What is Pure Computation

In this post, I'll respond to the comments and feedback to my recent blog on 2 misconcepts about functional programming.

I write the original blog because I found it funny that languages like XQuery and XSLT, which I feel so 'functional', is not actually referentially transparent. Yet monadic I/O as implemented in Haskell, which I feel so 'procedural' is referentially transparent. So my intension in the blog was to come up with something that can better depict the computation model of XQuery and XSLT. My approach was to keep the concept of RT, but stretch the concept of functional programming. From the feedback, I can see that people think that's not acceptable.

Candle's Procedural Design

In the previous blog, I've shared my thoughts about functional programming. Sometimes it looks like to me that procedural programmers are side-effect addicted, whereas functional programmers tend to have side-effect phobia. Hard-core functional programmers tend to spread the feelings that side-effects are evil. Yes, if the side-effects are unnecessary, then they should be eliminated by all means. But if the side-effects are desirable (by requirements), then there's nothing evil about them. There's no impurity in a program if it allows desirable side-effects.

Everyone agrees that side-effects should be controlled and carried out in a structured manner. The question is how. I think the Holy Grail of current programming language design is to find the best way to mix procedural programming with functional programming. In the following blog, I'm going to talk about the procedural design in Candle, and how it compares to similar designs in languages, like Haskell, Erlang and Clojure.

2 Misconcepts About Functional Programming (relating to context and monad)

It's interesting that almost every old and new language that claims to have some functional support has to go through the trial of its purity, like Perl, Scala, Erlang, Ruby, JavaScript and Python. Candle is no exception. Candle's last release announcement to LtU triggered some interesting discussion about whether the functional features in Candle are pure. I finally have sometime to share more about my thoughts on this issue. I'd like to talk about 2 misconcepts about functional programming. The first misconcept is that dynamic scope variables are not functional. The second misconcept is that monads are functional. I know the following writing is going to kick up the dust, but it is better than we settle without thorough thinking.