I find it hard to believe that HarperCollins isn’t going to be proofing its shipping product.

This statement appears to have been made by a non-Kindle owner. I can assure you that this is exactly what happens. I have purchased exactly 2 Kindle books from Amazon. One by Tor (Macmillan) and the other by Random House. Though not HarperCollins, I think they are sufficiently similar to extrapolate business practices. I can assure you they were not proofed in any way shape or form. I attempted to keep track of the number of errors I spotted while reading the Tor book and gave up when I crossed the 200 mark. I didn’t bother with the Random House one (and it is in much better shape, usually just extra or missing commas).

The Kindle itself is a fine device but the e-books I have seen are pretty terrible.

XML is like more-powerful HTML (in fact HTML is a subset).

Not true.

HTML is a subset of SGML.

XHTML 1.0 is an XML application of XML, most often served as text/html.

I wrote this book that talks some about these things.

Jeffrey Zeldman asked:
This is known? Admitted? Proven? Or just asserted?

Jeffrey Zeldman didn’t ask that.

Jeffrey Zeldman was quoting “God of Biscuits,” who was challenging Jeffrey Zeldman’s assertion that third-party conversion from publisher files to e-books to Kindle format introduced errors.

“God of Biscuits” was implying that Jeffrey Zeldman’s observations were invalid unless Jeffrey Zeldman could support them with such things as “Failure statistics on a run-rate basis.”

In quoting “God of Biscuits”‘s dismissal of Jeffrey Zeldman’s blog post, Jeffrey Zeldman was not questioning or referring to or thinking about XML workflow.

Likewise, “God of Biscuits” was not questioning or referring to or thinking about XML workflow.

I am not sure what is being questioned. XML workflow is known, proven and documented. CUPs XML workflow is explained here.

They say:

The fully XML-coded master files enable us to repurpose the material for whatever publishing needs Cambridge and you may have for the content in the future. This could be just the book, or web material, e-books and smartpdfs (which are a sort of simple e-book) for display on the web, CDs, new editions or marketing material.

XML is like more-powerful HTML (in fact HTML is a subset). In the same way that a (well-written) web-page will reflow and reformat when you resize your browser, so XML can automatically be reformatted and reflowed into any format.
It requires more information to be recorded in the document originally (just like converting plain paper into a web page) but gives more flexibility later. Whether publishers of fiction, for example, would make such an investment is a moot point, but comma-perfect automatic kindling is possible.

Straight fiction flows well into either a scroll or a page. Paragraph break and punctuation conventions were developed by printers to accommodate the page and adapt well to a scroll and most divisions in straight fiction use those conventions.

Guidebooks, how-tos, plays/screenplays, and other highly structured texts run into trouble on the Kindle (and on the web (a scroll)). I expected that, but I didn’t expect trouble with having multiple fonts in a single book. Code (if and when it appears in a Kindle book) appears in the same font as the paragraph text. This seems like a no-brainer, as does displaying boxed text, and pulls.

The paradigm confusion is evident where and when the Kindle designers tried to accommodate the page or the scroll: the percentage of book read indicator is supposed to make up for not feeling the weight of the pages transfer from right to left as I read as much as the white jimmy mouse substitute is supposed to give me more control over what happens within the scroll.

All that said, I’m glad that Amazon tried so hard to make room for the old ways, but I don’t think the e-book devices will succeed by simply matching the successes/requirements of all their predecessors, instead written works will need to adapt to the new medium as they always have from scrolls to illuminated manuscripts to the printing press to the web.

I’m speaking from personal experience, and from conversations I’ve had with book people. The personal experience prompted me to have those conversations. Like you, I was curious to find out if what I’d experienced was an anomaly or a common occurrence.

Those conversations led me to believe that there are known problems with Kindle conversions, and that publishers don’t always have the resources to identify and fix the problems before the electronic file gets sent to Amazon and sold to customers.

Publishers, as you may know, are not thriving at this time—not that they ever have. Especially at this time, publishers don’t have money to throw at lots of things, including salaries. Most people I know in publishing are in it for love. They are passionate about books and reading, about knowledge and sharing, and maybe even hopeful about the possibility of an informed democracy to bring some light and happiness to human life.

The global economy shows signs of recovering (Wall Street, housing market), but people are still getting laid off, and some sectors of the economy are still hurting very badly.

All of that is to say that it’s only to be expected that there aren’t necessarily staff proofreaders comparing Kindle editions to the original print files on a line-by-line and word-by-word basis. This might change as digital editions continue to grow as a percentage of a publisher’s market, and publishing companies reallocate budgets, but if you’ve worked in companies, you know that planning cycles lag—there may be a year or more between recognition of a change, acceptance of a plan to accommodate that change, and working implementation of that plan.

To return to your question, this is a blog post, not a feature story in The Economist. I don’t have statistics (don’t know that anyone does—gathering statistics is also an expensive activity), haven’t spoken to dozens of publishers, haven’t hired teams to run tests on and perform an audit of, say, 30 random Kindle editions. Just a guy talking here.

I can’t say what goes on at Harper-Collins. Wouldn’t know. I’ve had nice meetings with them (unrelated to this conversation), but never discussed Kindle editions with them.

I also, truth to tell, haven’t discussed this with my own publisher. Which is weird, now that I come to think of it. If alerted to a problem in a Kindle edition, publishers fix it in a subsequent update, just as they do with print problems. My own publisher, who are like family and wizards to me, are checking for problems in Kindle editions.

If it doesn’t happen during the conversions from publisher files to e-books and from e-books to Kindle, then the book fairies must be introducing the errors.

Then there’s evidence? Anecdotes? Failure statistics on a run-rate basis?

I can see OCR’d stuff being pulled in from old texts and shoved into ebooks that get dumped into the market without much QC, but I find it hard to believe that HarperCollins isn’t going to be proofing its shipping product.

Y’know, like they do for each printing of each edition of each format on paper?

However, I think it’s a reasonable assumption that the vast majority of people are using their Kindles for novels, short stories and so on; these are more or less straight text and as such are not nearly as prone to the trouble you describe. My wife – who is an editor by trade – has read hundreds of books on her Kindle and has very rarely encountered such issues, although (somewhat apropos) she’s currently reading a traditional book which has one signature bound upside-down.

However, if this problem is as widespread as you imply, with the recent introduction of the bigger Kindle (aimed at the educational/text book market) this is Amazon’s problem to fix, and soon. Hopefully bringing light to the issue at least ensures that DWWS 3e ends up Kindle-intact.

This is known? Admitted? Proven? Or just asserted?

If it doesn’t happen during the conversions from publisher files to e-books and from e-books to Kindle, then the book fairies must be introducing the errors.

I have no connection to CUP except that I index their books sometimes, so I see the extra work which goes into providing the future flexibility.