Lambda the Ultimate

Kendall Clark of the Data Access Working Group (part of the W3C's Semantic Web initiative) has posted regarding their work on a query language and access protocol for RDF. The DAWG has recently released the 2nd draft of its Use Cases and Requirements and is looking for community input. Their initial query language design, BRQL, looks SQL-like but is designed to operate on graphs/sets of triples.

I know that this was discussed on LtU already, but as "Why Types" discussion demonstrated, people can always come with new ideas on old topics.

From theoretical point of view, call-by-value and call-by-name are dual.

Does it mean that in practice PLs should support just one of them, and get "two-for-the-price-of-one"?

Or does usability demand supporting them both?

Is this decision similar to supporting just And-Not as opposed to full set of (redundant) logical operations?

Does the decision depend on type system, or is fully orthogonal to it?

We extend the "scrap your boilerplate" style of generic programming in Haskell to accomplish an additional range of applications. This includes several forms of serialisation and de-serialisation, test-set generation, type validation, and type erasure. To this end, we provide a well-designed reflection API for datatypes and constructors, and we also provide more general means of extending generic functions for given monomorphic or polymorphic types. The presented approach is readily supported in the GHC implementation of Haskell.

The previous "boilerplate" paper was discussed here in the past.

This is a interesting paper and there are many reasons why I should link to it, but I'll let you guess the number 1 reason (hint: check section 10).