Re: Next Steps (was: Re: LOD Cloud Diagram Update)

I wonder - would it be at all possible, or has it even been tried, to have
a public domain web crawler that updates an index of LOD datasets? Only
paying attention to pages/responses that contain instances of void:Dataset
for example?

Regards,
Frans

2018-04-30 15:05 GMT+02:00 Sarven Capadisli <info@csarven.ca>:

> On 2018-04-30 13:20, John McCrae wrote:
>
> > We have released a new update of the LOD Cloud Diagram at:
> >
> > http://lod-cloud.net/
>
>
> IMHO:
>
> I propose that the generation of the visualisation minimises human
> involvement - from the point of letting the index know about what's out
> there to actually verifying, and then generating the visualisation.
>
> One way to do that can be along these lines:
>
> https://linkedresearch.org/cloud
>
> In a nutshell:
>
> 1. Send a Linked Data Notification ( https://www.w3.org/TR/ldn/ ) to the
> cloud's inbox(es) indicating what, where..
>
> 2. Consume the notifications in the inbox and auto-generate the
> visualisation (as well as for other reasons for anyone to process the
> notifications).
>
> Many things to improve but the main architecture is there staring at us.
>
> But hang on, we're not done! There is a whole set of R&D challenges
> waiting for this community to pick up on:
>
> * Shape of the notification (what should it say?)
> * Filtering.. (eg. spam prevention)
> * Authentication/Authorization (consider possible restrictions..)
> * Inbox management (how many inboxes, how big.. paging?..)
> * Findability/Searching..
> * Offer "cool" interactive virtualisations
> * Creative reuse, analysis..
>
> I suggest that addressing the above should be community driven. Maybe
> even have open challenges/articles/demos at (Semantic) Web conferences ;)
>
>
> -Sarven
> http://csarven.ca/#i
>
>