keith said:
>   BB I can not believe you are serious.

is that so?  because i find your disbelief to be quite humorous!  :+)


>   1) Your critic fails all logic.

it fails _all_ logic?  i have a hard time believing that, keith...        :+)


>   Why in Gods name would anybody intermix scans
>   from more than one book in the same directory.
>   Their are more than enough files just from one book !

i wrote that huge post, and _that's_ what you took from it?

talk about missing the point.  you missed it by a mile, keith.
(a mile is about 1.6 kilometers, in case you are wondering.)

for the record, not that i think anyone else missed the point,
it might not be that you'd _want_ to put more than one book
in a directory, it's that you _could_ if you ever _did_ want to,
whereas, when all books are named p001-p999, you cannot.

the more important point is that, given the files for a book,
and for another book, you wanna be able to tell them apart.
all files for a book should be named with a common element.
and the name of every file should be unique from all others,
across your entire system.  this is nothing but common sense.


>   2) How is a sequence of five arbitary characters anymore
>   informative.  Or can you remeber 26^5 titles.

the characters are not informative in and of themselves, but
they become meaningful when all files from a book receive
the same prefix, because then you see, just from the names,
they go together.  and no, there's no need to remember them,
since the catalog will keep all of the information straight and
make the appropriate information available to the end-users.

but i suspect you knew all that.

***

but in order to see how someone might do it another way,
go look at the internet archive and their naming convention.

they went for longer names, hoping for _some_ meaning...
and, to a degree, they attained it, at a cost in convenience.

for instance, here's a subdirectory name:
>   http://www.archive.org/details/adventuresoftoms00twaiiala

that subdirectory maps onto another more-specific one:
>   http://ia331317.us.archive.org/1/items/adventuresoftoms00twaiiala/

so their "name" for this book is "adventuresoftoms00twaiiala".

therefore, you might guess -- correctly -- that this book is
"the adventures of tom sawyer".  but it doesn't inform you
_which_ edition of the book this is, or where it came from,
or if it is one of the several copies from project gutenberg,
or when it was published, or any number of details about it.
to get to that information, you'll have to visit their catalog,
and if you're gonna visit a catalog anyway, you might as well
visit the catalog to find out the 5-letter "prefix" of the book,
a prefix that's much easier than "adventuresoftoms00twaiiala".

and you better believe me, because it has happened to me,
once you get a lot of the archive.org files on your machine,
it starts to become very hard to discriminate names such as:
>   http://www.archive.org/details/adventuresoftoms00twaiiala
>   http://www.archive.org/details/theadventuresoft00074gut
>   http://www.archive.org/details/theadventuresoft07193gut
>   http://www.archive.org/details/theadventuresoft07194gut
>   http://www.archive.org/details/adventurestomsa02twaigoog
>   http://www.archive.org/details/adventurestomsa00twaigoog
>   http://www.archive.org/details/adventurestomsa00willgoog
>   http://www.archive.org/details/adventurestomsa01twaigoog
>   http://www.archive.org/details/adventurestomsa05twaigoog
>   http://www.archive.org/details/tomsawyer00twain
>   http://www.archive.org/details/adventuresoftoms20twai
>   http://www.archive.org/details/adventuresoftoms99twai
>   http://www.archive.org/details/adventuresoftoms00twai2
>   http://www.archive.org/details/tomsawyeradv00twairich
>   http://www.archive.org/details/advtomsawyer00twairich
>   http://www.archive.org/details/booki-export-the-adventures-of-tom-sawyer

so, for me anyway, a 5-letter prefix seems to do the job just fine.

***

likewise, we can look at the system used by project gutenberg,
where the "prefix" for the book is essentially its 5-digit name.

digits are, in some ways, even more convenient that characters.
the problem is, 5-digit names only work up to 99,999 books...
that's enough for now, for project gutenberg, so that's fine, but
i wanted more breathing room, so i chose 5-character names...

***

or let's take a look at youtube names.  here's a sample u.r.l.:
>   http://www.youtube.com/watch?v=sA_0cvd1EUM
>   http://www.youtube.com/watch?v=qybUFnY7Y8w

first, i'm not sure why they need that "watch" in every u.r.l.
surely "watching" a video would be the default action, not?,
so it seems to me they could have abstracted that out, but...

we find they're using an 11=character name, one that uses
_both_ uppercase and lowercase letters (i only use lowercase),
_and_ numbers, _and_ at least some other characters as well.
that's going to give them _many_trillions_ of possible names,
which i guess is how high you think if you sell for $1.6billion.

***

speaking of google, let's see their book filename convention:
>   http://www.google.com/books?id=3n4hAAAAMAAJ
>   http://www.google.com/books?id=Y7sOAAAAIAAJ

they've got a 12-character name, uppercase and lowercase,
plus numbers.  which, again, will accommodate lots of files.

***

>   Come On Man! Wake up.

well, it's after midnight my time, so i'm about to go to sleep;
but i will wake up tomorrow morning, all ready to post again.

-bowerbird