[IAEP] The Children's Library On OLPC project
acahalan at gmail.com
Sat Jul 25 04:37:07 EDT 2009
Jim Simmons writes:
> A Journal entry consists of a file plus metadata. There is no real
> advantage in NOT storing the book in the Journal. You can convert
> whatever book format you're reading into a zipped archive of same on
> reading it for the first time then mark the Journal entry with Read's
> activity id. This would give the Journal entry Read's icon and make
> it resumable by Read. I do something like this with Read Etexts when
> it reads a plain text file. I'm not trying to save disk space in this
> case; I need to add a pickle file to the archive to store annotations,
> so I create a new Zip file and store the text and the pickle in it.
This encapsulation makes it more difficult for people to share
books with non-Sugar users. If a Sugar user provides a PDF to a
Windows user, Adobe Acrobat should recognize it. Likewise for
sharing with MacOS X and GNOME users.
Putting a bit of non-critical metadata on a file is not a reason
to be changing the file format. Normally an xattr would be used to
store this data. (hopefully the Journal is xattr compatible)
> The XO does not have enough disk space to hold hundreds of books as
> PDFs. Plain text files would work, but kids like pictures and I don't
> blame them. As I see it, the child should choose what books go on his
> computer for himself, and delete books when he has lost interest in
This all depends greatly on the PDF generation tool.
Most are not focused on producing small files.
Text should be stored as text. It should not have fancy kerning,
because this causes bloat from constantly specifying coordinates.
It should use a standard PDF font. The font should not be embedded.
The PDF should be compressed. (not just the images)
Images should be stored as JPEG with an appropriate compression
level. Computer-generated line art should be in vector format.
A recent PDF standard revision should be used.
More information about the IAEP