The Technical Text

By Nirmaldasan

Technical writing is the art of communicating technical knowledge to a specified audience. The topic may be as simple as a recipe or as complex as an integral equation. Some of the common technical documents are business letters and user manuals. The nature of the subject and audience determines the style and structure in which technical content is packaged. The text may contain definitions of technical terms, descriptions of products, instructions and examples. In most cases, text is accompanied by graphics/tables/illustrations. Some of the text structures are a) inverted pyramid b) topic-development-conclusion and c) situation-problem-solution-evaluation.

Technical writing requires an understanding of three principles: a) brevity b) clarity and c) scannability. Brevity is the avoidance of superfluous words, phrases and sentences. Clarity is the unambiguous and logical presentation of data or ideas. Scannability is the use of lists, visuals and tables based on the principles of brevity and clarity.

Technical writers have to choose words with care. The P-F-S rule describes a hierarchy of Precise, Familiar and Short words. The order of choice is P-F-S, P-F, P-S, P. There may be more than one precise form of expression. Technical writers have to first seek a word that is precise, familiar and short. If not available, then precise and familiar word is the next choice. This is because a familiar word is better than a short but difficult word. If even a P-F word is not available, then go for the P-S word. The final choice is the P word. We have ignored the F-S words because in technical writing precision is more important than familiarity and brevity.

Technical writers should also take care in constructing sentences and paragraphs. They need to know basic grammar and the different kinds of sentence styles. Some of the common grammatical mistakes are dangling modifiers, misplaced phrases, subject-verb disagreements and wrong tenses. Some of the stylistic errors are redundancies, weak be-verbs, unnecessary expletives and overnominalisation. These problems are dealt with by David A. McMurrey in his book on technical writing. The book deals with all aspects of technical writing and is available online at http://www.io.com/~hcexres/textbook/

A well-written technical text, based on the principles stated here, may now be checked for readability using the FORCAST formula. Patrick FORd, John CAYlor and Tom STicht developed this formula for measuring readability of technical materials. Sticht, who headed the team, published an article about the formula in 1973. Here is the formula: 20 – N/10. N is the number of monosyllabic words in a sample passage of 150 words. The greater the score, the greater the reading difficulty.

If the FORCAST score is high, the text has to be rewritten. The formula seems to suggest that monosyllabic words enhance reading ease. It may not be wise to write to a formula. Moreover, the formula is only indicative of reading difficulty. Nevertheless, it may be stated as a principle that monosyllabic words are better than polysyllabic words. But brevity may be sacrificed if precision is at stake.

The rewritten text must now be reviewed by peers and then the final version lands on the desk of technical editors.

(presented at the SDNB Vaishnav College, 19 February 2005)

Compact Readability Chart

By Nirmaldasan

(nirmaldasan@ hotmail.com)

 

For those who fear a formula but still wish to know the grade-level of a text, Edward Fry’s Readability Graph has the answer. Though Dr. Fry recommends three samples of 100 words, we shall take just one sample to avoid the calculation of averages. All we have to do is count the number of sentences and syllables in a 100-word sample and look at the relevant zone in the graph.

 

The Compact Readability Chart (CRC) is a small attempt to tabulate the reading levels. A shortcoming of this simplification is that it does not show grade-levels; it shows on a seven-point scale whether the text is very easy or very hard or the other levels in between.

 

Take a sample of 100 words. Count the number of sentences (S) and excess syllables (X). Excess syllables can be easily counted by leaving out the first syllable of each word. The point of intersection of X and S shows the level of reading ease: VE (very easy); E (easy); FE (fairly easy); S (standard); FH (fairly hard); H (hard); and VH (very hard).

 

Sentences (S)       →	1	2	3-4	5-6	7-9	10+
Excess Syllables (X)↓
0-9	VH	FE	E	VE	VE	VE
10-19	VH	S	E	VE	VE	VE
20-29	VH	FH	FE	E	VE	VE
30-39	VH	H	S	FE	E	VE
40-49	VH	H	FH	S	FE	E
50-59	VH	VH	H	FH	S	FE
60-69	VH	VH	H	H	FH	S
70-79	VH	VH	VH	H	H	FH
80-89	VH	VH	VH	VH	H	H
90+	VH	VH	VH	VH	VH	VH

 

The CRC is the outcome of a close study of the reliable Flesch Reading Ease. I had problems in filling some of the squares. For example, I had a doubt whether it should be VE or E in the square of intersection between 0-9 (X) and 3-4 (S) and I subjectively settled for E. If you are not happy with the entries, you are welcome to change them.

 

A word about X. Flesch’s simple syllable-counting procedure adds X to the number of words to get the number of syllables. Irving Fang’s easy listening formula is based on X. And Dr. Fry has been for long counting X instead of syllables.

 

Understanding Smart Language

By Nirmaldasan

(nirmaldasan@hotmail.com)

 

2007 marked a milestone in readability research with the publication of William Dubay’s Smart Language in January and Unlocking Language in March. In Smart Language subtitled ‘readers, readability and the grading of text’, DuBay defines readability as the ‘ease of reading created by a literary style that fits the reading level of the audience’. The book discusses ‘reading ease’ and ‘reading level’ but leaves out ‘literary style’ — may be because it is governed by aesthetic principles which is beyond the scope of this book. It describes the course of readability research and explains the quantitative approaches to the creation of smart text which, according to Dubay, is that language adjusted to the reading level of an audience. “Language can be very well written — and very plain — and yet written at the wrong reading level,” he writes, obviously to drive home the point that plain English must also be smart language.

 

Smart Language is an instructive and useful book. Divided into two parts, the first gives a brief introduction to the adult literacy surveys and reading habits. The second part deals with the grading of texts and what science has learned about ‘making reading easy for different classes of readers’. DuBay passionately advocates the use of readability formulas as a first step — and not the only one — in assessing readability levels. However, he reiterates the old caution: “Never write to a formula.”

 

A very useful feature of Smart Language is the description of a number of readability formulas, including the Fry graph and the Dale-Chall formula. George Klare’s normed passages, presented in the appendix, may be used in the classroom to demonstrate the difficulty of grading texts.

 

The classic readability studies, which constituted just a chapter in Smart Language, is the subtitle of the second book Unlocking Language, in which Dubay has edited 10 landmark studies in readability. The first by Edward L. Thorndike titled ‘Word Knowledge in the Elementary School’ appeared in Teachers College Record (November 1921); and the last by Rudolph Flesch titled ‘A New Readability Yardstick’ appeared in Journal of Applied Psychology (3 June 1948). This book is chiefly for the scholar. DuBay introduces each of these writers and places each article in its historical context. He also discusses the contributions of other key researchers whose articles are not in the public domain: L.A. Sherman’s ‘The Analytics Of Literature’ (1893), Harry D. Kitson’s ‘The Mind Of The Buyer’ (1921), Gray and Leary’s ‘What Makes A Book Readable’ (1935), among others.

 

In the preface, Dubay writes: “The purpose of this book is to bring students of reading into contact with this introductory sample of the original articles, methods, and thinking of these educators. In all of them, we see the urgency and pragmatism of the times. I hope that reading them in context will highlight their special place in the story of our remarkable language.” Yes, it does. Both the books deserve to be read, studied and preserved. Purchase details are available online: http://www.impact-information.com

 

Flesch Reading Difficulty

By Nirmaldasan

 

Sixty years ago, Rudolph Flesch revised his original readability formula, which predicted difficulty on a scale of 0 (‘very easy’) to 7 (‘very difficult’), and made it ‘a more useful instrument’. In ‘A New Readability Yardstick’, published in the Journal of Applied Psychology (3 June, 1948), he replaced the count of affixes with a count of syllables, expanded the scoring system (100 points) and ‘turned around’ the formula ‘by reversing the signs to predict reading ease’.

 

The Flesch Reading Ease (FRE), in its general form, is calculated by measuring the average words per sentence (AWS), the average syllables per word (ASW) and substituting their values in the expression 206.835 – (1.015 * AWS) – (84.6 * ASW). Flesch suggests a simple procedure to count the syllables. He writes: “To save time, count all syllables except the first in all words of more than one syllable and add the total to the number of words tested.” However, he did not incorporate this idea of excess syllables in his formula.

 

Irving Fang’s Easy Listening Formula, measured by the number of syllables more than one per word in a sentence, seems to have been inspired by Flesch’s simple syllable-counting procedure. And, taking a leaf out of Fang’s book, Davis Foulgers incorporated the count of the excess syllables in the Flesch formula. Foulgers’ Simplification (FS) is 122.235 – (1.015 * AWS) – (84.6 * AXW). The new variable AXW is the average number of excess syllables per word.  The FS has achieved two things: first, the constant has been reduced from 206.835 to 122.235; and second, syllable-counting has been reduced to the counting of excess syllables. And, what is more, the FS gives the same answers as the FRE.

 

But one can ask for more! A further simplification is achieved by subtracting FS from 100 and ‘turning around’ the scoring system to predict reading difficulty instead of reading ease. Here follow the Flesch Reading Difficulty (FRD) formula and the ‘turned around’ scoring system:

 

FRD = (1.015 * AWS) + (84.6 * AXW) – 22.235

 

Scoring system: 0 to 10 (very easy, 5^th grade); 11 to 20 (easy, 6^th); 21 to 30 (fairly easy, 7^th); 31 to 40 (standard, 8^th to 9^th); 41 to 50 (fairly difficult, 10^th to 12^th); 51 to 70 (difficult, 13^th to 16^th); and 71 to 100 (very difficult, college graduate).

 

The FRE, the FS and the FRD yield the same answers. Sixty years ago, it was the FRE for calculating reading ease. And now, I think, it will be the FRD for calculating reading difficulty.

 

Measuring Listenability

By Nirmaldasan

(nirmaldasan@hotmail.com)

 

I think I should begin with a disclaimer that this article lays no pretence to scholarship. From William DuBay’s ‘Principles Of Readability’ and Jim Trelease’s ‘The New Read-aloud Handbook’, I have culled some information on listenability. We gain listening skills much before we learn to read. But as we move up from Grade 8, the difference between reading and listening skills decrease; at Grade 12, there is no difference between the two. And researchers have found that beyond Grade 12, it is easier to understand a text when read than when listened to.

 

The Reading Grade (RG) is measured on a scale of 1 to 17+. Let the scale for Listening Grade (LG) be 1 to 12+. I assume that there is a gradual and uniform decline in listening skills with LG = 0 and RG = 3 as starting points. I assume that if LG increases by 1, then RG increases by 0.75. Put the other way, if RG increases by 1, then LG increases by 1.33. This assumption ensures that at Grade 12, LG = RG. If this assumption is right, then here are my conversion formulae:

 

LG = (4/3) * (RG – 3)

RG = [(3/4) * LG] + 3

 

Since books are graded based on RG, it becomes easy for us to calculate LG. Suppose you pick up a Grade 7 book and wish to know its listenability. In this case, RG = 7 and hence LG = (4/3) * (7 – 3) = (4/3) * (4) = 16/3 = 5.33. The score tells us that the book may be read aloud to a fifth grader.

 

But what if the book’s grade is not known? We may use Harry McLaughlin’s Smog Grading for readability and Irving Fang’s Easy Listening Formula (ELF) for listenability. But to keep our calculations simple, I prescribe modified versions of these formulae.

 

Modified Smog Reading Grade (MSRG) = 3 * [(square root of P3) + 1], where P3 is the number of polysyllabic words in three sentences.

 

Substituting the MSRG in the conversion formula for LG, we get Modified Smog Listening Grade (MSLG) = 4 * (square root of P3).

 

Fang’s ELF is calculated by counting the number of syllables more than one per word in a sentence. ELF = S – W, where S and W are the number of syllables and words in a sentence respectively. This formula punishes every extra syllable. Now, if we assume that all polysyllables are equally difficult, then we get the Approximate Easy Listening Grade (AELG) = N + P, where N and P are the number of non-monosyllabic words and polysyllabic words in a sentence respectively.

 

I think I should end with a caution that these formulae are yet to be thoroughly tested.

 

J.M. Barrie’s Peter Pan

By Nirmaldasan

 

A children’s classic that you can return to, even after growing up, is J.M. Barrie’s Peter Pan, first produced as a play in 1904 and then as a story book in 1911. The eponymous hero of Neverland who never grows up, the ferocious Captain Hook hunted by a ticking crocodile that has swallowed a clock, fairies, mermaids and ‘gay and innocent and heartless’ children who fly away to Neverland make Peter Pan an enjoyable reading experience.

 

No doubt a book for children, its grade level may be determined by a readability formula. Let us take 10 samples of three sentences each and apply the Strain Index (S3/10, where S3 is the number of syllables in three sentences). Since the book has 17 chapters, we may take one sample from the beginning and one from the end of each of the following chapters: I, V, IX, XIII and XVII.  

 

Here are the scores for the 10 samples: 5.5, 7.9 (Chapter I); 6.0, 4.7 (Chapter V); 12.4, 14.6 (Chapter IX); 8.2, 5.9 (Chapter XIII); 6.2, 13.5 (Chapter XVII). Parts of the book are written at grade level 4 and parts at grade level 14. However, the average yields a Strain Index of 8.49. The mean deviation is 2.615 (sum of the deviations divided by the number of samples).

 

But why should the sample be restricted to three sentences? You may wish to take a longer sample. The general form of Strain Index = 0.3 * (Sn/n), where Sn is the number of syllables in ‘n’ sentences. I would recommend a sample of 30 sentences to simplify the calculation. Thus, for a longer sample, Strain Index = S30/100, where S30 is the number of syllables in 30 sentences.

 

Now in the 10 samples that we have taken, there are a total of 30 sentences and 849 syllables. Therefore, Strain Index = S30/100 = 849/100 = 8.49. Peter Pan is suited for students of Grade 8.

 

The Story Of Writing

By Nirmaldasan

 

Nita Berry’s The Story Of Writing, published by the Children’s Book Trust in 1998, is a delightful read for students of the higher grade levels. This is the way it begins: “Remember how long you took to start writing? When you were very young, you had to learn initially to grasp a pencil or crayon firmly between your fingers. Slowly, you began to draw clumsy lines, then circles and different shapes. Later, you wrote your ABC in capitals, perhaps a little shakily to start with, but yes, you were writing! And when you were a little bigger, you developed a beautiful running hand, and wrote words, then sentences at last!”

 

This sample has 5 sentences, 116 syllables, 373 characters and 80 words. The average number of words per sentence (AWS) is 80/5 = 16. Let us calculate the readability of this passage applying three formulae: Strain Index (SI), Character-count Index (CI) and the Higher Grade Level (HGL).

 

SI = 0.3 * (Sn/n), where Sn is the number of syllables in ‘n’ sentences. Substituting the values, we get SI = 0.3 * (116/5) = 6.96.

 

CI = 0.1 * (Cn/n), where Cn is the number of characters in ‘n’ sentences. Substituting the values, we get CI = 0.1 * (373/5) = 7.46.

 

HGL = W4/10 to W5/10, where W4 and W5 are the number of words in 4 sentences and 5 sentences respectively. Here, W4 = 59 and W5 = 80. Therefore, HGL = 5.9 to 8.0. It is interesting to note that the SI and the CI scores lie within this range.

 

And here is the concluding paragraph of Nita Berry’s book: “In an electric age today, when computers have taken over the task of designing letters and text, the ancient art of writing by hand has perhaps ceased to evolve. Yet, the value of the written word is cherished. The early lesson it taught us at the dawn of our history — of how to communicate and record — is pushing mankind towards new frontiers, and newer horizons on earth and even beyond.”

 

This sample has 3 sentences, 105 syllables, 320 characters and 70 words. And AWS is 70/3 = 23.33.

 

SI = 0.3 * (105/3) = 0.1 * 105 = 10.5.

 

CI = 0.1 * (320/3) = 10.66.

 

HGL = (0.4 * AWS) to (0.5 * AWS) = (0.4 * 70/3) to (0.5 * 70/3) = 9.33 to 11.66. Here again, the SI and CI scores lie within this range.

 

Of the two passages we have tested, the first appears easier to read than the second. It is better to test more samples before averaging the scores.

 

NOTE: The Strain Index, the Character-count Index and the High Grade Level are single-factor readability formulae like the FORCAST, the SMOG and the Easy Listening Formula. I have discussed the last three formulae in ‘Fascinating Approaches To Readability’; and the first three formulae in ‘Strain Index: A New Readability Formula’, ‘Character-count Formula For Readable Writing’ and ‘Readability Conjectures’ respectively.

 

The Average Sentence Length

By Nirmaldasan

 

Plain English recommends short sentences. Robert Gunning faults marathon sentences in his book How To Take The Fog Out Of Writing. Though he admits to the possibility of long sentences being balanced and readable, he notes that only highly skilled writers such as Charles Dickens and Thomas Wolfe can write a marathon sentence with clarity. He adds: “But even these accomplished writers produced marathon sentences only occasionally. On the average, they wrote fewer than 20 words per sentence.”

 

Jyoti Sanyal’s Indlish has an interesting and informative chapter titled ‘Shrink or sink’ on the length of the sentence. He writes: “Based on several studies, press associations in the USA have laid down a readability table. Their survey shows readers find sentences of 8 words or less very easy to read; 11 words, easy; 14 words fairly easy; 17 words standard; 21 words fairly difficult; 25 words difficult and 29 words or more, very difficult.”

 

Martin Cutts, in his Oxford Guide To Plain English, offers the following guideline: “Over the whole document, make the average sentence length 15-20 words.” And what’s the reason? He explains: “More people fear snakes than full stops, so they recoil when a long sentence comes hissing across the page.”

 

Plain English also recommends short words. Even if the average sentence length of a document is 15-20 words, readability is not guaranteed. Polysyllabic words are likely to make the meaning of the document difficult to grasp. So we also need a guideline for average word length.

 

Words have two units of measure: syllables and letters. I examined the word list of Voice Of America and found that the average word length in syllables is 1.74 and that in letters is 5.67.

 

Now, we have been measuring sentences only in words. But sentences have three units of measure: words, syllables and characters. And so we may take the following as the new guideline: “Over the whole document, make the average sentence length 15-20 words, 25-33 syllables and 75-100 characters.”

Fascinating Approaches To Readability

By Nirmaldasan

 

I have been fascinated, even inspired, by three readability formulae: SMOG Grading of Harry McLaughlin, Easy Listening Formula of Irving Fang and the FORCAST formula of FORd, CAylor and STicht. All these are single-factor formulae. Interestingly, none of them are based on sentence length, considered to be a key indicator of text difficulty.

 

To calculate SMOG (smoke-plus-fog as well as Simple Measure Of Gobbledygook), one needs a long sample of 30 sentences — the first 10 from the beginning of a text; the last 10 from the end; and the rest from the middle. If P is the number of polysyllabic words in the sample, then SMOG Grading = (square root of P) + 3.

 

The Easy Listening Formula (ELF) counts every syllable more than one per word in a sentence. That is, if W is the number of words and S the number of syllables in a sentence, then ELF = S – W.

 

The FORCAST formula needs a sample of 150 words. If N is the number of monosyllabic words in the sample, then grade level = 20 – (N/10).

 

Now, let us apply these formulae on a sentence from Martin Gardner’s ‘Mathematical Puzzles And Diversions’: “Flexagons are paper polygons, folded from straight or crooked strips of paper, which have the fascinating property of changing their faces when they are ‘flexed’.” Applying the ELF to this one-sentence sample is simple, but we need to suitably adapt the other two formulae that require a longer sample.

 

Since there are 4 polysyllabic words in this sentence, one may expect 4 * 30 = 120 long words in 30 sentences. Therefore, SMOG Grading = (square root of 120) + 3 = 10.95 + 3 = 13.95.

 

The sample contains 25 words and 40 syllables. Hence, ELF = 40 – 25 = 15.

 

For the FORCAST formula, let us consider only the first 15 words of the sample: “Flexagons are paper polygons, folded from straight or crooked strips of paper, which have the …” Since there are 9 monosyllabic words in a sample of 15 words, one may expect 9 * 10 = 90 monosyllabic words in a sample of 150 words. Therefore, grade level = 20 – (90/10) = 20 – 9 = 11. In short, if N is the number of monosyllabic words in a sample of 15 words, then grade level = 20 – N. 

 

Our sample sentence has yielded different results: 13.95 (SMOG), 15 (ELF) and 11 (FORCAST). This cannot be helped because each of the formulae used a different indicator of text difficulty. Instead of asking which is right, we may simply calculate the average. This works out to 13.31. And that, we shall say, is the grade level of Gardner’s flexagonal sentence.

 

Readability Conjectures

By Nirmaldasan

 

I have two conjectures: one for primary-grade texts; and the other for non-primary materials. The conjectures are based on the wish that the number of words should suffice to evaluate grade levels.

 

Let W3, W4 and W5 be the number of words in 3, 4 and 5 sentences respectively. Then, the grade levels can be easily calculated:

 

* Primary Grade Level (PGL) = W3/10 to W4/10

* Higher Grade Level (HGL) = W4/10 to W5/10

 

Let us test these conjectures.

 

‘Just A Few Words, Mr. Lincoln’ by Jean Fritz is the story about the Gettysburg Address written for grades 2-3. Here are the opening sentences:

 

“President Lincoln was one busy man.

“He had two big jobs.

“He had to free the slaves.

“And he had to win the war. The Civil War.”

 

The Strain Index = 2.1 (S3/10, where S3 is the number of syllables in three sentences). And the PGL = 1.7 (W3/10) to 2.4 (W4/10).

 

The last page of the book contains the full text of Lincoln’s Gettysburg Address. Here follow the first and second paragraphs:

 

“Four score and seven years ago our fathers brought forth on this continent, a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.

 

“Now we are engaged in a great civil war, testing whether that nation, or any nation so conceived and so dedicated, can long endure. We are met on a great battlefield of that war. We have come to dedicate a portion of that field, as a final resting place for those who here gave their lives that that nation might live. It is altogether fitting and proper that we should do this.”

 

Certainly, this is a non-primary-grade speech. The Strain Index = 9.5. And the HGL = 9.1 (W4/10) to 10.2 (W5/10).

 

We have successfully tested two samples: one for each conjecture. A single swallow does not make a summer, but surely is a sign of hope.