Something Sigil Can't Do?!?
 

A friend asked me to clean up his 26 MEGABYTE epub!!! I said alright figuring it was just a bunch of enormous images that could be squished down to a reasonable size and some nasty bloat like:

<div class="paragraph"><p class="p"><span><font class="normal">Blah Blah <span class="italic">BLAH</span><span class="normaltext">!</span> And furthermore <span class="boldfont">Blahdy Blah</span><span class="normaltext">...</span></font></span></p></div>

I could easily make into:
<p>Blah Blah <em>BLAH</em>! And furthermore <strong>Blahdy Blah...</strong></p>


But nooooo......what do I find......

307 .png images of each page of a book!!!

I would certainly HOPE that a "reputable" company would not create such a mess....or maybe it was someones miserable attempt at a fixed layout?? In any case, I told him to get his money back and buy a different copy from a different bookseller...maybe one that was a reasonable <1Mb file.

Soooo.....can we get Sigil to add a button/function to OCR all the selected images and put them in proper html/epub-ese??? Pretty Please???:rolleyes:
[/sarcasm]

Regretfully honesty, integrity, knowledge, and workmanship has slipped by the way in our new progressive world.

When a client has obsessed on getting an absolute fascimile of the printed layout, tested the epub on every possible device and complained about every minor difference...I'm sorely tempted to do just that.

What is the problem with people not wanting to do pdfs in these situations? Sized for the smaller windows, they will have the exact layout. They sure are a heck of a lot less work...a printer driver will do anything that can be printed on paper.

When making a publication for tablets, why bother with epubs? So much work for so little gain. Must be the fad of the year, I guess.

Don't get me wrong, I love epubs in my Sony Reader whose battery lasts a week or two between charges and is scarcely larger than a trade paperback and holds 300 books. For the others, as I said, why bother?

In the above case, a text pdf would not have been gigantic either, but the disreputable so and so's would have just done what they did to make a giant epub, make an image only pdf.

You're sure this was a stock retail epub?

Sounds more like someone took the raw SVG "pages" of a DeDRMed Topaz ebook and tried to make a "quick & dirty" ePub out of it (converting SVG to PNG), if you ask me.

No, I'm not "sure". I just opened it in sigil, saw the single XHTML page with links to each of the images and closed it. I didn't waste any time trying to figure out where it came from. I just started ranting about all the evils of substandard/shoddy workmanship and how he shouldn't "buy from that company" again! I told him to return it and buy from a different company. ;)

Convert all the PNGs to JPGs and you'll probably have a smaller epub ;)

Nah, jpeg will be larger. Palleted png will be really really light on file size (if it's just greyscale, 8 colour palette is good enough), and gif a few bytes smaller.

But none of those matter as they are always going to be decompressed for reading, and as such eat up a whole load of memory and be generally crap for performance (most 'good' readers cache a few past pages and the next page or two.). Problem readers, webkit-based reader apps/readers(since they don't really paginate) and some ADE bugs however will mean that the images just sit around in memory all the time, and depending on the reader can starve the rest of the OS/framework leading to sudden reboots or hangs. Fun times. Also really bad for load time.

Probably because PDF's cannot be sold as "ebooks" through any of the major retailers, would be my guess.

I don't know of any reputable conversion house that would have done this, but I do know of several lower-tier scanning houses that would output something like this from Abbyy as an "ePUB" or "eBook" for a client. I could (and won't) name at least two from which I've seen/received/fixed similar work.

Hitch