Re:Is Datalog a good language for authorization?
by flurie on Saturday February 19, 2022 @08:02PM
(#30402178)
Sure it is, because the language of Open Policy Agent, called Rego, is a Datalog variant. I find a researcher’s failure to mention this concerning.
[ Parent
| on HN ]
Re:Is Datalog a good language for authorization?
by Nullabillity on Saturday February 19, 2022 @08:17PM
(#30402274)
The second sentence of the post links to OPA as an example of a system using Datalog for authz.
Also, "X uses Y for Z" doesn't imply that Y is good at Z. It can be the start of an argument about how that choice is turning out for X, but it's in no way an argument by itself.
Re:Is Datalog a good language for authorization?
by flurie on Saturday February 19, 2022 @09:38PM
(#30402726)
My assertion is glib, but your response is wholly insubstantive. Can you to point out how this piece treats Rego as an actual example of a real-world Datalog? I don't see any link when mentioning real-world Datalogs.
The fact that this piece makes no mention of it but provides a drive-by link to it suggests that the author is not actually interested in evaluating the problem in a realistic scenario. The piece the literary equivalent of a reminder that one can harm oneself with a gun without ever actually discussing how one should implement gun safety.
Re:Is Datalog a good language for authorization?
by 0x0000000 on Saturday February 19, 2022 @08:16PM
(#30402273)
The second link in the article is https://www.openpolicyagent.org/docs/latest/policy-language/