
[Archivos] .PAR y .P*
#1
Escrito 09 June 2002 - 11:45 AM
#2
Escrito 09 June 2002 - 11:54 AM
El poder de las Matematicas unido a la informatica!!!
#3
Escrito 09 June 2002 - 11:57 AM


Alesgar: Se me saltan las gráminas...
Hitomi ho tojireba hohoemu kao ga mieru
#4
Escrito 09 June 2002 - 12:11 PM
#5
Escrito 09 June 2002 - 12:20 PM


Alesgar: Se me saltan las gráminas...
Hitomi ho tojireba hohoemu kao ga mieru
#6
Escrito 09 June 2002 - 12:49 PM

Los ficheros PAR
Son ficheros de paridad q se usan como comodines (o sea, q si una película tiene 20 partes y no nos hemos podido descargar las partes .3, .8 y .19, con tres ficheros PAR y el correspondiente programa, estas tres partes se recuperan de forma automática).
El programa para usar/generar los PAR es el FSRaid (se puede descargar de http://www.fluidstud...com/fsraid.html ), hay otros programas aparte de este, pero bueno, la cosa es q se bajan tantos ficheros PAR (.P01, .P02,.....etc) como partes nos falten, se guardan en el mismo directorio en el q están el resto de los ficheros q se descargaron, se da al fichero .PAR y el programa comprueba q los q se bajaron están bien y dando a "Repair" regenera los q faltan sin importar el orden de los mismos (si en vez del .3, .8 y .19 fuesen otros tres distintos, con los tres primeros ficheros PAR se arreglaría el .avi igualmente).
Para generar los PAR, se le da a la opción "Create", se eligen los ficheros de los q se van a hacer los PAR, se desliza la barra más o menos a la izda, según el nº de ficheros PAR q se quieren generar, se da a aceptar y ya. El primero q se crea (el .PAR) sirve de fichero de verificación (como los .SFV) y se sube para comprobar que lo q se bajó se bajó bien, y cuando empiezan a desaparecer partes, según se van pidiendo, se sube el .P01, el .P02....etc.
De cualquier forma, el FSRaid tiene un buen manual .html q cuenta todo esto, pero bien.
Ale, espero que os haya sido de ayuda. Saludos y si hay más dudas, avisen.
Frases célebres:
by Sunburn
No tenia saldo. Estoy estudiando. Adios.
Post-it humedos con es.// escrito con tinta resistente al agua por el lado de la pega.
by Deadsunrise
En casa sólo, se folla [...] de puta madre.
#7
Escrito 09 June 2002 - 07:40 PM
#8
Escrito 09 June 2002 - 08:11 PM

Quick run through, remember QN I think said it was the summed parity
of all the bytes of a set of files. Well it was like that, now it is
different.
The old way (hex numbers (base 16 or 4bits), 2 hex numbers is a byte
(8 bits) :
blah.r00 = 12 34 56 78 9A BC [soy subnormal].
blah.r01 = 01 02 03 04 05 06 [soy subnormal].
blah.r02 = 02 04 06 08 0A 0C [soy subnormal].
the bytewise sum in MOD256 (or bytewise XOR) is the p01 file =
blah.p01= 15 3A 5F 84 A9 CE .......
So p01=r00+r01+r02 and conversely r00=p01-r01-r02 or for any one of
the 3 rxx files, you subtract the other 2 from the p01 file to get it
back. It only works for one missing file. Basically 15-02=13 but
13=12+01 or 11+02 or etc....
You can prove this to yourself with the wincows calculator, it does
hex in scientific mode. It works. MOD 256 means only take the 2 least
significant numbers from the answer. Like FF+1=100 but in MOD256, it
equals 00.
So this was the old way, which is where the "you can only be missing
one file of the set" was from. The new way, as was said by John and
Don, is by matrix method called Reed-Soloman coding and I have no good
way to explain it, nor did I look into the math too hard (just haven't
had the time). I think the basis is that each pxx file is a function
(or equation) of the bytes of all the files of the post, not just a
subset. So each pxx is an equation of all the files, and your missing
2 files, so you get 2 pxx files. What you now have is 2 equations 2
unknowns. Now it is simple algebra.
All right, for my next trick, I will attempt a simple mathmatic proof
of the Reed-Solomon coding method. Mrs. P blindfold please......(make
drum roll sounds while you read it)
OK class, for this example we will use the 3 data files as our post
and make 3 pxx files from the following matrix (why? because I have an
example of a 3 by 3 Reed Solomon matrix and don't want to calculate
the Vandermonde matrix for another example, but don't post 3 pxx's for
3 data files, it's just wrong)
The matrix
1 1 1
1 2 3
1 4 9
So if p1, p2, p3 are parity files and d1, d2, d3 are data files then:
p1=1(d1)+1(d2)+1(d3)
p2=1(d1)+2(d2)+3(d3)
p3=1(d1)+4(d2)+9(d3)
So say we are missing d1 and d2 so we get p1 and p2, we get 2
equations and 2 unknowns:
d1+d2=p1-d3
d1+2(d2)=p2-3(d3)
so subtracting the first equation from the second removes the d1 and
yields:
d2=p2-p1-2(d3)
and subtracting the second from 2 times the first cancels d2 and
yields:
d1=2(p1)-p2+d3
So d1 and d2 are now known. As can be seen, this would work with any
combination of 2 of the px files and will work for loss of any 2 data
files.
Lets do the earlier example rxx files:
blah.r00 = 12 34 56 78 9A BC [soy subnormal].
blah.r01 = 01 02 03 04 05 06 [soy subnormal].
blah.r02 = 02 04 06 08 0A 0C [soy subnormal].
then p01 and p02 are
blah.p01= 15 3A 5F 84 A9 CE .......
blah.p02= 1A 44 6E 98 C2 EC .......
so using the solutions for d1 and d2:
byte1 of blah.r00 = 2(15)-1A+02 = 12
byte2 of blah.r01 = 1A-15-2(02) = 01
The rest of the bytes will work out the same. As for the way the
technique does it, the solution comes from a matrix equation. The
matrix equation for missing d1, d2, and p3 is:
0 0 1 ^-1 d1 d3
1 1 1 * d2 = c1
1 2 3 d3 c2
which should give you (though I can't remember how to do an inverse
matrix to prove it)
1 2 -1 d3 d1
-2 -1 1 * c1 = d2
1 0 0 c2 d3
This matrix gives us the 2 equations we had and another d3=d3 (no way,
duh). This is how it would solve for more than 2 unknowns. I would
have to do a couple of iterations.
Note the matrix for the par files is the number of data files wide
(call this n) by the number of par files tall (call this m). The total
files n+m must be less than 2 to the number of bits. For mirror the
number of bits is 8 (a byte) so the max number of files plus pxx must
be less than 256 (due to the MOD256 stuff). And note the matrix for
the par files is 1,2,3,4,5,.... carried to the power of the row-1 for
each row. This, I believe, makes the equations reduce easier and
probably makes the matrix stuff easier.
Well now you know more than you wanted to. Automan, if you want to,
you can put this in your explanation of the par files. I might throw
this to the mirror creators for their use too. It is a lot less math
than explaining the matrix calcs for the layman. I don't even
understand that fully.
Well, perhaps math-magical matrix was a better explanation after all.
Dudeman
"Techi-puke at your service"
¿Cómo se puede creer en semejante hijo de puta?
#9
Escrito 09 June 2002 - 08:38 PM

Frases célebres:
by Sunburn
No tenia saldo. Estoy estudiando. Adios.
Post-it humedos con es.// escrito con tinta resistente al agua por el lado de la pega.
by Deadsunrise
En casa sólo, se folla [...] de puta madre.
#10
Escrito 09 June 2002 - 08:38 PM




Alesgar: Se me saltan las gráminas...
Hitomi ho tojireba hohoemu kao ga mieru
#11
Escrito 09 June 2002 - 08:50 PM
buenisima la explicacion tio.. me la he leido y +- entendido y todo (jejej, como se nota que me interesa y no lo que doy en clase...)
#12
Escrito 09 June 2002 - 11:03 PM
#13
Escrito 09 June 2002 - 11:05 PM
#14
Escrito 10 June 2002 - 12:00 AM
Matematicas+binarios+ingles=


#15
Escrito 10 June 2002 - 01:00 AM
#16
Escrito 10 June 2002 - 12:14 PM



twitter]scarletlayer[/twitter]
Vaya mierda con las firmas....
#17
Escrito 10 June 2002 - 05:46 PM
#18
Escrito 10 June 2002 - 08:07 PM
#19
Escrito 10 July 2002 - 10:52 PM
Un programa para usar los par
SmartPAR
Tuto sobre ese programa
Komo usar el SMARTPAR?
Otro programa para usar los par
FSRaid
Y su korrespondiente tuto
Komo usar el FSRaid?
Kon estos tutos no tendreis ke tener ningun problema son mu faciles de entender
#20
Escrito 12 July 2002 - 01:57 PM
yo m he descargao los episodios d crest of the stars y m han fallao X en cada episodio como narices los consigo si es q solo sale el .par y no hay ningun .p01 .p02 ...
#21
Escrito 12 July 2002 - 02:17 PM
1 usuarios están leyendo este tema
0 miembros, 1 invitados, 0 usuarios anónimos