or number'
> > and not ('followed by' value 'Microsoft 2003 version')))
> > Largely because the reasoning over individuals is different. This may
> > not work after all. need to play some more.
> > James
> > On 10/06/2011 10:43, Andy Brown wrote:
> > > Hi,
> > >
> > > Looking at the competency questions collected from our workshop
> that related to
> > > life cycle, there seem to be several bits of information that
> people would
> > > like that relate to maturity. As well as any official statements about
> > > maturity (such as a beta tag), we can make judgements about the
> status of some
> > > software by looking at these bits of information: software which has been
> > > around for some time, with a recent release, active support
> forums, etc. might
> > > be considered 'mature', while software labelled 'beta', with a
> version number
> > > <1, and only 2 releases might be considered 'immature'. I've
> come up with the
> > > following statements that could be made about software, that
> could be combined
> > > to deduce its state of maturity and, as usual, questions arise
> about how we
> > > handle versions!
> > >
> > > - first_released (point to the version class, or just date)
> > > - latest_release (point to version class, or just a date? This will need
> > > maintenance)
> > >
> > > My inclination at the moment, is to just have first_release as
> a data property
> > > pointing to a date, but have latest_release pointing to the
> class for that
> > > version, in which case we also want release_date as a property
> of software.
> > >
> > > Other statements we can make are the following properties:
> > >
> > > - has_declared_status (alpha, beta, obsolete)
> > > - has_disucssion_forum_url (point to forum; do we want to
> distinguish between
> > > official support and other forums?)
> > >
> > > Do we want to think about a metric for giving the maturity of a piece of
> > > software, or let people make up their own mind based on the
> data given? I
> > > guess the latter.
> > >
> > > Any thoughts before I implement these?
> > >
> > > Cheers,
> > >
> > > Andy.
> > --
> > European Bioinformatics Institute,
> > Wellcome Trust Genome Campus,
> > Cambridge, CB10 1SD,
> > United Kingdom
> > Tel: + 44 (0) 1223 494 676
> > Fax: + 44 (0) 1223 494 468
Reader in BioHealth Informatics
School of Computer Science
The University of Manchester
+44 161 275 6251
[log in to unmask]