more experimentation on "gods and fighting men".

i ran out a .pdf version specifically for the iphone:
>   http://z-m-l.com/misc/14465-320*440.pdf

the iphone has a display that is 320*480, but i
shortened up the height because when you read
a .pdf in safari, there's a bar across the bottom.

speaking of which, the .pdf reader in safari sucks.

the first problem is that the .pdf is presented in
one long scroll window.  it would be much better
if it was _paginated_, and a left/right swipe would
navigate from page to page, a la iphone typicality.

the second problem is that you cannot advance
the scrollbar in long jumps to remote locations;
you must move through the .pdf incrementally...
when a .pdf has 530+ pages, like this one does,
that becomes a significant pain in the derriere...

i'd hope that other .pdf readers (a la goodreader)
have overcome these constrictions in their apps...

***

so now we have three different versions...
>   http://z-m-l.com/misc/14465-320*440.pdf
>   http://z-m-l.com/misc/14465-5.5*8.5.pdf
>   http://z-m-l.com/misc/14465-8.5*11.pdf

aside from some differences in the _margins_ and
the leading within the text-block, these .pdfs are
extremely similar.  they each have the same lines
on the same pages, and that is easy to recognize...

what you will find, moreover, is that the _pagesize_
doesn't make much difference at all -- almost none!

if you pull these 3 .pdfs into the same view-port,
the "zoom-to-fit" display option will make them
appear to be quite similar.  indeed, on the iphone,
which has smarts that ignore whitespace margins,
the different versions appear virtually identical...

reflection shows that this is what i was talking about
when i discussed how the "pocket dictionary" and the
oversized version shared the exact same pagination
as the "regular" version.  the pages are just "zoomed"
or "reduced" as necessary to fit the paper being used.

so, if you create your .pdf _correctly_, it can be used
on a wide variety of screensizes without any problem.
the secret, however, to making the .pdf "correctly" is
to make sure the fontsize being used is large enough.

when i created the 8.5*11 .pdf, i had to use a fontsize
of 17 points, which will seem "too big" to many people.
it's not "too big".  it creates a 66-character line-length,
and the resultant .pdf is viable across many screensizes.

and this is why the .pdf that most people create is bad,
and not useable on smaller screens.  they use a pagesize
of 8.5*11 with text that is sized at 11-point or 12-point.
that creates lines too long for best readability, and text
that's too small to read when shrunk to a small screensize.

***

however, even though i use phrases like "best readability"
and "too small to read", the most important takeaway here
is that digital text gives people the ability to make a .pdf
that they can _customize_ to their own unique preferences.

questions like "what pointsize gives best readability?" and
"what's ideal line length?" and "what should the leading be?"
all presume that there is _one_ answer that applies to all...

that's a presumption that has some validity when we are
creating _one_ version of the book that _everyone_ reads.

but that's no longer the case, thanks to print-on-demand.

now, everyone can make their own version of the book, and
the only "ideal" variables are the ones for _you_personally_.

some typographer might say that garamond is the "best font"
for this book.  but if you like palatino better, use palatino!

and who cares if most people prefer 12-point type the best?
if you like 11-point, use that to print your copy of the book.

if you want 72-character lines, fine!  66-character lines?  ok!
54-character lines, no problem!  40-character lines, great!

4-inch lines, fine!  or 4.5-inch, 5-inch, 5.5-inch, whatever!

the "standard" leading is 120%.  but if you like 150% better,
and you don't mind that it'll take more pages to print it out,
then go for it!

***

there is one caveat, however.  it's fine for you to create
your own one-off of a book set to your own preferences.

but don't inflict that book on the world at large, please...
keep it to yourself.

the world doesn't need 14,926 versions of every book.

because here's the problem.  some of those versions
will have errors in them.  and other versions will have
other errors.  and other versions will have still others.

sorting out these 14,926 versions will be a nightmare.

that's why -- in my first experiments with this book,
i retained the p.g. linebreaks.  there's no reason to
introduce new linebreaks to the world unnecessarily.
it just complicates things, for very little good reason.

as long as you retain the p.g. linebreaks, it'll be easy
for users who wanna track the provenance of the text.
once you introduce different linebreaks, even if the
text is the same, ease of comparison is jeopardized.

so go ahead and change the linebreaks for yourself.
but don't inflict that changed version on the world...

***

of course, there's little reason that project gutenberg
should be introducing a new set of linebreaks itself...

it would be far better if it kept the p-book linebreaks.

those paper-books are already out there in the world.
they are the canonical version we should adhere to...
their linebreaks should be the linebreaks that we use.

so let's take a look at the paper copies of this book...

internet archive has scanned four copies of this book.
here are scans for pages 122 and 123 in those books:
>   http://z-m-l.com/misc/godfm00gregrich122.jpg
>   http://z-m-l.com/misc/godfm00gregrich123.jpg
>   http://z-m-l.com/misc/godfm00greguoft122.jpg
>   http://z-m-l.com/misc/godfm00greguoft123.jpg
>   http://z-m-l.com/misc/godfm00yeatgoog122.jpg
>   http://z-m-l.com/misc/godfm00yeatgoog123.jpg
>   http://z-m-l.com/misc/godfm01yeatgoog122.jpg
>   http://z-m-l.com/misc/godfm01yeatgoog123.jpg

as you can see, each of these scan-sets is consistent;
linebreaks and page composition are the exact same.

character-count on these lines is about 54 characters.
page 122 and other full pages have a line-count of 36.

my next experiments will be to reproduce this format.

jim, if you want to make yourself useful, you could do
a reformatting of the project gutenberg e-text so that
it matches the linebreaks from one of these scan-sets.
i will give you a couple days time to pull off that task...

-bowerbird