almost all words marked as mis-spelled - but only in one book ?

cybmole · 10-13-2015, 04:58 AM

please see attached screen grab.
I cannot fathom this, in just this one book sigil says that even "a" is mis-spelled !

it happens with both my portable & regular sigil installs, is something in the epub book files telling sigil to not expect English ?

how to debug ?

PS the source is an azw, just bought from amazon & converted to epub with calibre

the free sample is available here if anyone can dupplicate or debug.
http://www.amazon.co.uk/Mass-Effect-...es+mass+effect

UPDATE
I think it must be something in the book, because calibre editor also highlights most words as mis-spelt; and it is NOT a conversion glitch because calibre editor on the source .azw format does the same thing.

cybmole · 10-13-2015, 05:42 AM

update, well I have sort of diagnosed it: staring closely at the text, some letters did not look quote right e.g. look at any "u" in the image.

so using copy + paste + replace all , I replaced all the "u", "v" and "a" letters with normal equivalents. But I am still mystified how it made it thu amazons validation & release procedures while full of weird alternative characters.

the book blurb says it is a hot seller in ibooks so maybe the author messed up moving it from there to amazon ?

DiapDealer · 10-13-2015, 06:20 AM

Quote:

Originally Posted by cybmole

But I am still mystified how it made it thu amazons validation & release procedures while full of weird alternative characters.

To my knowledge, there is no "validation" procedure for selling books through Amazon (not for flagging any "weird alternative characters" anyway). If kindlegen successfully produces a kindlebook from the sources, then it's a valid kindlebook. The only way a kindlebook would get flagged/pulled for errors is if enough customers complain about errors. Since it took you a long time to even notice the alternative characters (which are perfectly valid , by the way) WERE actually alternative characters, it's not likely the average reader would ever take notice.

It's a little weird, granted, but not really so mysterious. Ebook creators can do all kinds of weird things to their ebooks that most customers will never notice if they never "pop the hood" (which means the vast majority of customers). Perhaps it's a form of watermarking, or perhaps a personal quirk of the author that the ebook's creator simply obliged. Either way, it's not something that Amazon's release procedures would "catch".

Doitsu · 10-13-2015, 07:30 AM

I did a quick spot-check and found out that a couple of letters have been replaced by similar looking letters from other alphabets:

a = а (Cyrillic small letter a)
o = ο (Greek small letter omicron)
u = υ (Greek small letter ypsilon)
v = ν (Greek small letter nu)

Since they were consistently replaced throughout the sample, they're probably not used for watermarking, unless the author replaced other letters in the epub versions sold by other sites.

Whatever the motivation, it's a stupid idea, because it'll prevent readers from looking up words.

cybmole · 10-13-2015, 07:53 AM

yep that is it. they are in the amazon format original like that also.

I can't come up with a scenario for how that could happen accidentally though unless the author is Greek or has a Greek keyboard???

my first thought was that the language definition in the book's metatdata was messed up, but that looks OK

DiapDealer · 10-13-2015, 08:15 AM

My guess is intentional artistic stupidity license.

Doitsu · 10-13-2015, 08:24 AM

Quote:

Originally Posted by ⅮіарⅮеаⅼеr

My guess is intentional artistic stupidity license.

It's also possible that the author is a big fan of visual spoofing with homoglyphs.

P.S. Did you spot the non-Latin characters in your user name?

DiapDealer · 10-13-2015, 09:16 AM

Quote:

Originally Posted by Doitsu

P.S. Did you spot the non-Latin characters in your user name?

In my username?

Doitsu · 10-13-2015, 09:43 AM

Quote:

Originally Posted by ⅮіарⅮеаⅼеr

In my username?

I meant your MR member name (ⅮіарⅮеаⅼеr). I had replaced all characters, except for the final R with lookalike characters.

DiapDealer · 10-13-2015, 10:35 AM

Nope. I never noticed!
Just goes to show you: it's only an error if I notice it.

cybmole · 10-13-2015, 10:46 AM

in this instance, going back to the original source on my Kindle, the a and the o are indistinguishable, but the strange lower case u is easy to see , and annoying once you know it's there.

the "breakthrough" was from thinking what do all the mis-spelt words have in common and the non-misspelt ones don't. so once I fixed the "u" and some of the words then became OK, the hunt was on

10-13-2015, 04:58 AM	#1
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	almost all words marked as mis-spelled - but only in one book ? please see attached screen grab. I cannot fathom this, in just this one book sigil says that even "a" is mis-spelled ! it happens with both my portable & regular sigil installs, is something in the epub book files telling sigil to not expect English ? how to debug ? PS the source is an azw, just bought from amazon & converted to epub with calibre the free sample is available here if anyone can dupplicate or debug. http://www.amazon.co.uk/Mass-Effect-...es+mass+effect UPDATE I think it must be something in the book, because calibre editor also highlights most words as mis-spelt; and it is NOT a conversion glitch because calibre editor on the source .azw format does the same thing. Attached Thumbnails Last edited by cybmole; 10-13-2015 at 05:03 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Mis próximas lecturas	cristbo	Cafetería	3	09-01-2021 05:39 PM
Dictionary - stored words are marked as wrong	rubeus	Sigil	2	12-05-2014 09:17 AM
Free Book (Kindle) - ?Como salgo de mis deudas? (UK Only)	koland	Deals and Resources (No Self-Promotion or Affiliate Links)	0	01-20-2011 09:14 PM
Free Book (Kindle) - Marked	koland	Deals and Resources (No Self-Promotion or Affiliate Links)	1	09-15-2010 11:22 AM
font mis-representation in Mobipocket Reader	cyberbaffled	Kindle Formats	2	06-24-2010 12:02 AM

10-13-2015, 05:42 AM	#2
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	update, well I have sort of diagnosed it: staring closely at the text, some letters did not look quote right e.g. look at any "u" in the image. so using copy + paste + replace all , I replaced all the "u", "v" and "a" letters with normal equivalents. But I am still mystified how it made it thu amazons validation & release procedures while full of weird alternative characters. the book blurb says it is a hot seller in ibooks so maybe the author messed up moving it from there to amazon ?

10-13-2015, 07:30 AM	#4
Doitsu Grand Sorcerer Posts: 5,685 Karma: 24031401 Join Date: Dec 2010 Device: Kindle PW2	I did a quick spot-check and found out that a couple of letters have been replaced by similar looking letters from other alphabets: a = а (Cyrillic small letter a) o = ο (Greek small letter omicron) u = υ (Greek small letter ypsilon) v = ν (Greek small letter nu) Since they were consistently replaced throughout the sample, they're probably not used for watermarking, unless the author replaced other letters in the epub versions sold by other sites. Whatever the motivation, it's a stupid idea, because it'll prevent readers from looking up words.

10-13-2015, 07:53 AM	#5
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	yep that is it. they are in the amazon format original like that also. I can't come up with a scenario for how that could happen accidentally though unless the author is Greek or has a Greek keyboard??? my first thought was that the language definition in the book's metatdata was messed up, but that looks OK

10-13-2015, 08:15 AM	#6
DiapDealer Grand Sorcerer Posts: 28,364 Karma: 203720150 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	My guess is intentional artistic stupidity license.

10-13-2015, 10:35 AM	#10
DiapDealer Grand Sorcerer Posts: 28,364 Karma: 203720150 Join Date: Jan 2010 Device: Nexus 7, Kindle Fire HD	Nope. I never noticed! Just goes to show you: it's only an error if I notice it.

10-13-2015, 10:46 AM	#11
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	in this instance, going back to the original source on my Kindle, the a and the o are indistinguishable, but the strange lower case u is easy to see , and annoying once you know it's there. the "breakthrough" was from thinking what do all the mis-spelt words have in common and the non-misspelt ones don't. so once I fixed the "u" and some of the words then became OK, the hunt was on

Advert

Advert