Terry Given wrote:
[...]
its a scanned image PDF. I use a neural net to do my searching of books
(I read them, then remember whats in them and roughly where).
I used to be able to do that with books. But now they are disk files and
there are too many
Besides, when you are looking for info on a tone control circuit, you
automatically know it will be in the thick, short red book called
"Radiotron Designer's Handbook". Filter stuff will be in the huge medium
thickness book by Zvrev, perhaps also with a red cover. So the
associations are easy to make.
But it gets much more difficult when you have hundreds of thousands of
files scattered through thousands of directories. They have no color or
weight, and each one basically looks exactly the same as the others until
you open them and start browsing.
So I wrote a special operating system that runs in dos and attaches a
descriptive phrase to each file. Yes, hundreds of thousands of them.
Fortunately, the software can sift through html, c, pascal, and plain
text and extract some meaningful information about the contents. But I
can override that whenever I wish.
The description goes with the file when it is moved or copied to another
directory. I then generate an index that contains all the descriptions,
and can search for a key word or phrase that appears in the index. It
takes just under two seconds to find all the files on my hard disk that
match the search terms. I can hilite a file with the cursor and scroll
through the list.
When I find a file that looks interesting, a single keystroke takes me to
the drive and directory where the file is located, and places the cursor
of the main program on the file. Then I can read it with any of a number
of different programs, depending on the file type.
The program keeps the files separate by naming them with a hexadecimal
representation of the current date and time. Since this never repeats, I
can copy or move files anywhere without worrying about overwriting
another file with the same name. Except, of course, if it's app notes by
different companies that happen to have the same name. The program warns
me this is about to happen so I can check if I really want to do the
operation.
But I end up never having to type in a directory or filename. And the
system is hundreds of times faster than any gui OS - Windows or Linux!
I am not lucky, I am sneaky. I contract to a company in Auckland, NZ (1
week per month), and that contract specifically allows me to bill them
the NZ$90/month for my broadband internet connection. Otherwise, it'd be
dial-up for me too
Sneaky. I'm using Win 3.1, so there's no way I could get broadband. The
necessary api's won't run on my machine, and there is only one person in
the world still using it so nobody is going to write anything for it.
Including viruses. Yea!
And Trojans. Yea!
And popups. Yea!
on a related note, few things are more annoying than boneheads who think
giant images are necessary for websites. Or worse still, animations.
Cheers
Terry the luddite
Yes - let's hear it for Flash! It won't even run on my machine
I am still astonished at the way people will make 350k gif files that
could be shrunk to 10k if they knew how. Or 2 page pdf files that are
mostly text but take up 3 megs. These people should be given a DOS
machine and floppies for a year until they figure out how much this costs
the other people who have to download and store it.
Mike Monett