[Fwd: Readability indices]

[Michael Lewis <lewism -at- BRANDLE -dot- COM -dot- AU>]

Geoff has trouble sending to the list, and has asked me to forward this
on his behalf. He says some nice things about one of my postings, so how
could I possibly refues?

Geoff Hart wrote:
>
> Greets!
>
> Bang on the head! You put the problem of readability indices into
> very nice perspective. I'd like to elaborate on what you said, but
> can't post directly to the list. Would you mind forwarding this for
> me? I still can't post directly. Thanks!
>
> --Geoff
> ********
> Michael Lewis hit the nail on the head when he pointed out that
> syntactic analysis is much more valuable than simplistic readability
> indices, which are one of my personal hobgoblins. Frankly, none of
> the common readability indexes based on word or sentence length or
> word complexity are worth using; worse, they're usually dangerously
> misleading. Strong statements? Let's do a reality test.
>
> Sentence length: Is "eschew obfuscation" really simpler for most
> readers than "write simply and clearly and don't confuse your
> readers with convoluted wording"? Is "follow standard maintenance
> procedures" more helpful to readers than providing a reminder of
> what these represent via a short list of the procedures? I think
> not.
>
> Word length or complexity: Is "neurobiology" any more complex than
> "the study of nerves"? Not to a neurologist. "Teleoperation" isn't
> much help to most of us, but makes perfectly good sense to robotics
> wonks. I don't think anybody really understands "existentialism"
> <g>, but philosophers at least know what the word represents.
> Speaking of which, you can't get much simpler than the word "I",
> but would any two techwhirlers agree on what the word means? Is the
> length of the word itself important? I doubt it.
>
> The key to readability is threefold. First, choose words that your
> audience understands, and use explanations or "wordier" phrasings
> where the words aren't going to facilitate comprehension. Second, use
> simple, unconvoluted structures, and where it's helpful, things like
> bulleted lists, subheadings, and so on. Third, test your assumptions
> on your audience. Nothing else will work reliably.
> --Geoff Hart @8^{)}
> geoff-h -at- mtl -dot- feric -dot- ca

--
Michael Lewis
Brandle Pty Limited, Sydney, Australia
PO Box 1249, Strawberry Hills, NSW 2012
Suite 8, The Watertower, 1 Marian St, Redfern 2016
http://www.brandle.com.au/~lewism
Tel +61-2-9310-2224 ... Fax +61-2-9310-5056