See also: IRC log
<bhyland> http://www.w3.org/2011/gld/meeting/2012-08-30
http://www.w3.org/2011/gld/meeting/2012-08-30 -> Minutes of last week
bhyland: Last week we talked about what we need to do to improve our messaging
<BenediktKaempgen> +1
<gatemezi> +1
bhyland: Do we have approval for last week's minutes?
<bhyland> Thanks. Minutes for last week approved.
Resolution: Minutes approved
bhyland: We haven't revisited this for quite a while (January :-( )
<bhyland> http://dir.w3.org
bhyland: Little has happened in
terms of evolution of it
... It was updated to the latest version of Callimachus
<bhyland> http://dir.w3.org/rdf/2012/directory/categories-list.xhtml?view
bhyland: If you go to the
directory statistics http://dir.w3.org/rdf/2012/directory/statistics.xhtml?view
... we have about 70 organisations, for example. Getting a
member request about once a fortnight
... It's perhaps looking out of date in some places. Usability
is not as good as we'd like
... So I want to get some feedback and get some volunteers to
join a meeting to flesh out some requirements to make it more
compelling etc.
... I asked folk to come prepared today with the URL of a
directory that was useful and usable
<bhyland> http://www.w3.org/2011/gld/wiki/Community_Directory_Requirements
bhyland: Any comments?
... Appeals for honesty...
<bhyland> PhilA: The Crossover Directory is different but has some of the same issues, namely lack of current content.
<bhyland> … PhilA suggested some of the human aspects of directories ...
<bhyland> … common problem is people working in isolation, collectively not achieving the desired outcome.
bhyland: The one I think of is
Slideshare. I don't like the proprietary, locked down aspect
but it's mighty useful
... Getting Linked Data info into people's hands is important,
no?
... What would make a difference to DERI, for example?
cygri: Hard to say... one thing
that strikes me as important is ability to showcase projects
and products that we have done
... We want to say more than "hello, we do LD research" - we
want to say we have this OS software, as have produced X
etc.
... So that level of detail would help
bhyland: Hadley - in organising events etc. Would it be helpful to have a list of organisations who are doing work in this field?
<HadleyBeeman> hi
<HadleyBeeman> sorry
<HadleyBeeman> As with the community directory, if we take on the challenge of creating and maintaining a list of events, it's a resource commitment.
<George> http://opencorporates.com/
<George> "We have information on 44,411,651 companies"
bhyland: Basically, how can we make it useful for the thing we want to promote
<HadleyBeeman> Also, there are others chasing lists of events of open data, linked data, government data… I'd have a look at lanyrd, data.gov.uk, the OKFN, etc.
George: Raises us of business vocabulary by Open Corporates and its 43 million entries etc.
bhyland: You said it's more about the fact that you can list the info than having our own directory?
George: In the spirit of LD we want it to be easy to incorporate data etc.
bhyland: You say that if we
produce a standalone data we should be able to include data
from Open Corporates etc. - so we can take data in and
out
... So the data is two way
... But the other piece is the usefulness of what we're
capturing - if we agree we need it then we need an effort to
make the site usable (incl. baic things like specifying a
country)
<bhyland> PhilA: We touched on this in June 2012, the Comm Directory could act as a showcase of good linked open data.
<bhyland> … for example, the bathing (swimming for us yanks) water visualization was wonderful.
<bhyland> … Consider focus on visualizations.
<bhyland> … Data portals, like data.gov.uk, will include sections on apps that are on that specific data portal site, e.g., "latest", "good visualizations".
<bhyland> … We need apps that will make policy makers stand up and say, 'that is cool, I want that for my agency!'
<bhyland> … Add some info about who created the app, etc.
bhyland: Any more comments?
cygri: I guess what we're finding
is that we're relying on people entering data about themselves
and their projects
... So the data ends up being variable in quality to put it
mildly
<bhyland> cygri: If we rely on people entering info themselves, the info is both variable in quality and potentially uncompelling.
<bhyland> … Do we need a curator?
cygri: So do we need someone to
be a more active curator?
... It's difficult to get to a good level of collection of data
without some level of curation
... Someone to highlight certain things, important aspects
etc.
... This raises the question of who, of course...
... At the moement, it's not going to make or break someone
whetehr they are or are not in the directlry
<bhyland> … it isn't going to 'make or break anyone's business if they are in the directory today ...'
cygri: So few people will care
about it enough to give it sime
... If the info there isn't complete then there's no incentive
for maintaining it
bhyland: I agree with all of your
points
... I'd like it to be on people's radar
... Make it somewhere for people to look for help making the
business case or speakers etc.
... A pharma co. has anounced it's putting money into orgs like
DERI rather than doing it themselves. So having a go to place
with the weight of the w3.org namespace can be compelling
... the variability of the daat doesn't help
<HadleyBeeman> I know I'm not fully in the conversation today, but I think we should take care around the line between building the tools and standards to make government linked data useful and usable — and protecting individual companies, business models, marketing, etc.
<cygri> biplav: people will be interested in datasets that are avaialbe; value that it can provide; and who are people i can contact for my problem
<bhyland> Biplav: Wants directory to list: What are the datasets, what has been the impact, what have been the benefits? Better visualization is compelling.
<cygri> biplav: concrete measurements, dollar impact, to grab people's attention
bhyland: You're saying things
like concrete measurements etc. is what anyone would need to
make a decision on spending money
... we need to move away from research and more into enterprise
way of thinking
<gatemezi> How could you see the benefits of datasets different from that we have for e.g. here in http://thedatahub.org/
<bhyland> +1 to Biplav
Biplav: We have a transformation department. If I tell them they should use LD, they'd say why? They need a quantification of what was possible before and what would be possible afterweards
<Zakim> sandro, you wanted to muse about selling dogfood
<Biplav> transformation => transportation
sandro: I was concerned about the
directory - us being caught up in a dog food project. But...
maybe we could focus on the data and not the UI quality?
... Maybe something like an ISWC challenge? Ask people with
relevant expertise so we'd be the showcase for their
work?
... so we'd be a feed of data about open data projects
... I'm scared about diverting energy but it might work?
bhyland: Interesting
George: The challenges I've been
involved with have been regulated by the America Competes
Act
... need to reach out to the communities that would be
interested
... lots of otehr less-regulatory driven challenges and then
can be pretty high touch. Someone has been heavily
involved
... i.e. it takes person power
<Biplav> What we are discussing is a good topic for a workshop at a conference like ISWC, I think.
George: Being able to repurpose the data is good of course. Maybe we could set up a SPARQL Push ??
<bhyland> No, not a workshop, a potential challenge targeted at the right group of developers once we are further along ...
<Zakim> cygri, you wanted to suggest focusing on a small area and aim for completeness + high quality
PhilA: Sounds like a sponsorship/funded project oppoertunity??
cygri: Repurposing hte data is interesting but it only becomes interesting if the data is good and interesting. And that's the challenge. Worrying too much about the tech side could be a bit of a distraction if we don't have that quality
<BenediktKaempgen> +1 to having a challenge to find interesting content for community directory
<George> PuSH - DERI did some work that does a SPARQL query on a PubSubHubbub (PuSH) feed hub
<George> so
cygri: So I'm a little worried about going in that direction
<bhyland> +! to what cygri is saying that the data needs to be more complete & is a valuable resource that people *want* to combine for some interesting insights
cygri: One idea for how we may
improve the data quality might be to limit the scope to a
particular topic
... The 'open gov data' space is very big. Maybe we can focus
on a domain?
... Try to establish completeness and quality?
... and then expand, based on that success?
... No idea what that topic could be
PhilA: Agriculture?
sandro: I'd say that if even one of us were motivated to spend ~ 4 hours per week on digging up info - we could have a pretty high quality data set. It seems to me that Gov LD is only a couple of hundred people
bhyland: If we just limited it to 4 hours then that sounds like an excellent job for a new hire/intern
<scribe> ACTION: bhyland to get a Computer Science Intern to spend ca. 4 hours a week on gathering data for the directory [recorded in http://www.w3.org/2012/09/06-gld-minutes.html#action01]
<trackbot> Created ACTION-75 - Get a Computer Science Intern to spend ca. 4 hours a week on gathering data for the directory [on Bernadette Hyland - due 2012-00-13].
<sandro> bhyland: I'll see if I can find an intern to do this kind of work, eg 4hrs/wk
<sandro> bhyland: At the end of the day, the technology doesn't matter very much. If we produce turtle & human-readable-stuff, that's enough.
<sandro> bhyland: The point is it's a data-driven app, and some better tool comes along, they can slurp up the RDF.
bhyland: It shouldn't be about proving RDF, that's a given...
<sandro> bhyland: It's a given that we produce human + machine-readable output from this.
Biplav: If you could share what you are thinking of the Intern doing I might be able to help with resources :-)
bhyland: I'll get a description
and if you can help find an alternative or additional person
that wouldbe great
... no modelling or vocab development etc.
<bhyland> @PhilA, actually there is some work to be done by the editors …
<bhyland> Agreed, we all want the vocabs to move to LC by TPAC (late Oct 2012(
<bhyland> thanks PhilA and DaveReynolds for concise action request.
DaveReynolds: Comfortable with the schemas going in place. Looks to me as if, modulo my comments, the spec docs look good for FPWD
Proposed Resolution: RDF schemas fopr ADMS and Buisiness Vocab move to /ns space. *ADMS is an update)
<DaveReynolds> +1
<cygri> +1
<cgueret> +1
<bhyland> +1
<BenediktKaempgen> +1
<gatemezi> +1
<bhyland> Love this bias towards action!
<George> +1
<BenediktKaempgen> Announcement: WS related to QB work (2nd Workshop on Semantic Statistics for Social, Behavioural, and Economic Sciences: Leveraging the DDI Model for the Linked Data Web ) - http://www.gesis.org/en/events/gesis-workshops/ddi/
PhilA: Thank you DaveReynolds for your helpful comments
<bhyland> thanks everyone, very productive discussion. Appreciate it.
bhyland: If you had a directory URL to point to anotehr directory, please add it :-)