PDF haters? Perhaps they exist. But if so, the only reason you hear of them is because they seem to be so much louder than the host of PDF lovers, at least if the Manybooks.net format statistics are anything to go by. In the past few days, 4,500+ PDFs were downloaded from the Manybooks website, with eReader coming second with less than two thousand books downloaded.
We’ve talked about Manybooks.net before; a website run by Matthew McClintock that distributes Project Gutenberg and Creative Commons licensed ebooks, presented in a wide range of (automatically generated) formats: eReader, iPod, iSilo, iSiloX, Librie, Palm Doc, PDF, Plucker, Rocketbook, TCR, and zTXT.
Matthew tells us that he has no idea why PDF is more popular than all others combined. In an e-mail conversation he writes: “I’m not sure what the attraction is with PDF, to be honest. I added it because it was easy for me to do, but I didn’t think folks would go for it — I thought the PDA-centric formats like Doc, Plucker, or iSilo would do better. eReader makes a little more sense, as it’ll run on Desktop (PC or Mac), Pocket PC, or Palm about the same, but PDFs just don’t work well on the handheld devices I’ve used.”
“I forgot to archive the old format count page before resetting the counter [on Jan. 1, 2006], but they were essentially the same ratio,” Matthew continues, “PDF is, and has been, the most popular, with 2 or 2.5 times the downloads as the next most popular, eReader. The surprising change with the count reset this year is Plucker’s slight drop in popularity, if I recall the old statistics correctly — perhaps it’s due to Plucker format availability on gutenberg.org?”
A poll is now running at Manybooks.net’s homepage to help determine what device visitors use to read ebooks with. Could it really be? Are desktop PCs being used as ereaders?
Thanks for the info, Branko. Methinks that many folks could be printing out the PDF files–a GOOD use for PDF. Also, Matt lacks the software to do Mobipocket format, which I suspect would steal away many fans of “pure” e-reading. – David
maybe they’re (ugh) printing them out!?
I’ve heard that this is done.
If one looks at the list of formats, the only one which is PC-based (Windows/Mac/*nix) is PDF (it’s also designed for printing) — all the others are designed for PDA or dedicated ebook hardware. (Of course, a couple of these formats have ported readers for Windows desktops/laptops, but few people know these exist.)
Also, to be sure, I downloaded a PDF and noticed it was formatted for printing — definitely not optimized for reading on PDA devices. Also, that PDF was not tagged, so it’s not as “reflowable” as a tagged PDF would be.
Many ebookers tend to assume that the overwhelmingly vast majority of readers of ebooks do their thing on PDAs and dedicated ebook readers. I believe this to be incorrect — the percentage of ebook buyers who use a laptop or desktop is higher than many believe, although I don’t know what it is (I’m one of them!) Also, I know that a lot of people buy PDF ebooks for printing and reading them the old-fashioned way.
In fact, when I added my datapoint to the Manybooks.net poll on what platform I use to read ebooks, about 20% of the respondents noted desktop/laptop, which is fairly high given this site is unabashingly oriented towards those who read ebooks on PDAs and dedicated ebook readers.
As an aside, I’m perplexed why Microsoft LIT is not supported. It’s easy to generate (I’d be happy to share my OEBPS to LIT converter with Matthew) and it is readable on both Windows mobile devices and regular Windows desktops and laptops. It need not be DRM’d, if that is a fear. And people need not activate Reader on their PC to read non-DRM LIT files. (If Matthew generates OEBPS, it’s then possible to generate the Boo format for the Cybook, I believe — need to talk with Laurent Picard on this.)
There is absolutely nothing wrong with pdf when it is used for the applications it was designed for. Preserving formatting for paper documents was its original design goal.
You can create tagged text documents in pdf that will reformat gracefully on devices with constrained displays, but there are definitely better formats for that.
One e-book application pdf is well suited for is image files of printed documents. I have a number of old paperbacks and pulp magazines that are falling apart and are unlikely ever to be reprinted. Scan them in as image files in pdf, compress a bit, and they read great on tablets.
Ah, yes, print! I have heard of that! 😀
You’ve all raised some interesting points – and I’ll follow up with Jon for details on how to generate LIT (that’s a re-occuring request from visitors!). It was my understanding that LIT could only be generated on Windows, but if there’s a way to do it using linux I’ll do it 🙂
As for Mobipocket, I’d be glad to add that as well, if I was able to do it with Linux software. As far as I know, however, it’s only possible using Windows (my site runs on linux, and I don’t have the time or interest in running Windows at this point). I’ve written to the Mobipocket folks a few times, and their response has always been that they’re “considering the creation of a linux publishing tool.”
But I should probably update the “Help” section of the site to let folks know that eReader can be used on the desktop, too (there are desktop apps for Windows and Mac), even if most folks just stick with what they’ve already got — I’m guessing that most folks are familiar with PDF, and will just pick whichever format they already know rather than look into the alternatives… At the same time, I feel like the site should cater to what folks are looking for rather than what I think they should use, so it’s a bit of a bind.
Once I’ve got a clearer idea of which devices folks are reading with perhaps I can reformat the PDFs to suit a little more closely, with a version for PDA use in addition to the print oriented version. Honestly, when I added PDF it was as an afterthought — it’s just not on my radar as something that’s terribly useful on a PDA!
Unfortunately, Matthew’s assumption is correct. One must use a Windows system to generate LIT since the underlying engine for producing LIT is litgen.dll which is not available as source code; it also requires MSXML for validated parsing.
The key, however, is to produce an OEBPS 1.0.1 version for each ebook (e.g., conform each ebook to XHTML 1.1), and, as needed, add a couple MS-specific things to the OEBPS Package. At that point it becomes “push button”, and maybe someone can offer their Windows box to do the conversion (I’d do it if the number of ebooks numbers less than one hundred or so.) I’ll be glad to advise/help on this.
Ah, well I was afraid that the Windows requirement might be a sticking point — there are 12,000+ files, and the current system builds them on command (when the first user requests the file, which is then stored for future use)… so it looks like Mobipocket and LIT might have to wait until I’ve got a Windows server at my disposal. Thanks for the offer, though!
O.k., let me know when you are able to do the mass conversion to Microsoft LIT (and before that, to generate valid XHTML 1.1.)
Manybooks.net is a wonderful resource. Keep up the great work, Matthew!