Sunday, 24 April 2011

Typography is about reading – and so are ebooks

Amazon’s latest UK Kindle ad shows how little care it takes over the typography of the ebooks that it hosts. This is their latest full-page ad in the Guardian. I can just about forgive the opening words in ALL CAPITALS, because, although letterspaced SMALL CAPITALS are much nicer, even William Morris resorted to chapter openings in the all-up style. The real give-away is the confusion of the hyphen with the dash. My first reading of the opening sentence assumed a Joycean style, mimicking ‘riverrun, past Eve and Adam’s’, followed by ‘doubt-but’ (what a disbelieving goat does?). But no, I was supposed to read those tiny, close-up horizontal marks as dashes – and then the parenthetical clause should become clear. Except it doesn’t. And William Boyd is a parenthetical stylist, a major dash-monger, with three on the opening page of Ordinary Thunderstorms, all decisively mangled by Kindle. He also believes in the use of the hyphen when creating compound adjectives: ‘pale-faced’, ‘even-featured’, ‘charcoal-grey’. Oh dear, they get mangled too, as Kindle thinks that they are non-breaking hyphens, and so studiously avoids correctly justifying the line preceding them to ensure that the compounded words are not (as they should be) divided by a line break. Here by the way is the printed original:

How can so little care be given to the presentation of text on a[n electronic] page? Do publishers care, or even realize, what is happening to the texts they lovingly commission, copy-edit, and proof-read, when they enter the electronic domain? 
I wonder, especially if they sub-contract the ebooking of their print files to Amazon, rather than apply quality control themselves. Here, for example, is a complaint on Amazon about a student edition:
‘The reason for the poor review is not anything relating to the translation, but the rendering of text not only in this book, but many of [this publisher’s] translations released in this series. The main problem is that the text is not rendered in a clean manner, so that, for example if you highlight “Livy” in the introduction of this text it is rendered “Liz7"” when you look under “highlights and annotations”. This means that if you search for “Livy”, your search will not find this word. This has been a constant problem in a number of this publisher's books. I have contacted them about the issue but have received no response. I would recommend [another publisher’s] text which provides a clean rendering of the text as well as hyperlinks to the footnotes which this publisher does not provide.’
The image below is not from the text criticized, but shows what happens when a printed book is scanned, word by word, to produce a searchable PDF. 
And the original print edition:
Note how the running headlines, which orient the reader so successfully in the print edition, are very little use in the ebook.
Not that other publishers are blameless: the following screenshots are from an edition of the Canterbury Tales, where we can see
  1. headings not identified as such, because they do not ‘keep’ with the following text; 
  2. headings (such as ‘NOTES’ which are at the wrong level in the heading hierarchy;
  3. misaligned note cues causing text misalignment;
  4. verse not correctly formatted, so that a turn-line does not indent, nor does the start of a line assigned a line number align correctly.

And here is the original printed page, showing text and notes simultaneously:
But at least this edition of Chaucer has line numbers. Another publisher's edition of Wilde simply omits them, leaving the reader floundering helplessly in relation to a citation in any other edition or critical text, or indeed in relation to the edition’s own notes, which are identified by line number, but are not accessible by hyperlink from the text.

Let’s look at the original print edition, and see how effectively the reader is oriented by indentation, headlines, line numbers, and page numbers; and how easily one can look up a note at the back of the book, and know where to return to in the main text.

In comparison with the inadequacies of consumer electronic texts, printed books are still miracles of compact, considerate texts that care for their readers, and how they are used.


Here’s an interesting link to a TeX implementation of justification on a Kindle.

Edit 2

Thanks to all the readers who picked up my misreadings and misspellings.


Ian Dennis said...

Get at them Paul!

Saw the ad this morning and thought the same - what a grey, horrible, unappealing page.

Ian Dennis

Dr Ian Hocking said...

Thanks for an engaging post, Paul. It's good to see a little empirical light shed on this issue. There is a compromise to be struck between what the author/publisher wants the book to be and the limitations of the Kindle reader itself, but half decent mark-up can get us most of the way there. One persistent issue these days seems to be the failure to omit indent at the beginning of new sections. Can't stand it.

Anyway, thanks for the post.

Andrew said...

Which is exactly why I can't bring myself to buy an e-book reader. I love properly typeset text.

Anthony said...

1) As an avid reader and not a typography enthusiast, it appears to me that typography is not about reading, but design, except at a very basic level (Font and size).

2) In a very practical sense, some of the things you mention (Dashes versus hyphens, for example) don't affect my reading experience one iota. I'll admit that others might bother me if I read books in which they were present (Variations in the handling of footnotes).

3) Unless I have trouble making out what a word is, there's very little Amazon or publishers could do to properly cause me problems. Then again, perhaps my practical approach to reading equips me well to ignore any of these shortcomings. I may not be a very good gauge of the average reaction to these typographical missteps.

4) Incidentally, the font choice on this blog does more to hinder my reading of it than anything I've come across while reading on my Kindle.

On the other hand, were all of these problems addressed, and Kindle books ideally formatted, I'm sure the reading experience would be noticeably improved. This is simply because while many do not notice something that is a little lacking, they often do notice something that has seen a lot of attention to detail.

Anonymous said...

You fontsters are *weird*.

Franz said...

Great piece. I've had to return more than one e-book because of the shoddy workmanship. Let's hope complacency doesn't win out. Some more thoughts on the state of e-books at

Anonymous said...

Might it be the simple fact that most epubs are created using MS Word .doc files as the standard input file, the epitome of bad layout and design.

Am in the process of producing epubs of my own books and shudder to think of what will happen when I have to use that abysmal program to birth them.

Anonymous said...

I would just be thrilled to have a a ragged right option instead of full justified. It would come in especially handy on the smaller screens of the iPhone and iPod touch. Most iBook ebooks come this way already, but I haven't bought a Kindle book that was formatted that way yet.

name99 said...

"In comparison with the inadequacies of *consumer electronic texts*, printed books are still miracles of compact, considerate texts"

It is important not to throw the baby out with the bathwater. Your complaints reflect (badly) on Amazon and (especially) commercial publishers. But they are NOT a generic feature of electronic publishing, or ebook readers as a general class of devices.

I now do most of my reading on an iPad, and most of what I read is technical material in the form of PDFs. It should be noted that every element of the chain that produces this experience is very different legally and commercially from what you're complaining about, but essentially the same at the technical level.
The works I read are pretty universally created using LaTeX which (unlike the abominable MS Word and pretty much all its commercial equivalents) produces high quality typography, including mathematical typography.

They are published as PDF (yay, Adobe --- this one success does a lot to make up for your notable screwups in other media), and Apple then tries hard to make PDF work well on iPad. (As opposed to the eBook reader publishers, who seem to begrudge PDF's existence, and do a lousy job of supporting it.)

The reading software I use (Good Reader, $5 at the App Store, and worth every penny ten times over) provides both powerful capabilities to organize MANY PDF files (I have thousands on my iPad) and good (if not yet perfect) support for cropping the margins of PDFs so that the readable content fills the screen area with no wasted space.

Finally Apple give us a device with a screen that makes sense in reading PDFs and a screen resolution that, while not yet perfect, is acceptable for now.

As I said --- by all means complain about the crap that is being fed you by big publishing. But the appropriate response is to harness the technology to create a better eco-system --- which for the works being discussed here, all of which seem to be out of copyright --- should not be difficult from a legal perspective.

Look at what the math/physics community has done, from creating LaTeX to creating arxiv, because they appreciate that surrendering control of what they love most to commercial entities would be a disaster; and learn from the experience.

Anonymous said...

Thank you for drawing attention to this problem. Your analysis is kind of high-brow (and I mean that as a compliment) but even the recent current mass-market bestsellers from Kindle, Kobo et al. have so many problems with typos, weird hyphen/dash issues, mismatched smartquotes, dumbquotes mixed with smartquotes, and so-on that although I love ebooks I dread buying them, as a little over half the time I'll find them unreadable and will have wasted my money. And Project Gutenberg stuff is even worse, which is hard to understand as they are generally each worked on by groups of people -- one assumes they're just people not that into these kinds of details. (Gutenberg stuff available thru Kindle is just as bad, in spite of the fact it's not free.)

Soen said...

Is it ironic that this piece on poor typesetting on the Kindle is actually poorly typeset and is pretty horrible to look at in Chrome on Windows?

Anonymous said...

In addition to the typography issues with eBooks, there's also a lack of proofing. I have never seen so many typos, punctuation problems and spelling errors that crop up in eBooks. I wonder if anyone proofs the books at all before they're pushed out the door to the consumer.

Amy Alkon said...

Thank you so much for this post. I was shocked by how unreadable the Kindle text was (in the UK ad). It's disappointing to see how ebooks seem to be an annoyance and an afterthought to publishers -- just as they are starting to outsell printed books on Amazon. Yet, they are a mess in so many ways: For example a professor friend who wrote a book with footnotes said she had to make them endnotes because footnotes wouldn't show up on the Kindle. Nuts.

Anonymous said...

I bought an ePub copy of Glen Cook's Lord of the Silent Kingdom (second in a series) from the iBooks store and it has line-break hyphens embedded as part of the text. The result is that when I resize the text to make it larger I feel I'm reading a parody of 19th century writing as a character worries that the Grail Empress might "take ad-vantage" or wonders about the plans of the "Patri-arch."

Interestingly, the first book of the series is not available through the Apple Store, and the third is and has no problems with random hyphens, so this seems to be a problem they corrected as time went on.

Jace said...

Just wanted to point out that your first quote about "Livy vs Liz7" is not about typography. It's about OCR errors.

Blueye said...

Cool article. One interesting notice though. It appears that Boyds original typeset is using a space + endash + space layout.

So as abysmal as the e-reader is using a hyphen, the above approach looks almost as incorrect to the more appropriate (word emdash word) without spaces style. At least that's how this designer with his editor wife in toe has always set type. .

Anonymous said...

Thank you, kind blogger, for pointing out the elephant in the eRoom! The state of typography and page layout in ebooks is abysmal! Currently ebooks seem to be a mere afterthought, a "freebie" that they can produce with the simple click of an "EXPORT" button. I hope that before too long, publishers will give the layout of ebooks the attention they deserve.

Damian Cugley said...

I was going to comment that so much of this was stuff that Knuth managed to get right in 1982 with TeX—but then I see you have a link to TeX on the Kindle at the bottom of the article already! :-)

Anonymous said...

banned your site within my browser, because clicking on the pictures opens a new browser advertising window.

Mike Topping said...

Great piece. Interestingly, the dash/hyphen confusion is absent from the copy of Ordinary Thunderstorms I have on my IPhone Kindle app. Not sure whether the file has been updated by the publisher, or the original file is just better rendered by the app than on the Kindle itself (I don't have a Kindle so can't check).

Anonymous said...

Talking of good typography, why are you lines about and inch long? Makes it very hard to read...

Shawn said...

Good article, and definitely a subject that needs more attention. In you opening paragraph you quote the text as "double-but" when it is actually "doubt-but".

ches said...

I completely agree with your comments, which is why I have been working on an iPad app "iTeX" that is a reader that handles LaTeX documents specifically formatted for the iPad, with separate layouts for portrait and landscape presentations. This lets the typographer create exact images for a device, with known dimensions and resolution.

A new version should be up in a week or so.

Bill Cheswick

Douglas Edric Stanley said...

Good thing I have the Safari Reader button to read this (excellent) article, given that the default layout on blogger is so utterly horrid.

This whole field of digital typography is still very much a work-in-progress.

Luc said...

It is sad, but the new generation were never thought the art of typography. Because computer. Because easy. Because ignorance. Because "I don't care". Sad.

Burton Hanson said...

I'm visiting your site via MetaFilter's posting about it. I know that right-justification, if done properly, looks better to the eye than no-right-justfication(and worse if done improperly), but, and perhaps I'm the exception, I don't mind reading text that isn't right-justified. Be that as it may, assuming one is preparing text for publication as an e-book by Amazon, is there a web site that helps one do it in a way that will look its best?

Sean Harrison said...

I use a jail-broken Nook color for most of my reading. For many purposes, a PDF of the same typesetting that is sent to the printer is in fact the best option for an ebook. Turned sideways on a seven-inch screen, most books can be read quite comfortably this way.

Richard said...

Thank you for explaining what I realized instinctively when I first tried to read a book on the kindle. My own poor description was that it was boring! A completely inadequate description. I thought it was the font but now realize it's the whole presentation. Thanks to you I'll be able to articulate my disappointment with the kindle (and every other electronic reader). Please continue to press the issue, perhaps these companies can be shamed into correcting the problems.

Anonymous said...

I agree with your point but I also find ereaders either make text unreadably small or flow resized pages badly and screw up diagrams and maths. And do not get me started on the document management interfaces which are best described as primitive.

The goal of the ereader is not to provide a good reading experience but to get you to buy ebooks. In my case the poor typography and user interface have put mne off ereaders.

Anonymous said...

Great post. I use the Kindle app on my iPad and every book I have is littered with examples of shoddy workmanship. More often than not it looks like Amazon have shoved a scanned .pdf through ABBYY FineReader Express and then done a cursory proof to check for obvious mistakes.

In fact if you want to see an abysmal example which is unreadable download "A Compendious Anglo-Saxon and English Dictionary" by Joseph Bosworth. It is pretty much unreadable.

Anonymous said...

Never use any compressed form for graphics supporting text. None of the text examples are readable. Your formatting is worse in this sense than what the kindle did to the other text.

Nathan Derksen said...

Much better, thanks.

Anonymous said...

Thanks for freaking out all my clients. Do you live in LA? I'm coming to find you.

But in all honesty, you do have to take special care. Just sending them off to a third party formatter is a disaster.

We make all our eBooks fresh, right here in LOS ANGELES! BOOM BITCHES!

Anonymous said...

And yet Kindle notified me because they corrected the (likely) little read book, "Independent People" that I had ordered and read months earlier. The text in the first version had several typographical glitches.

I have long noticed the same issue when Word or InDesign products are converted to pdf. Often the same careful attention to detail in the print version is pointless once you post it in pdf. The side-by-side comparison shows a distinct lack of respect to precision in the coversion process.

Are we on the cusp of change? Many young communicators (my field, but I am old)seem to think that we will all have to drop our pedantic positions and embrace "twit" speak or txt. For the sake of all those who made English their profession in one way or another, I hope this isn't true. But maybe it is just a matter of time before sloppy presentation of information by electronic media is not something you'd blog about.

Anonymous said...

"Should becomes clear" ... ?
If you're going to rant - - - - do so using good grammar and English.
This may be an interesting read.
I don't know.
I stopped there.
If only you had proofed it before hitting "publish" ...

Anonymous said...

Take a look at some of the million books that Google has scanned. The OCR artifacts are much worse.

Amazon has a continuum of strictness they could enforce. The more book available the less scan precision, The more precisely copyedited and proofread, the fewer books. At this point they offer about 10 percent of their catalog in e-books. That's just enough that you can usually find good books on most topics.

They had to jump-start a brand new market. In the end things are not going to be solved until the publishers themselves take responsibility for their own books, which won't happen until the e-book edition is the primary, first, most important edition, not an afterthought. That can't happen without a thriving e-book market, for which we will have Amazon to thank.

As for justification, I think they should just default to quad left, rag right. In all honesty, most justified books, even classic typography, are pretty terrible. The only way to get even word spacing is quad left. To get decent word spacing out of justified text you need to have 80 or more characters per line, which in these days of speed reading is too wide. People don't read the same way they did in Gutenberg bible days. They scan vertically.

Mike said...

Take a look at some of the million books that Google has scanned. The OCR artifacts are much worse.

Amazon has a continuum of strictness they could enforce. The more book available the less scan precision, The more precisely copyedited and proofread, the fewer books. At this point they offer about 10 percent of their catalog in e-books. That's just enough that you can usually find good books on most topics.

They had to jump-start a brand new market. In the end things are not going to be solved until the publishers themselves take responsibility for their own books, which won't happen until the e-book edition is the primary, first, most important edition, not an afterthought. That can't happen without a thriving e-book market, for which we will have Amazon to thank.

As for justification, I think they should just default to quad left, rag right. In all honesty, most justified books, even classic typography, are pretty terrible. The only way to get even word spacing is quad left. To get decent word spacing out of justified text you need to have 80 or more characters per line, which in these days of speed reading is too wide. People don't read the same way they did in Gutenberg bible days. They scan vertically.

Nicholas Bodley said...

Your logon interface rejected a "yes" reply -- I was at Princeton for one term. Whether "United States", or the default "Yes" caused me to be rejected, I don't know. That green "1" in the area- code box is strange. Moreover, why were initial caps converted to small (Unicode term)


Regarding the topic: It's sad to see what a mess has been made. Hyphenation at the ends of lines was worked out beautifully for computers roughly 20 or 30 years ago; try DEC (before HP bought it, and while Ken Olsen was still in charge).

As to dashes: Ordinary typewriters didn't have them; now that they are available, they are still often too hard to insert (Wikkipedia editing is an exception).

As to literacy: What do you expect for typography in a country with an appalling literacy rate, so bad that lots of people can't spell "off" correctly, or don't care! A country so sub-literate that two-letter postal codes (a legacy of pre-WW II data processing -- fixed-length fields) are conidered to be abbreviations (even the USPS eventually decided that they are abbreviations).

We're a country that occasionally starts businesses with misspelled company names, such as the Millenium Hotel in NYC, or Symantec (which should have been "Semantec", as a literate employee of that company told me).

We're also a country what finds the surnames "Swain" and "Murray" too difficult to spell or pronounce.

Then, in about the past decade, unsubscripted numerals have become acceptable (even by the BBC, at times) in chemical formulae, e.g. "CO2", "H2O". This usage is not only subliterate, it's ridiculous when subscripting (and generic superscripting) are hardly difficult in any respectable text-composing interface.

I haven't studied the examples in detail, yet, but I'm interested in typography, and what you report looks truly disappointing. It will deter my purchase of an e-reader until I can obtain decently-typeset e-Books.

I'm sorry, but I think basic literacy might be more fundamental than typography. Typography, after all, is a civiliized activity, and civilization is being redefined in a hurry, to be kind.

I'm entrusting these comments to be edited, and (very likely!) pruned severly, thereby.

Thank you for permitting comments without the hassle of registering!

Anonymous said...

Dude... Seriously ?

Pat Hobby said...

Paul, I was very glad to read this, and can only hope that others do, and feel the same way. I assume that we saw the same advertisement, though I can't remember the details - because I couldn't get beyond the two hyphens masquerading as en-dashes. I just thought, You're kidding me? You've taken out a full-page ad that's (after a fashion) unreadable unless you're clued up enough about dashes to know what mistake has been made and correct it in your head. Doesn't Amazon employ anyone who knows/cares about these things? I would like to think that conventional (I think the word is 'legacy') publishers are not allowing their ebooks to carry such obvious mistakes, but if you assume that an average ebook would cost, say, £250 to proofread (properly), and publishers are bringing their frontlist and hundreds of backlist titles out as ebooks then ... well, I wonder if they want to swallow that cost. Like I say, hope your piece gets read, Best, Steve Guise

Mike McNamara said...

Great article, when I first saw the advert I also could not believe how wrong it looked. Your comments really point out one of the big failings of current eBooks.

If only publishers were as caring about how their digital products were presented as they have been in the past about their print products then we would see some improvements. But until Publishers bang their drums much louder, the eReader developers don't really have to listen.

If a Typesetting/Printing company did this much 'damage' to a publishers 'book' then they would end up with no further contracts.

Anonymous said...

What is really bizarre is that the John Lewis advertisement for Kindle uses the same extract fronm Ordinary Thrunderstorms but with correct English punctuation. Presumably, it must be something to do with the edition that has been converted. If so, how can one discover before purchase whether an edition uses English punctuation?

JDJ said...

Platforms and publishers can share the blame here.

On the platform side, every last app and device enforces its own ideas about style. Even a brilliantly designed EPUB file is going to have its margins altered on Kobo, or its alignment screwed up in Stanza, or its justification altered in iBooks, or fake page numbers pasted over the text in Adobe Digital Editions. There is little consistency between e-readers, made worse by the current poor state of those readers. All the publisher can do is complain, because even a sincere effort to make a decent ebook will meet failure.

That said, the quality of ebooks coming from big publishers is almost universally bad. (Maybe this is why the apps and devices try so hard to assert their own styles; they can't count on the majority of ebook files to get it right alone.) Production processes at most publishers are still very much focused on the paper, since that's what all the old codgers know best. When the ebook is the afterthought, a task sent to some nameless company in India for bottom dollar, you can expect the results we see.

Worse still, Amazon only accepts ebooks in their own stupid file format even though most of the world is doing the production in EPUB. Many of the books on Kindle have been exported from an EPUB file at some point, which is yet another source of poor quality.

(Most ebooks in general have been exported from an ill-suited source, like Word or a PDF.)

Unless we get a clean slate with EPUB 3, ebook typesetting is going to be terrible for the next five to ten years, at least. Let's get used to it.

Anonymous said...


That's why I'm not buying an e-reader yet.

I'm pretty sure they are aware of the issue – the problem, I guess, is that there are three main popular e-reader-brands (kindle, sony reader, nook) and even those change resolution and format with every new version. Under these conditions caring for properly typeset text becomes uneconomical.

Maybe, at some points, there will be some market-wide standards for these matters (format / resolution). Before that, I don't think publishers will care to properly set their ebooks.