US5584023: Computer system including a transparent and secure file transform mechanism
Posted 23 Apr 2002 at 11:30 UTC by lkcl
came across this patent. it claims a patent on any transparent
mechanism - encryption, compression, encoding, translation and
and presumably also virus scanning and blocking - between a computer
and its storage media.
what are the implications for open source?
delphion URL, posted 23 Apr 2002 at 11:32 UTC by lkcl »
If you intend to fully comply with all relevant patents in your free
software work, then you will not get any work done, because it will take
all of your time and then some to study the patents which might apply,
determine whether they do apply, decide which ones are unenforceable due
to prior art, and so on.
Seriously, even if you spent 40 hours a week on the "patent
compliance project", there are probably new patents being filed so fast
that you will not have finished analyzing all possibly applicable
patents for decades (by which time the first ones that you started with
will have expired).
Stop wasting your time getting all excited about whatever patent
randomly happens to cross your radar and go back to coding please.
5584023 i would hazard is much more general. it refers to "transform"
mechanisms which would include stacker, doublespace, microsoft's
encrypted file system, linux loopback filesystem, open source
transparent encrypted fileystems - the whole works.
6185681 appears to be specifically related to applications _on_
computers, and additionally covers exactly what 5584023 covers
[transparent "transforms"] but only cryptographic transforms.
i think this is fascinating.
zooko, i assume that you are talking rhetorically and
political-soap-boxey. you can't be referring to me: i have not written
a single line of open source code for over six months, and am proud of
the fact that i am such a scary person that no US-based corporation will
sponsor me to do open source coding. enjoy! :)
You're right, I meant "THE READER" as the one who should stop
freaking out about random patents and go write code.
You (lkcl in specific as well as "THE READER") are welcome to
contribute code to the Mnet project
but we're definitely not going to give you money.
mounts, posted 23 Apr 2002 at 23:19 UTC by splork »
: break out of the "linux == innovation" mindset
(its not true in most cases). loop mounts have been in BSD since the
early ages. Also look at BSDs union mounts (mounting one FS on top of
another; getting some combination of both through).
This patent is so bogus as to be completely laughable,
and there's tons of prior art. Authors of open source
programs shouldn't worry about it. Those that are nervous
should just look for techniques that existed before the
filing date: patents can't protect prior art.
The very first Unix filesystems violate the patent: you can
read the disk device in raw mode, or you can read it via
the filesystem interface. That's a transformation.
"But I can't take that kind of legal risk!" you say.
Fair enough; in that case, you must halt all open source
development, because there are thousands of bogus, extremely
broad software patents on the books, covering so much that
almost every significant program you write will infringe
on at least one.