Re: How to publish SPARQL endpoint limits/metadata?

Hello Ghislain,

On 8-10-2013 12:16, Ghislain Atemezing wrote:
> Hi Frans,
>
>> I am experimenting with running SPARQL endpoints and I notice the 
>> need to impose some limits to prevent overloading/abuse. The easiest 
>> and I believe fairly common way to do that is to LIMIT the number of 
>> results that the endpoint will return for a single query.
>
> What about providing a dump page where you put a VoID or DCAT 
> description of the ALL dataset… (here describing the actual size)
> And for the SPARQL endpoint , just upload a *SMALL* size of it (with a 
> LIMIT configuration here )

I am not sure if I understand what you wrote. Are you saying that I 
should not make all data available through the SPARQL endpoint?

>>
>> I now wonder how I can publish the fact that my SPARQL endpoint has a 
>> LIMIT and that is has a certain value.
>
> I guess it will all depends on your endpoint backend (fuseki, 
> virtuoso, OWLIM ,etc). Some comes with configuration files where you 
> can set the limit of the maximum results.
My question was not about how to configure limitations on the endpoint 
backend, but how to make those limitations known to users of the backend.

Perhaps I need to elaborate on what I am thinking about... Let's say a 
user agent that is browsing or crawling the web comes across a dataset 
that is somehow interesting. Using the VoID metadata the agent can 
discover the URI of the SPARQL endpoint. With that knowledge, the agent 
can start querying the dataset for those data that it is interested in. 
Now it would be very nice to know the service limitations before the 
agent starts querying. Perhaps the service has some kind of access 
control, perhaps it is scheduled to be offline at the moment, perhaps it 
can return no more than 100 results per request... If limitations like 
these are known to the agent, the changes of successful communication 
will be much increased. And perhaps it will take some weight of the 
server too.

Regards,
Frans