Hi,
I agree with Yves. Of course there is a trade-off between lowering the entry
cost for consumers and lowering it for publishers.
If you have a Web site around your data already and just want to export it as
LD as well then you probably already have search functionality and it should be
easy to integrate it with the LD export.
However if you are just about to publish data that hadn't been published before
then, to say it with TimBL's words, GIVE US THE RAW DATA! :-)
Shiny Web pages around it? Later!
Search? Later!
SPARQL endpoint? Later!
Sitemap? ...
...
Of course each of these steps makes the data so much more useful and makes it
so much easier to start using it. However we can't require everything at once
or no-one will do it. Instead we should get people enthusiastic about the
*idea* of LD so that they start publishing their data soon. Every additional
bit of data is useful and we should welcome it all instead of picking on the
publishers for details.
Then, when this first step is done, we should educate them about what else they can do to
make their data more discoverable and useful ("Hey, if you also put the data as RDFa
in your pages, people can copy it to their clipboard and re-use it elsewhere!").
This step is very important as well but it shouldn't be a requirement for publishers
because otherwise they won't be accepted in the higher planes of the ultimate perfect LOD
cloud or something. We can't afford that luxury.
Regards,
Simon
