Problems with Sigil Files on Nook BNRV500

[MoReader92019]

Hi Calenorn, DNSB:

I deleted all files from Nook (including annotated files), copied in my epub files back in and let Nook do the rest. The file that was used to make Trump’s Tariff was still in generic book form. When I booted it Trump’s Tariff came up. Nook is confusing the original with Trump’s Tariff. That has happened before. I thought I might fix the problem in Trump’s Tariff. Whether starting from html or using a template published ebook, I still run into these problems. However, even in html generated, I copy over from templates to conform to the epub validator OK. The html generated files have inline stylesheets, and page after page of epub validator errors, so many that it is depressing and overwhelming. The object code must have something embedded that is being copied over. This leaves me the option of hand typing a file from a model to hopefully get clean object code.

I checked Metadata and OPF in the two mixed up files studied so far and find nothing out of sort. The Navpoint ID in Trump’s Tariff are different and more complicated than the original. I either generated the file in Sigil or may have simply altered an ncx in place. I cannot remember what I did. Whatever I do, this thumbnail and slow boot problem remains, in spite of epub validation showing no errors. Any hints would be helpful to get a quick fix.

[MoReader92019]

I can also nuke or text my copy over stuff before entering it into a scratch file to make my own clean (sic) copy and get away from doc file problems. I thought maybe Sigil would not have that problem of embedded stuff, showing it in html. Of course, that is wishful thinking. Always looking to get things the quick and easy way.

[DiapDealer]

To my knowledge, Sigil simply doesn't have the ability to make an epub "load slow." If it validates with epubcheck, then it has to be a Nook quirk that needs accounting for.

Perhaps it's the cover image itself? Many rendering systems have issues with how jpegs get created/saved (compression, optimized/not optimized, progressive/not progressive, smoothing, exif/xmp data, etc...). The size of the image in the attachment you added itself is quite excessive. It alone is much larger than most full novel-sized epubs are altogether.

[MoReader92019]

Well, yeah, but if the object code Sigil saves has embedded information that confuses Nook, what can Nook do but slow down, make an if mandated decision to go with a default thumbnail? And if two thumbnail options are there, what does Nook do, a timeout pick and choose? I'll just start from scratch, learn how to do epub right and upload a clean copy of my files that cannot possible have embedded stuff from previous files. I have only tried this with the last several of my files anyway. My last epub used Smashwords with a few flaws that I could have fixed with Sigil, but even these seem to fool NOOK.

[DiapDealer]

Originally Posted by MoReader92019

Well, yeah, but if the object code Sigil saves has embedded information that confuses Nook, what can Nook do but slow down, make an if mandated decision to go with a default thumbnail? And if two thumbnail options are there, what does Nook do, a timeout pick and choose? I'll just start from scratch, learn how to do epub right and upload a clean copy of my files that cannot possible have embedded stuff from previous files. I have only tried this with the last several of my files anyway. My last epub used Smashwords with a few flaws that I could have fixed with Sigil, but even these seem to fool NOOK.

There's no object code involved. Sigil saves a zip archive that contains xhtml/xml/css/image files. I'm not sure what "embedded information" you're referring to. I don't know the least little bit about how Nooks work, but I DO know Sigil. And there's no embedded anything anywhere (that somebody didn't put there themselves).

You have some crazy long file names that in turn result in crazy-long links, but other than that, there's nothing hiding in that epub anywhere. There's a couple of small html files, a reasonably sized css file and a gigundous cover image that's comprised entirely of text.




Maybe it's not the size, but some of the image metadata. Should be easy enough test. Replace the existing cover-image with a much, much smaller one (or get rid of it even) and see how the epub performs on the device.

5000x5000 ppi to display an image full of text is a bit much. Drop the resolution a little. I know for a fact that some RMSDK-based renderers will barf on images that have a side much longer the 2000 pixels. Your image is 2050x3050 pixels. I don't know how much ram a Nook has available, but this image is occupying way more of it than it needs to.

EDIT: I also notice that image itself is not tagged with the "Cover" property. That's not spec-related so much, but perhaps the Nook relies on that particular bit of OPF metadata?

[Calenorn]

A jpg 600 x 400 pixels makes a fine thumbnail for the NOOK.

[MoReader92019]

Calenorn & DiapDealer,

Thanks for your info. I have nothing to do with the size of thumbnails. I must meet Smashwords requirements for images and these far exceed 600x400. It is Nook that deals with that as the case warrants. I use 5000 resolution put in by Gimp because my covers always look like crap, even just text files at default 72, and not so great at 5000 either. By crap, I mean resolution, not my primitive graphics. I have never met a graphics program that I like and lack tablets, pens and quality resources. I do not know why other peoples covers look great and mine never do whether with or without images. When I used OpenOffice as a source for epub, they would always lower my resolution to conform to 200 dpi print even though it is an ebook, ruining my files. When I turned to print, Kindle would always reject my images due to less than minimum 300 dpi. No matter what I did with changing the resolution in the print/print file of Writer, OpenOffice still sent Kindle 200 dpi. Now I use Word and now have a problem with xfering my OpenOffice styles. Office is worse than before after each upgrade. Why OpenOffice is screwing with what is basically an image matter, and not leaving dpi to the print function of my printer is their business. Microsoft seems to control images too as to what is passed along. My graphics always look bad. I have no control over this with the software that I have. It does not matter what I do in Gimp, it still fails. Because of small sizes of readers, I am limited to about 550 width for them to fit in the screen. That does not include the cover image which is bigger and conforms to the reader page as required as the sole image there. The trouble is I have more of a textbook subject. I now avoid graphics and use word pictures for the most part. My graphics are never able to give me the resolution to do this. I do not have good drawing or editing equipment or apps. They cost money and very little money comes in. My readers prefer free stuff, and frankly so do I. Thousands of downloads come to just about nothing. I have also been hacked a lot and must consider tampering of apps at all levels. Well, try and prove it. Slow torture by degraded performance is much more irritating than just a crash. Who knows? Equipment breaks down, intermittent disruption, decaying drives, hackers too clever for my defenses.

I will reduce the size of my cover images on my own files to see if Nook fits them in then, if that is the problem. This is easy enough in html, from the source 2000x3000.

My Nook has taken some falls too and may not be working quite right.

As for object code, I am simply saying that all computers compile programs whether in the old fashion way or on the fly. Those files you are talking about are just a bunch of binary bits and bytes. That is how computers work. If these binary bits and bytes get into a file cross-eyed, corrupted, then you need a debugger to detect them, remove or alter them, and to know where to look.

Thanks for your time. Hope you got something out of it too. I think the matter is at an end. I will deal with things as they are and come. There is always letting Smashwords deal with this, and just get my Word styles to where they need to be. Most of my problems on that score were settled years ago. But I wanted to go direct epub to learn the ins and outs. That takes time. So long.

[DiapDealer]

Originally Posted by MoReader92019

As for object code, I am simply saying that all computers compile programs whether in the old fashion way or on the fly. Those files you are talking about are just a bunch of binary bits and bytes. That is how computers work. If these binary bits and bytes get into a file cross-eyed, corrupted, then you need a debugger to detect them, remove or alter them, and to know where to look.

That's not object code by any definition I'm aware of. There simply is no compiled code in an epub. But I won't argue that a corrupt file can't cause havoc.

And I was criticizing the cover image artistically. Its size and resolution just seem like a ton of overkill to display what is basically text. I can easily see resource-limited devices (I've no idea what the specs are on your Kobo, by the way) needing to chew on something like that for a bit when loading.

[DNSB]

I took a quick look at the epub and made a few changes. Among others, I shrank the cover image and split the text files into separate files for each segment. Could you check it on your Nook and see if it is any happier.

[MoReader92019]

I erased all references to any Trump's Tarrif in all files on Nook. DNSB's file did run and booted fairly quickly. The Thumbnail showed up in Library and Recent Read. Next, I saved DNSB's file under a different name and saved it on my computer with my Sigil, but not making any changes. I cleared my Nook and uploaded to the Nook. The file worked with thumbnails showing.

So! The book is very small for so many files? Smashwords often breaks books into smaller files (those very long names come from Smashwords that DiapDeaer was complaining about. I modify files often and saves time to not change things.) But often files are broken in funny places. My Table of Content title shows, and then a page skip to show the content, and page breaks whenever a new file is jumped to. Hey, live with page breaks? I read lots of epubs with them. Yeah, I get a hair up my ass on little things. I know long files run better with multiple files for the body of the book.

OK, I will look the file of DNSB over carefully, compare it to my Smashwords files that also work, and then to mine that do not. What can I say? It takes years and years of experience.

Thanks for the help: Roger, Wilco and Out