domingo, 6 de noviembre de 2011

The unfilled promises of declarative programming

I've just read in a paper that Constraint Logic Programming languages,
Mermaid Render - Front

because of their declarative nature, [it] is enough to state what has to be solved instead of how to solve it [...] the approach seems to be very friendly for a company real-life and day-to-day decision-making
It ringed a bell, where did I listen to such a promise? Oh, yes, some years ago I did a short Erasmus project on "Metaprogramming in Gödel", at Roskilde Universitetscenter. Afterwards, I produced a student research project named "Logic Negation and SDLFA" for Euskal Herriko Unibertsitatea/Universidad del País Vasco. We faced tough difficulties to go from what has to be solved to the actual solution to anything. As Prof. Lloyd put it, "Logic is wrong". Years afterwards, declaring "I want to migrate my ASP web to ASP.NET" doesn't produce any meaninful solution, either. We must accomplish it by hand. So my advice is, beware the promises of declarative programming, if you want to listen to them, better be bound to your ship's mast.
Udysseus (Ulysses) Mosaic From The Bardo
By the way, the paper is from a quite specific field, scheduling for project driven manufacturing. It could meet the task, not sure, but... good luck!