Sigh. I tried to make two basic points, both of which were apparently missed
by most of the people who replied.
"Stanislav Malyshev a.k.a Frodo" <[EMAIL PROTECTED]>:
> AM>> What most of you are missing in your zealotness is that the discussed
> AM>> service is based on reparse points, a new technology in Windows 2000.
>
> What you missing in your zealotness is that I'm not zealot and never was.
> So please hold you tongue and avoid sticking labels.
>
> AM>> This is a MUCH cleaner solution than using kernel modules that trap
> AM>> file-related system calls. It also, by design, allows several such
> AM>> modules to be loaded concurrently.
>
> In fact, it *is* using NT kernel modules to trap file-related calls,
> AFAIK. Only those modules are integrated.
I am not interested in defending a technology I know very little about; that
would be silly.
In fact, it would be as silly as ATTACKING a technology I know nothing
about. Yet, when the Microsoft press release was forwarded here, most of
the posters here did just that.
I was trying to show that any groundless attack can be countered by a
different speculative argument. Only I took the time to read up a little
about SIS, so what I said had something to do with reality.
So these discussions, technically, are worthless; the act of starting such a
useless discussion can be attributed either to cluelessness or zealousness.
The following is a good example for this.
> AM>> That's a stupid assertion to make if you don't know what sort of
> AM>> hash they're using. Do you?
>
> Didn't they learn you in your math class that *any* hash that isn't equal
> by length with hashed objects will collide objects? It's called Dirichlet
> principle, IIRC - if you have 3 objects and 2 signatures, 2 objects must
> have identical signature. Think about it.
Irrelevant.
Can you even precisely DESCRIBE the following
1) the SIS file comparison algorithm
2) any other SIS algorithm
3) the exact usage they make of the ``fingerprint''
4) the size of the ``fingerprint''
Or are you using a press release as the basis for your speculations?
As it happens, the answer to 4) seems to be 128 bit. Without knowing
the answers to 1), 2) and 3), you still can't assert that SIS is ``unusable
for any serious data storage''.
Moshe Zadka <[EMAIL PROTECTED]>:
> Please learn basic combinatorics. It doesn't matter what kind of hash they
> use.
Please learn to read. Of course collisions are possible; that's not a matter
of opinion. To assert based on this that SIS is ``unusable for any serious
data storage unless it compares files bit-bit every time'' without knowing
the kind of hash, and how it's used, is stupid.
=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
