azazil.net

I am still alive! Thanks for sticking around through the quiet period.

Even though my posts here have been, ahem a little thin on the ground since our Tennessee road trips a few months ago, I’ve been active on twitter and I’m now blogging professionally for the Haxor Network since the beginning of this month. Please, do have a look at Mac Haxor and Linux Haxor and vote for my Mac and Linux tips! Ogle at the sexy lady screenshot. Sign up for the rss feeds etc etc.

If you’re interested in being paid to post, or guest blogging for the Haxor Network, Pavs is still searching for contributors, so head on over and express an interest in the comments. The site gets over 15,000 page views per day, so posting there is a great way to raise your profile!

Although I’ll be posting shorter Mac and Linux tips and news articles at least twice a week for the Haxor Network, I’ll continue with travelogues, martial arts commentary, opinion pieces, personal productivity items, and the occasional more in depth article on computing here at Azazil.

Over the last few days I’ve been reading the wonderful TextMate documentation in parallel with James Edward Gray II‘s excellent TextMate: Power Editing for the Mac. One of the (very few) shortcomings of TextMate is a distinct lack of many of the Emacs modes I’ve grown to rely on when working on GNU Software.

Unquestionably, TextMate is Emacs reimagined for the Mac, and it’s about time I stopped chickening out and running for Emacs when I can’t quite bend TextMate to my will. So, I plan to write the missing modes and distribute and maintain them from here. I picked on what looked to be an easy start with a Texinfo mode, although it turns out to have such a huge number of directives (each with slightly different syntax from the others) that it became a much larger project than I had anticipated.

I see that the TextMate wiki have an outstanding request for implementation of a texinfo mode, and I hope this goes some way towards filling that need. Please bear in mind, gentle reader, that this is my first TextMate mode, and as such is certainly rough around the edges. I would, however, be delighted to receive contributions to improve it in the form of patches and/or constructive criticism. At the moment it provides (very comprehensive) syntax highlighting, a template document and support for the symbol list to jump directly to any node in the current file. There are not yet any commands or macros.

Download the Texinfo bundle here

Unless I am very much mistaken, you simply unpack the archive and then double click on the icon install it into your personal TextMate bundle library.

This is the third part of a short series of articles about implementing Phomnemonal, my experimental typing shorthand. The other parts are listed at the end of the article. If you’re not interested in the theory, please jump right to the abbreviations.

An English Spelling Reform Called What?

Last time I wrote an article in this series, I was so busy patting myself on the back on the cool name I had invented, I didn’t notice that using standard English (an oxymoron if ever there was one) to spell it made it impossible to figure out how to pronounce the thing. Using the sound associations I tabulated last week helps a little, if I respell it founemonl, but with four syllables to pronounce it’s still not particularly clear. When I say it, Phomnemonal rhymes with phenomenal, only where phenomenal emphasizes just the second syllable, Phomnemonal emphasizes the first two syllables equally… but enough of that :-D

Diphthongs in Phomnemonal

Although I’ve already catalogued 12 vowel sounds needed to speak English, many words have not only long and short vowels, but also vowels that change sound in the middle: Diphthongs. Since Phomnemonal (are you getting used to saying it yet?) assigns the short vowel sounds to a, e, i, o and u, we can spell most diphthongs using a pair of short vowels. Table 1 shows the final nine sounds (beyond the 30 covered last week) needed to pronounce British English, as an extended version of the vowel table in the previous article.

Once you understand how the diphthong sounds are just two regular vowel sounds smushed together, and that they are always spelled by writing the letters for those two sounds together, last week’s Egzampl might be (a bit) easier to follow. Some linguists categorize the long-vowel diphthongs here as triphthongs, but I’m definitely talking about English not Welsh, no matter what last weeks comments say! ;-)

Increasing Typing Speed by at Least 10%

In the last article I set out an alphabet along with examples of the sound for how each letter is always pronounced. Purely because of the baroque way we spell with (particularly in British) English, even when using no abbreviations at all and simply spelling words as they sound, the result is an almost 15% reduction of the number of keypresses required to write English prose.

That does require quite an investment of effort to memorise not only what sound goes with each of the 30 letters, but to relearn all the spellings of common words you write. You do, in fact, end up typing slower while you pause to figure out what that shorter spelling was… better, instead, to make a more modest start that will give some pay off right away with minimal effort up front.

Employing Zipf’s Law, it turns out that in typical English writing, more than 80% of text is comprised of only 1000 words or so. The corollary of that fact, however, is that most of the interesting information is actually conveyed by the other 20% of the words in that text. To some extent English is already optimised so that frequently used words (a, at, the, in, on, for, i, he, it) that certainly play an important part in the syntax of the language, do tend to be shorter, where the infrequently used words (important, syntax, language, infrequently, shorter) tend to be longer.

Since Phomnemonal has a 30 letter alphabet, there are 30 possible one-letter abbreviations, and another 900 possible two-letter abbreviations, which gives us the means of writing 930 of the 1000 words that comprise 80% of written English using only one or two letters. Using phonetic spellings as a base, I’ve assigned each of the one-letter abbreviations to appropriate English words, being careful to provide mnemonics for all but the obvious ones.

Table 2 shows each of the one-letter abbreviations:

An Example

In contrast to the hard to understand phonetic shorthand I used for last week’s example, this time only the one-letter abbreviations above are used:

Our apartment z less than an hour r Cape Canaveral, d we ` hoping u visit \ Kennedy Space Center today, but while we ` g \ Florida Mall o Sunday we experienced \ most incredible thunder storm I p ever seen. S we parked our bikes outside, \ air around us actually crackled f a fraction v a second n anticipation v \ loudest peel v thunder I p ever heard. Less than a minute later, s we dashed f cover n \ Mall itself \ heavens opened f sheets v driving rain peppered with cracks v lightning almost directly overhead. I h no intention v trying u actually ride n that kind v weather, so Monday morning’s severe weather warning was enough u disuade us r chancing \ round trip u \ Space Center.

This time, that same paragraph is 126 characters shorter than the original and, thanks to having selected the highest frequency words to abbreviated, the paragraph is 22 characters shorter even than the phonetic example from last time… without sacrificing readability very much at all. Put another way, using only the 28 one-letter abbreviations (a and i don’t count) above, I pressed almost 17% fewer keys than writing that same text out in longhand.

In the Next Installment

I chose the text for the two examples more or less at random, but even for text that uses fewer of the abbreviated words than average I haven’t found any natural prose that doesn’t save at least 10%. Even if you go to the trouble of learning these abbreviations well enough that your finger muscle memory types the abbreviation without conscious thought on your part, that doesn’t help out too much when you want someone else to read your writing. That is what computers are for! In the next installment, I’ll show several ways of setting your computer up to expand the abbreviations as you type: your fingers press \, but your computer types the for you.

I’ve had my MacBook automatically expanding these abbreviations for me for about 2 weeks now, and although I’m still hesitating a little on plenty of the abbreviations, it has certainly helped speed up my prose typing noticably, and I expect to reach the full 10-15% speedup in another couple of weeks time.

Related Articles

1. Typing at the Speed of Thought
2. Phomnemonal: Analysis of Sounds in English Speech

The second part of my Linux security article is the top featured artice at IBM developerWorks Linux Zone today. If you’re at all interested on how to push machine lockdown to its absolute limits, then please do check the article out (registration required, sorry).

Where the first part laid the ground work by explaining physical security, and how to remove bash (and other shells) from your Linux installation to take away the easiest way for infiltrators to execute non-validated code on your network; this part of the tutorial is where things get really interesting, and shows you how to build and administer a binary signing system that prevents the kernel from executing any library code or applications that were not signed by you.

Related Articles

1. Industrial Strength Linux Lockdown

This is the second part of a short series of articles about implementing Phomnemonal, my experimental typing shorthand. The other parts are listed at the end of the article.

What’s in a Name?

Already, I’m tired of calling this thing my experimental typing shorthand, but luckily I had a flash of inspiration yesterday and decided to christen the system Phomnemonal — a contraction of phoneme, and mnemonics, that looks a bit like phenomenal at first glance. Man, I should have been in marketing! :-D

A Phoneme Based Shorthand

I figured that part of the reason that there are so many letters in English spelling is that the language is crazy about retaining the etymology of words. But they are variously descended from the dozens of different languages of all the countries that invaded us over the last few thousand years, kicked our ass and influenced the words we use, so that now there is no consistency in the sounds made by different letter combinations: What is the point of the letter ‘c‘? Sometimes it sounds like ‘k‘ as in curtain, and sometimes like ‘s‘ as in certain. The ‘f‘ in often makes a different sound to the ‘f‘ in of, but the same sound as the ‘ph‘ in phoneme. And then there’s rough, cough and though… the list goes on.

As a first step towards shortening the number of letters typed for a given word, I’ve analysed the sounds we need to make to speak the most commonly written 4000 or so words of English. There 21 distinct consonants plus 12 basic vowel sounds. Here they all are, shown as an example of a word that uses that sound (apologies for leaning towards my own accent here, please point out any problems in the comments):

IANAL (I am not a linguist!), so these sounds probably aren’t grouped correctly and/or have high-falutin’ names that I’m not aware of. I think group 4 are the nasal consonants for instance. Feel free to educate me in the comments.

Short vowel number 6 is the indistinct neutral vowel sound when pronouncing say the, often called a schwa by linguists.

Orthography

In order to make typing as efficient as possible, having one letter represent each sound would be ideal. Unfortunately, there are already 33 sounds in the tables above, so in the first instance, I need to combine any similar sounds or find shortcuts to writing them. The easiest one to eliminate is the schwa, since we often don’t write it anyway, as in peop{}le. Since the aim here is to produce a typing shorthand, there’s no need to distinguish between thy and thigh…

Also, with the Dvorak keyboard layout I have easy access to some extra symbols which can be made good use of, so I have 30 unshifted symbols at my disposal. Anyway, after much shuffling and deliberation this is what I came up with:

This table is in asciibetical order, where most (21) of the roman letters correspond to the sounds we already use them for in long-hand English. Here are all the rules needed to respell words phonetically using the Orthography in table 3:

1. 16 consonants (b, d, f, g, h, j, k, l, m, n, p, r, s, t, v, z) represent the sounds you would expect from normal English.
2. The traditional vowel letters (a, e, i, o & u) represent the short vowel sounds from table 2.
3. Each of the corresponding long vowel sounds is represented by a remaining letter or symbol (c, q, y, ‘, w).
4. The remaining burn vowel is represented by `.
5. Any schwa sound between consonants is not written at all. A schwa at the end of a word is written with a `.
6. The thy and thigh sounds are both written with the \ grapheme.
7. The not grapheme is used to write words like didn’t (did- in shorthand).
8. The letter j is also used for the chew sound in addition to jew, since they are difficult sounds to tell apart anyway.
9. The letter x is used for the shoe sound, like some romanisations of Chinese.
10. Since we’ve totally run out of letters, the wing sound is still represented by the digraph ng.
11. Uppercase versions of each letter are written by pressing shift before typing the key for the letter itself. That is, uppercase `, ‘ and \, are ~, “ and | respectively.

An Egzampl

From Wednesday’s Gatorland post:

C apctmnt iz les \n an au` from Kaip Knavrl, and wy w` houping tu visit \ Kenudy Speis Sent` tudei, but wail wy w` at \ Fl’id` M’l on Sundei wy ekspyryunsd \ most inkredibl \und` st’m I hav ev` syn. Az wy pckd c baikz autsaid, \ q araund uz akxly krakld f’ a frakxn ov a seknd in antisipaixn ov \ laudist pyl ov \und` I hav ev` h`d. Les \an a minit lait`, az wy daxed f’ kuv` in \ M’l itself \ hevnz oupnd f’ xytz ov draiving rein pep`d wi\ krakz ov laitning ‘lmoust dairektly ouv`hed. I had nou intenxn ov trying tu akxly raid in \at kaind ov we\`, so Mundei m’ningz svia we\` w’ning woz ynuf tu disweid uz from jansing \ raund trip tu \ Speis Sent`.

Compared to the original long-hand, this is already 104 characters (or almost 15%) shorter. That’s despite the fact that a few of the phonetically spelled words are actually slightly longer than their long-hand counterparts. And we haven’t even started working out the abbreviations yet!

In the Next Installment

Now that we have a complete alphabet along with the sounds made by each letter, in the next part I’ll tackle Diphthongs (the example above shows several of these already if you can’t wait), and the first 30 abbreviations in Phomnemonal (pronounced founemonl). Those abbreviations will be 30 words chosen from the 50 most commonly used in written English to be represented by a single grapheme from the orthography in table 3 above. Even if you only learn only those 30 abbreviations and nothing else, you will gain more than 10% in your typing speed.

Related Articles

1. Typing at the Speed of Thought