Wednesday, July 04, 2007
Why isn't the open source community embracing declarative programming?
Declarative programming is where problems are described, or conditions on a solution are described, and the computer finds a solution. Often it involves the separation of "facts" from operations on the facts...
Too many folks are focused on enterprise worthy languages such as Java and/or Ruby with those in the lunatic fringe focusing on languages such as Smalltalk who don't have built-in support for CardSpace, OpenID, XACML, SPNEGO, etc while not focusing on the fact that the better answer is that folks shouldn't be writing code and instead should be thinking of better ways of describing the structure of the problems they face and feeding this information to an engine for consumption.
The blogosphere lacks a conversation around declarative programming with rare conversations on this topic by James Taylor, James Tarbell and Scott Mark. Likewise, I haven't ran across any industry analysts discussing the promise that business rules engines from Fair Isaac, iLog and others bring to the table. Sadly, this technology hasn't been widely adopted and therefore analysts who are historians in disguise don't provide much coverage when it reality, this technology could be liberating.
In the security space, I think that XACML meets the declarative spirit of security, by allowing folks to declare policy without having to worry about how it will be enforced. BPM also seems to have some flavor of declarativeness yet they still need considerable work in terms of maturing BPMN.
Todd Biske asked whether Enterprise Architecture should be declarative or imperative, why can't we ask the same questions about the languages/approaches used within the enterprise?
| | View blog reactionsToo many folks are focused on enterprise worthy languages such as Java and/or Ruby with those in the lunatic fringe focusing on languages such as Smalltalk who don't have built-in support for CardSpace, OpenID, XACML, SPNEGO, etc while not focusing on the fact that the better answer is that folks shouldn't be writing code and instead should be thinking of better ways of describing the structure of the problems they face and feeding this information to an engine for consumption.
The blogosphere lacks a conversation around declarative programming with rare conversations on this topic by James Taylor, James Tarbell and Scott Mark. Likewise, I haven't ran across any industry analysts discussing the promise that business rules engines from Fair Isaac, iLog and others bring to the table. Sadly, this technology hasn't been widely adopted and therefore analysts who are historians in disguise don't provide much coverage when it reality, this technology could be liberating.
In the security space, I think that XACML meets the declarative spirit of security, by allowing folks to declare policy without having to worry about how it will be enforced. BPM also seems to have some flavor of declarativeness yet they still need considerable work in terms of maturing BPMN.
Todd Biske asked whether Enterprise Architecture should be declarative or imperative, why can't we ask the same questions about the languages/approaches used within the enterprise?
