New patent, claims to store 1.28GB on a floppy :-)
Sachin Garg
The patent claims a new digital data compression method providing the
means to transport parallel digital data over a serial medium using a
"Frequency-Time Based Data Compression Method".
Read more at http://www.c10n.info/archives/439
It supports the compression, encryption, decompression, and decryption
and persistence of many binary digits through frequencies where each
frequency represents many bits. Am I the only one who couldn't
understand all this? ;-)
Sachin Garg [India]
www.sachingarg.com | www.c10n.info
Carsten Neubauer
the link to p. web services in your article leads to
a page containing the trojan Dldr.HTML.Agent.R,
perhaps you better don't link directly from there.
Greetings,
Carsten Neubauer
Sachin Garg
Carsten Neubauer wrote:
> Besides the obvious fun or nonsense in that claim,
> the link to p. web services in your article leads to
> a page containing the trojan Dldr.HTML.Agent.R,
> perhaps you better don't link directly from there.
Thanks, I explored a bit more, its just a tracker on that site that
causes this problem. Your antivirus/firewall says it's a trojan because
it tries to read the information your browser sends, more detailed than
ordinary trackers do. Nothing should be damaged or spied out.
btw, there is more than just fun in this. Most such patents/claims are
attempts to get money from people stupid enough to invest in something
they don't understand. Author will almost certainly try hard to sell
this to someone, hopefully that post will show up if anyone considering
this patent cares to google about it.
Sportman
> The patent claims a new digital data compression method providing the
> means to transport parallel digital data over a serial medium using a
> "Frequency-Time Based Data Compression Method".
Jeffrey Fries patent principle sounds similar as Jan Sloot's
principle with the difference that Jan Sloot used colors as frequency
to store in three dimensions, a sort of color math for digital coding.
Is it normal that the time between filling a patent (August 27, 2002)
and getting it (August 22, 2006) is almost 4 years?
Matt Mahoney
I read their patent no. 7,096,360
http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=7,096,360.PN.&OS=PN/7,096,360&RS=PN/7,096,360
Their "compression" method is multilevel pulse code modulation, e.g.
encoding 2 bits in one pulse by selecting among 4 frequency levels.
(They neglect such details as how you store analog signals on a
floppy.) They also claim an encryption method which is essentially a
Caesar cipher.
One wonders how it took the USPTO 4 years to decide there was no prior
art.
-- Matt Mahoney
Sachin Garg
Sportman wrote:
> Sachin Garg wrote:
> > The patent claims a new digital data compression method providing the
> > means to transport parallel digital data over a serial medium using a
> > "Frequency-Time Based Data Compression Method".
>
> Jeffrey Fries patent principle sounds similar as Jan Sloot's
> principle with the difference that Jan Sloot used colors as frequency
> to store in three dimensions, a sort of color math for digital coding.
Its a red flag common to many hoax claims and fraud inventors. They
refuse to use the established terminology of domain and try to use
their own confusing terms to jazz up their descriptions, and sound more
unique.
> Is it normal that the time between filling a patent (August 27, 2002)
> and getting it (August 22, 2006) is almost 4 years?
4 years does seems a bit too long, I googled a bit on this.
"The average prosecution time for a US patent is 3.4 years while the
average biotech patent is 4.4 years."
http://www.iploft.com/PTO-FDA.pdf
And
"The number of applications awaiting action by an examiner increased
from 107,000 in 1994 to 243,000 in 1999. In 1994, the average patent
application took 16 months to process. In 1999, the average processing
time was 25 months. The delay is worse in certain technology fields. It
took an average of 31 months to issue a patent on communications or
information processing technology compared to an average 26 months for
all technologies."
http://www.phil.frb.org/files/br/brq101bh.pdf says:
Now that its 2006, 4 years doesn't seems that long. Hopefully the
recent changes introduced/proposed by USPTO (like the wiki based
collaboration platform) will help speed up the process.
Sportman
> Their "compression" method is multilevel pulse code modulation, e.g.
> encoding 2 bits in one pulse by selecting among 4 frequency levels.
> (They neglect such details as how you store analog signals on a
> floppy.)
For example when a zero (0)=1070 Hertz and a one (1)=1270 Hertz and a
phone line frequency decoder needs 3,33ms to detect the frequency then
you can send 1/0.0033=+-300 frequencies (bits) in a second.
When a traditional phone line has a bandwidth from 300-3400 Hertz=3100
Hertz with Jeffrey Fries inventing you can transform 11bits data
(0-2047) in a frequency between 0 and 2047 Hertz and add 300 Hertz to
fit in the phone spectrum so 300-2347 Hertz. When this frequency again
needs 3.33ms to be present at the line to be detected well at the other
side, now it's possible to transmit 300 frequencies*11bits=3300bits
in one second.
This is the primary working of his inventing, it's clear that this
method work fine for lines where frequencies can be used to send more
bits in a time period but the transformation from bits to numbers
(frequencies) and storage of this numbers on a storage medium shall not
(or almost not) give any gain.
The second working of his inventing (where the trick must be) is only
named at 2 sub 25; "storage of massive amounts of data by storing
frequencies with each frequency representing a collection of data"
and at 7 sub 124 where figure 8 is explained; "These transitions (one
frequency to the next frequency) are used to synchronize an internal
clock. The internal clock is used for repeat frequencies. If the
frequency is repeated, then the data it represents is also repeated.
This occurs for the number of frequency detection periods as is
necessary to indicate that data is repeating...."
At figure 7 (sheet 7) of his patent the compression is done before the
frequency encoder (14) and the decompression after the frequency
decoder (18) and round the persistent storage (124) is not written that
there compression is used. Only at the decompression site 5 sub 1 and 8
sub 65 (f) is said; "The bits are adjusted to xx bits (if necessary)
either by truncating or zero filling the most significant bits..."
"...until the numbers of bits zero filled are equal to the said data
gathering rate".
Figure 5 look not to match the description at 6 sub 25, some
frequencies are reserved as control frequencies 5 sub 1 and 8 sub 20
for signaling like start of data, end of data, data value zero etc. In
this figure I see repeat 1 and 2 frequency what must be the signal to
repeat a part of the data related with that frequency but I can't
find a explanation about it. There is also a little mistake in 10 sub
25 (j and k) where is referred to numbers instead of letter must be
1=a, 2=b etc.
It looks like Jeffrey Fries did the same as Jan Sloot in his patent,
describe a working variant without (or almost without) compression but
"forget" to describe the magic algorithm of the synchronization and
internal clock what can for example be used in combination with a sort
of build up dictionary to send repeat signals to decrease the amount of
frequencies send over the serial line or stored at a medium.
Sportman
> Its a red flag common to many hoax claims and fraud inventors. They
> refuse to use the established terminology of domain and try to use
> their own confusing terms to jazz up their descriptions, and sound more
> unique.
http://www.fly-pegasus.com/resume.html I see an experienced man who
probably build his inventions first as modem (electronics + digital
techniques) and then did a discovery that streaming data with (big)
numbers has an advantage when using a sync/clock signal for
compression/repeating code purposes.
It's no coincident that Jan Sloot, Adam Clark and Jeffrey Fries build
all three first a streaming data device before they came to an other
application of their inventing that it can also be used to store
incredible amounts (1.000-1.000.000 times) more data on a little media
as a floppy or smart card then was possible before.
> 4 years does seems a bit too long, I googled a bit on this.
Thanks for the information I didn't know it could take so long.
Matt Mahoney
> Sachin Garg wrote:
> > Its a red flag common to many hoax claims and fraud inventors. They
> > refuse to use the established terminology of domain and try to use
> > their own confusing terms to jazz up their descriptions, and sound more
> > unique.
> When I check Jeffrey Fries resume
> http://www.fly-pegasus.com/resume.html I see an experienced man who
> probably build his inventions first as modem (electronics + digital
> techniques) and then did a discovery that streaming data with (big)
> numbers has an advantage when using a sync/clock signal for
> compression/repeating code purposes.
This is just another example of foolishness from the USPTO for
patenting what modem makers have already been doing for 20 years.
Almost as foolish as anyone who believes that you can extend this
method from a few bits to 1000 bits using 2^1000 frequency levels and
not run into noise problems. Of course his resume shows no analog
design exerience.
-- Matt Mahoney