Typos: Who Cares?

[Luffy]

I understand completely. I used to work as a data entry clerk, and after OCR-ing a document one might have the task to reconstruct the words from imperfectly gathered letters. Sometimes errors did occur on that level, and they were quite hidden in plain sight.

[Sirtel]

Originally Posted by Luffy

There are a lot of OCR'ed books out there that are illegal and full of typos.

Nowadays most of the pirated books are retail editions. The time of illegal OCR-d books is long past.

Before you ask how I know, yes, I visit a couple of pirate sites regularly. To see what new books have been released. They're neatly grouped by genres there and it's easy to get an overview, much easier than at any legal stores, including Amazon. Who insists on lumping all paranormal romance in with SF and fantasy. Pirates don't do that.

It's off topic of course, I apologize.

[Luffy]

Originally Posted by Sirtel

Nowadays most of the pirated books are retail editions. The time of illegal OCR-d books is long past.

Before you ask how I know, yes, I visit a couple of pirate sites regularly. To see what new books have been released. They're neatly grouped by genres there and it's easy to get an overview, much easier than at any legal stores, including Amazon. Who insists of lumping all paranormal romance in with SF and fantasy. Pirates don't do that.

It's off topic of course, I apologize.

Agreed. Better than waiting for emails which are ill forwarded and formatted. It's fascinating to hear others's take on the subject. And what is happening on the pirate sites.

[DiapDealer]

It all depends on what/how/how long I've been reading something as well. I'm more likely to notice a misspelling in the middle of a short news article than I am in the second hour of a reading session of a novel I'm thoroughly enjoying.

[Frenzie]

Originally Posted by JSWolf

Given that we have a spellchecker in Sigil and Calibre, it's unforgivable for even a single typo. But what is abominable is when the eBook has errors the pBook doesn't have.

How about the classing Nook-ing? https://languagelog.ldc.upenn.edu/nll/?p=3994

I assume that particular example is probably fairly rare, but the power of search & replace has likely struck plenty.

[arjaybe]

Originally Posted by DiapDealer

OK. Thanks.
I certainly wouldn't want to read a book's-worth of it, but that little experiment is a perfect example of why I may not even see a lot of typos whenever I read. I'm hardly ever breaking a word down into its individual letters when I read. My brain is willing to see the correct whole word if there's enough markers/context in place to allow it to do that.

This "ability" also contributes to proofreaders missing errors.

[DiapDealer]

Originally Posted by arjaybe

This "ability" also contributes to proofreaders missing errors.

Exactly. Which is why all book[s] did, do, and always will, have some errors.

[Frenzie]

Originally Posted by DiapDealer

all book

That kinda jarring though.

(Hopefully, so is something above. )

[DiapDealer]

Can't even give a guy 2 minutes for corrections?

That's another part of the problem.
People don't only have to get it right. They have to get it right right now.

[ilovejedd]

Originally Posted by DiapDealer

Did you have any difficulty reading the jumbled passage?

I can understand it but not as fast as I'd normally read. I certainly wouldn't want to read a whole page that's written entirely like that.

I'm quite forgiving of occasional typos (or bad OCR). As mentioned, the brain's interesting and shows us what we expect to see. Too many errors and I start finding it quite distracting and takes me out of the story.

[DiapDealer]

Originally Posted by ilovejedd

I'm quite forgiving of occasional typos (or bad OCR). As mentioned, the brain's interesting and shows us what we expect to see. Too many errors and I start finding it quite distracting and takes me out of the story.

Agreed. I think that describes most people. But I DO think there are more than a few people who believe that getting through a book without any distractions means that there weren't any mistakes in it. We miss and forget probably more than we'll ever know. Unless we stop and make a note of them when we encounter them, I guess. Talk about taking someone out of a story.

[ilovejedd]

Originally Posted by Luffy

There are a lot of OCR'ed books out there that are illegal and full of typos.

It's not just illegal ones.

I think publishers may have lost the manuscript or don't have an electronic copy for some of their catalog titles. I've bought several catalog releases that had a number of OCR errors. Often, these were correctly spelled words that were out of context in the sentence/paragraph. If I have a copy of the print book, I'd sometimes check if I couldn't figure out the sentence correctly from the context.

[Luffy]

Originally Posted by ilovejedd

It's not just illegal ones.

I think publishers may have lost the manuscript or don't have an electronic copy for some of their catalog titles. I've bought several catalog releases that had a number of OCR errors. Often, these were correctly spelled words that were out of context in the sentence/paragraph. If I have a copy of the print book, I'd sometimes check if I couldn't figure out the sentence correctly from the context.

Thank you for the specifications. I'm not quite knowledgeable about these peccadilloes. This aspect of publishing seems to be a bad thing for the future. I usually just post at the lounge forum. I am a part-time lurker, you know?

[Frenzie]

Doesn't need to be publishers as such either. Over at https://dbnl.org/, for example, you'll find a fair amount of texts that were OCR-ed a few decades ago. The biggest tell is e/c/o confusion.

Given that a fair number of the texts I'm referring to are in Middle Dutch, where spelling was quite variable, it can sometimes be more confusing than you might expect. (Of course the OCR was done on paper editions of the original manuscripts from within the past century or two; OCR-ing the original manuscripts with such minor mistakes would be quite a feat!)

I hasten to add that they provide a very valuable service. Don't take any of this negatively. It's just a minor beauty flaw.

[Fiat_Lux]

Originally Posted by Pajamaman

Have you ever written a book? Have ever proof-read a novel-length book? Those little mistakes can be easy to miss. Words that are spelled right, but are the wrong word so the spell-catcher misses them. Missing words. It is hard to catch them all over 100,000 words.

This is why a good word processor has the ability to turn on/off each of the five spelling dictionaries per language that are needed:

• Domain specific spell checking;
• 500-1,000 most commonly misspelled words. Ideally, user-customization is offered;
• The 100 most common homophone pairs/triplets, flagging each occurrence as a potential error;
• 5,000 - 10,000 most common words;
• The big list --- 100,000+ words. This is run only after the other spell-checkers have cleared the list;

This is also why your grammar checker needs to be able to:

• Use N-Gram checking;
• Construct sentence diagrams;

Tools that one usually has to run as stand alone, on a text file are:

• Concordance creation with frequency data. This picks up correctly spelled wrong words;
• List of character pairs, triplets, quads, and quins, with frequency data. This picks up strange/unusual occurrences of words. This tool can also be extremely difficult to use effectively. (I used this when writing a book, in English, about the Tao Teh Ching. Picked up every instance where «Teh» and «The» were used incorrectly;

Run all of those, correct the errors they pick up.
Then run them a second time.
Then use a text-to-speech program, to read the document to you. As oft as not, you will hear errors that made it past the other tools that you used.

However, all of those tools are prone to failure, when proofreading your own work. Which is why it is vital to have at least two other people do the proofreading.