I spent a couple of weeks tinkering with a shell script that would
download the 10000 vital articles from wikipedia and turn them
into ebooks. My book reader is quite old, so I turned the html into
txt and then back to epub in order to get out something usable.
For whatever reason I can't get a version with images out of
either pandoc or calibre's ebook-convert.
So, now I have some 130 books. Since the book reader doesn't have
a folder structure, this pile of wikipedia is drowning all the
other books I have. And then there is this feeling of being
overwhelmed. Without a way to search or to link from book to
another it seems just too much. And now I am not even reading
the books I was reading before I started this project.
I wonder what is the amount of information that makes going
through it seem manageable? I bet it is somewhere close to the
size of an average book. Maybe I was aiming too high. Maybe I
should try to get the 1001 vital articles. Maybe only their
summaries? It seems there could be some python way of doing this
that might save me a lot of sed pipes.
Well, anyway, just giving this example as a possible avenue for
some investigation. It seems to me that the learning that you can
get done with a traditional book is qualitatively different than
what can be done online by just searching blindly. So, the vital
article list seems like a very useful place to start. It is
basically like a huge book, since it has a structure that comes
about by the selection process. But it is just too big to be
squeezed into a handy ebook. So, there is a need for some
experimentation.
Maybe I could first of all try those summaries and see how many
can be packed into a regular book. I could have several
perspectives of summaries in different books (like people, math,
history etc). The amount has to be kept small enough and the
scale large enough so that an area of study is illuminated
"completely" but dimly. Then, the second round of books could
go into more specific areas.
