- From: Andy Seaborne <andy.seaborne@talis.com>
- Date: Sat, 10 Apr 2010 18:33:27 +0100
- To: Stephane Corlosquet <scorlosquet@gmail.com>
- CC: semantic-web@w3.org
> I guess as long as nobody's complaining we can just keep using these > public endpoints... :) Sure :-) I try to keep sparql.org up but it's on a best effort basis. It does get trashed sometimes - the logs suggest errant programs rather than anything malicious. Occasionally what looks like robots seem to GET the query page which must be quite boring for them. There is a limit on the size of graphs it will load from the web just to stop accidentally trying to FROM <dbpedia>. Queries are logged, but not who they are from the logs not kept for very long. The most interesting things are the errors, for example, attempting to load HTML pages, or file: URLs (it blocks access to the local disk :-), and especially failing to rad the whole response. It would be good if more general SPARQL processors were available to spread the risk of any one of them being down. It's running Joseki fronted by Apache and has been excellent for hardening the code for HTTP handling. There is also a query parser/validator. Andy
Received on Saturday, 10 April 2010 17:34:07 UTC