Fighting spam with pure functions

> and i don't want to be that guy,

Nah. Be that guy. They are all valid questions.

> unless the type system (not mentioned here)

I had bits in about the type system but it got cut cause, well, it was too long. It has static types, (not many, the basic types required to parse json) plus first order functions. There are no user defined types beyond ad-hoc via maps/vectors. And it has "strong" typing and does type inference.

> but surely this screams out to be implemented in lisp

One of the reasons this developed into its own language is historical. It started off very, very simple and slowly evolved into a domain-specific language. But there are a set of design requirements that complicate the situation: 1) easy to deploy changes and patch running instances (response time is critical), 2) embedding in a service, interoperatibility with the service, and so on.

> want to expose a limited interface to the users?

Yes, indeed. Part of the goal is to make the language as simple as possible for analysts to use. Obviously you need to be technical to some degree to write in a functional language, but we were trying to make it a business-logic layer on top of the infra.

You can find more details in http://research.microsoft.com/en-us/projects/ldg/a10-stein.p...

But any time we get good enough at catching an attack on a given channel, the attackers will switch to a different one - often times one where we haven't seen abuse before, and maybe don't have much good training data. In this case, it takes time to gather the right data and train new models, whereas analysts and engineers can do a reasonable job of stopping the attack in a faster timeframe.

Though interesting, all of this is somewhat orthogonal to what this article is about, though. FXL is the way we define features that are fed into classifiers, and its engine does all the data-fetching necessary to get the values of those features for classification. It also works for just putting rules on top of or next to the ML.