Next meeting on Wdenesday February 27th

[Erik de Castro Lopo]

Hi all,

The next meeting will be help on Wednesday February 27th.

This month we have two talks:

Jed Wesley-Smith: Why "Names Don't Matter" Matters

Barry Jay: Computing with Trees not Numbers

Abstracts below.

The RSVP is here:

https://www.meetup.com/FP-Syd/events/vcqlmpyzdbkc/

Doors open at 6pm, meeting proper starts at 6:30.

Everybody who intends to show up on the night shoud RSVP ASAP to allow our
hosts Atlassian to sort out the catering. People who said they were
going but now find that they can't should likewise update their status.

Cheers,
Erik


Why "Names Don't Matter" Matters
--------------------------------
Increasingly, people are popping up stating the controversial opinion that
names don't matter. To most programmers this statement is obviously untrue,
as naming is seen as difficult, and critical to semantic expression and
communication in their code. To make matters worse, these same people
demand fidelity to obscure and difficult names from mathematics, leading to
charges of hypocrisy – if names don't matter, why do they care we all use
any particular ones?

We'll try and tease out these seeming contradictions. We'll look at why
"names don't matter" is an important idea, and why its more obvious
superficial interpretation is specious.


Computing with Trees not Numbers
--------------------------------
Tree calculus is a calculus of combinations with a single ternary operator
(for building nodes) that has three reduction rules, according to whether
its first argument is a leaf, stem or fork. Its normal forms are the
unlabeled binary trees, which provide a better basis for computation than
the natural numbers: there is no need to encode and decode programs as
numbers, as the binary trees are simultaneously functions to be applied and
structures to be traversed. There are compositional, reduction-preserving
embeddings of the traditional models of computation into tree calculus. Also,
it supports self-interpretation without the need for quotation to convert
programs to trees. Tree calculus supports a type system that satisfies
subject reduction.

--
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/