When I hear they're adding features to C++

When I finally grokked delimited continuations

(thanks to the example in A Monadic Framework)

When it comes to my mind that the 1990s project of moving OSes to dynlangs failed and now we're all using Unix

When I'm about to replace an important subsystem in my VM

When somebody says 'but Clojure will bring more people to Lisp'

When I see that Ruby's parser is 10,000 lines of C

When somebody tries to explain JavaScript's prototype system

When I see that a new language makes a difference between statements and expressions

When I mention first-class environments and fexprs to an academic Schemer

When I try to structure data in Scheme

When a language designer talks about their ideas of improving the rules of lexical scope

When a code base uses a directory structure deeper than one level

When I read about Perl 6 semantics

When I try to design an extension or alternative to S-expressions

When I see that the designer of a new dynlang knows the CLHS and RnRS by heart

When somebody asks me how I added a DSL to my Lisp program in 3 minutes

When somebody says he doesn't like exceptions because he likes to handle errors where they appear

(This is the first post of a new series called "This PLT Life", inspired by This Advertising Life.)

What I've been reading

If you want to impress your coworkers with terms like cobordism: Physics, Topology, Logic and Computation: A Rosetta Stone:

A simple diagram ... can now be seen as a quantum process, a tangle, a computation — or an abstract morphism in any braided monoidal category! This is just the sort of thing one would hope for in a general science of systems and processes.

An Applicative Control-Flow Graph Based on Huet’s Zipper. Somebody actually used FP in the real world -- well, to write a compiler -- and can report that it did indeed in fact work. Related, here's my favorite zipper tutorial.

Resumable exceptions can macro-express delimited dynamic variables. Because you can never test your intuition enough.

More links: Ian Piumarta's liboop; Eric Torreborre on total FP; VeriML; automatic scoping and destructor functionality in C.

HT Paul Snively, Daniel Yokomizo, Patrick Thomson.

PoshLusT

PLT is a resentful business - there are so many accolades for blowhards, it can be hard to stomach.

Thankfully, the writer Vladimir Nabokov offers a remedy: the word poshlust:

Let me refer to one more method of dealing with literature- and this is the simplest and perhaps most important one. If you hate a book, you may still derive artistic delight from imagining other and better ways of looking at things, or, what is the same, expressing things, than the author you hate does.The mediocre, the false, the poshlust-remember that word-can at least afford a mischievous but very healthy pleasure, as you stamp and groan through a second-rate book which has been awarded a prize.

He also has advice for dealing with the good stuff:

But the books you like must be read with shudders and gasps. Let me submit the following practical suggestion. Literature, real literature, must not be gulped like some potion which may be good for the heart or good for the brain-the brain, that stomach of the soul. Literature must be taken and broken to bits, pulled apart, squashed-then its lovely reek will be smelt in the hollow of the palm, it will be munched and rolled upon the tongue with relish; then, only then, its rare flavour will be appreciated at its true worth and the broken and crushed parts will again come together in your mind and disclose the beauty of a unity to which you have contributed something of your own blood.

What PLT made you shudder and gasp? For me, the first things that come to mind are Kernel, the MOP, Ωmega, and A monadic framework for delimited continuations.

[Meta] Comments widget

I've added a recent comments widget to the sidebar of the blog.

There have been many excellent comments recently, and AoE completists may want to check them out.

[Unfortunately, the links don't take you to the comments themselves but rather to the page. Investigating.]

Manuel, have you lost interest in Kernel?

... I was asked yesterday two times on #interactiveprogramming.

No, absolutely not!

I simply got stopped in the tracks trying to find a satisfactory way of adding named arguments to it. I need that feature to consider it complete.

However, two new old topics will probably be raging in my skull in the future:

• Static types, because defining red-black trees that are balanced by construction makes so much sense (and it may finally help me understand those pesky trees).
• Interactive execution, because Roly Perera.

#interactiveprogramming IRC

For the discussion of interactive programming languages, I'm now hanging out on #interactiveprogramming on chat.freenode.net IRC.

You're all happily invited to join!

Back to the future

Roly Perera, who's doing awesome work on interactive programming is apparently taking up blogging again. Yay!

Back to the future

I know not of these “end users” of whom you speak. There is only programming. – Unknown programmer of the future 

Jonathan Edwards recently wrote that an IDE is not enough. To break out of our current commitment to text files and text editors requires a new programming paradigm, not just a fancy IDE. I feel the same. I’ve spent the last two and a half years moving my interactive programming work forward, and I feel I’m now ready to start sharing. Bret Victor blew us away with his influential, Jonathan-Edwards-inspired, mindmelt of a demo but was conspicuously quiet on how to make that kind of stuff work. (Running and re-running fragments of imperative code that spit out effects against a shared state, for example, does not a paradigm make, even if it can be made to work most of the time. “Most of time” is not good enough.) I won’t be blowing anyone’s minds with fancy demos, unfortunately, but over the next few months I’ll be saying something about how we can make some of this stuff work. I’ve been testing out the ideas with a proof-of-concept system called LambdaCalc, implemented in Haskell.