miniblog.

I have tried, but failed to persuade the denizens of Stack Overflow to add highlighting to #Emacs questions. http://t.co/rOmXAdEbgR

Memacs: An implementation inspired by the Memex concept, using #Emacs and org-mode:

I've written a Python program that *requires* TCO. I'm going to have to rewrite in continuation passing style. Argh.

Describing a program as 'just a very large number written with 0s and 1s' is rather like describing a book as a very long word.

Dark, unsound corners of the Haskell type system: http://t.co/9JrcMtQ4co

I was rather surprised to learn that some lisps have separate types for identifiers and symbols. The difference is subtle.

Improving an average project is a matter of learning from the alternatives. To push the state of the art you need to consider fundamentals.

With a little care and experimenting with re-builder, you can use query-replace-regexp to do syntax transformations on old code. Wonderful.

If you're using expand-region in #emacs, you probably want to require the mode-specific expansions too. E.g. `(require js-mode-expansions)`

'Why I'm Ditching CoffeeScript': http://t.co/uvhisjNrA5 Interesting discussion of array length caching gotchas in CS.

I am regularly impressed with range of dialects available in Racket. For example, strongly typed without parens:

Dired tip: rather than pressing RET on a directory, press `i` instead. This will show the directory contents in the same buffer.

Seeing people tweet screenshots of paragraphs of text. It's a workaround for the 140 character limit, which sadly will never change.

I'm really impressed with flycheck. The code is very clean and has perhaps the most readable regexps I've ever seen:

Another improvement in Python 3: `assert (1 == 2, "some boring message")` raises a SyntaxWarning that you're shooting yourself in the foot.