Post by nemo_outisPost by MyNymBy all means, people should analyse and compare products. But in my
view, the comparisons should be based on fact and not speculation.
Nevertheless, if negative speculation is to be the order of the day,
then in all fairness, it should be applied to all.
Regards.
No, NOT only "facts" should be taken into comparison. When evaluating
security software the criterion need NOT be "beyond reasonable doubt" or
even "balance of probablilities." No, the test for most is the Caesar's
wife one that goes beyond even "a shadow of a doubt" to "above
suspicion."
I agree, on the basis that security software is directed to serious
usage where there can be no chance of antagonistic involuntary
decryption. From production secrets to national security, there must
be no chance whatsoever of compromise.
Post by nemo_outisNor is any consideration of "fairness" to all comers applicable. A
security program - any particular program - may be rejected for any
reason or no reason.
The objective is to determine a subset of security programs that meet the
user's needs, not some theoretical "balanced" approach that would please
others or satisfy some abstract ideal of "commercial justice" (if there
is such a concept!).
Now, of course, it is entirely possible that application of these
criteria may result in the set of acceptable security programs being very
limited or even being the "null set." In such conditions the user may
decide to relax some or all of his selection crtiteria or, perhaps, to
take an entirely different approach (e.g., eschewing all current security
software in favour of other methods - sometimes called the Osama approach
:-)
Moreover, in the case of Truecrypt, while there may not be fire there is
considerable smoke. First, Truecrypt is NOT open-source - while the code
is available for inspection the licence hedging it is far from being real
open-source.
Look, for such advanced code, I seriously doubt anyone at all has sat
down and understood all of the code. A few lines hidden amongst
hundreds of thousands, perhaps millions, of lines of code would be to
most experts totally concealed. Or, some calls to routines which call
to other routines perhaps many many times over may be just as good at
concealment. Until the encrypted OS falls into the hands of the NSA
after a raid on terrorists who plan destruction somewhere in some
city, *then* NSA pulls the plug and opens the backdoor by entering
some line of instruction somewhere in the myriad lines of code that
has never been found.
I mean, who has sat down and followed each line of code comprising the
TC executables? Each line, and each call, and so on? Anyone?
Schneier hasn't that's for sure. I am sure Schneier reads this, how
about it Bruce, who has examined the TC executables?
Post by nemo_outisFurther, the (two?) developers have gone to extraordinary lengths to hide
their identity (although I have ferreted out some significant leads in
this respect which I believe identify them). Fair enough, I suppose, but
it means that one avenue for possibly instilling trust - the reputation
of the authors - is not available. Moreover, the authors have taken
additional measures to "close" Truecrypt, one of which is the application
of copyright.
Copyright won't prevents reverse engineering for examination of
backdoors. I don't believe there are only two devs. I think the
software was released as open source and free of charge to effect the
biggest uptake throughout the world, so it would be the OTFE package
of choice. Hopefully, of choice by significant criminals such as
terrorists and the like. Makes sense doesn't it?
Post by nemo_outisMore sinisterly, the authors have gone to extraordinary lengths to
"purge" all traces of earlier versions of the program's source and binary
code from the internet (not just Wayback but many others). However, after
considerable effort I have assembled a complete library of source and
binaries for all previous versions which I will make available to those
who request them. This attemopt to "bury" previous versions makes
comparison of change history difficult (and could be viewed by the
paranoid as a deliberate attempt to impede discovery of later compromised
code).
Fair enough, but if this project has been underwritten by the NSA or
someother dodgy agency, then the backdoors have been around since
V1.0. So, I don't think the code became compromised later, if it is
compromised, it has probably always been compromised.
I don't know why they have purged the old versions though, in that
case.
Post by nemo_outisYet further, the principal avenue for interaction regarding Truecrypt,
the Truecrypt forums, are rigidly controlled by the developers and/or
their henchmen - to the point where bug lists are not public and where
adverse comments frequently "magically disappear" and are purged from the
forums.
This is well known, see my earlier comments in earlier posts. IPs
banned, nicks banned etc if too many questions are asked in the
forums.
Post by nemo_outisThe net effect of this secrecy, rigid control, and chilling atmosphere is
to discourage many academics and other qualified crypto people from
examining the code and promoting Truecrypt. Remember: the principal
alleged benefit of open-source software (although Truecrypt is only
quasi-open-source) is the "many eyes" principal. In the case of
sophisticated security software, however, this principal is greatly
diminished since there are so few "eyes" which are qualified to examine
the subtleties of security software. All the more true when one may be
looking for, not mere bugs, but carefully camouflaged backdoors. (For
those who doubt that compromised code can exist despite being "in plain
view" and open-source, I invite them to review the Obfuscated C contests
of past years).
Agreed. It would not be difficult for the magicians at NSA to hide a
few lines in all that code.
Post by nemo_outisIn short, I'm suspicious of Truecrypt. Suspicious enough that I don't
use it. Do I have hard proof? No! But I don't require hard proof to
disqualify it. Suspicion is sufficient.
BestCrypt Wholedisk and FreeOTFE. My way may not be your way, but this
is what I have chosen.
As far as I know, only TC and DCPP have the capacity for a hidden OS
which I use because it is the best sandbox possible for sensitive
information (along with Tor, Privoxy and SMAC 2.0). I used to use
DCPP until it became unusable because of its reliance on FAT32 for the
initial setup (which fails Win7). I now use TC because while I am
deeply suspicious of TC and its origins, I know that only in utter
emergencies would the NSA (or equivalent) pull the plug on it and use
the backdoor, and my use does not qualify as an utter emergency. My
use has no illegal or national security taint at all, so for my
purposes, TC is ok (I need the hidden OS sandbox). However, my
suspicions remain.
PS How do you know Sarah Dean is not on the payroll of the NSA? Just
kidding, I have followed her career over the past decade from a
student dabbler to wherever she is now. She is very very good, but no
hidden OS.
regards
thang