Chris also recently commented that it is surprising that various operating systems do not automatically incorporate encryption into your basic profile. The following was my reply.

As to why windows *did* not, remember that of the various versions of windows that came before win 2k and XP, only NT and “windows for workgroups” (not well known, it came between windows 3.1 and windows 95, never sold much) had a true concept of multiple users. Notably, windows 98’s multiple users functionality caused all sorts of settings and information to “leak” across accounts. Hardly the ideal environment for any type of automatic encryption use.

In today’s world, where even substandard systems like windows have a decent concept of multiple users, incorporating encryption into the system’s storage is far more possible. I am not sure how *effective* it would be though. Remember, in order to read and write to your encrypted files, the computer would have to “know” the key, that is store it in memory. (alternately, it could store the *files* in a ram disk, and then only need the key to sync the ram disk to the encrypted versions of the same files, but this is *far* more fragile as a computer crash would cause incalculable data loss, as the old pattern of “save often” would not *necessarily* save to non-violate memory (aka the hard drive).) The key, or the pass phrase (depending on the encryption scheme), is then vulnerable to attack by anyone else on the system with access to your memory (IE the person who hacked your computer last week because you run windows, or the system admin, or a virus). And losing a key causes far-reaching headaches, as I am sure you are aware.

Better than any pervasive scheme of encryption then, I think, is simply better (I.E. more widespread) understanding of it. Companies such as Tiger Privacy[2] provide easy to use, and sufficiently secure, encryption for email that *could* be easily and widely deployed without too much trouble for users. While this has other problems (For example, it requires POP3, and thus emails are stored on the local drive, vs the (presumably) more trust worthy (I.E. more likely to be a RAID) server), it does demonstrate that encryption is not necessarily orthogonal to ease of use.

[1] http://www.schierer.org/~luke/log/view.php?date=20050525-1009
[2] http://tigerprivacy.com/

Last week Chris sent me an article talking about the use of encryption in the Revolutionary War and colonial periods.[1] Turns out that various encryption methods were very well known and quite widespread for government, business and interpersonal correspondence. So wide spread was the use of encryption, that the courts “have not treated those persons who have used encryption, ciphers, and codes with any presumption of illegality.”[2] In criminal cases, “evidence that a defendant used a cipher or encryption has been allowed (as in United States v. Burr) as proof of the means used to commit an illegal act, but as proof of an illegal act by itself.”[3] This is significant to remember in light of a Minnesota case where the existence of a PGP encryption program on the defendant’s computer was deemed evidence, by both the trial court and the appellate court, of criminal intent.[4]

A caveat here: The crime the defendant is accused of is horrific, and if guilty, he should be locked up for life with no chance of parole.

Still, it is disturbing to see that an encryption program, with no evidence of encrypted files, or its use in any way related to the crime, is evidence that you intend to break the law. There are any number of good reasons to use PGP, or its free clone, GPG, beyond illegal activities. For instance, many open source packages come with GPG signatures to help prevent a man in the middle attack from substituting a modified version of the package either in the mirror you are downloading from, or via a transparent proxy. Similarly, PGP/GPG signatures are used to verify identity in email correspondence, and, just as it was in the colonial period and the early days of the country, used to encipher perfectly legal, but personally sensitive information to protect it from friends, family, neighbors, and random people out on the Internet who should not be reading it. It is disturbing to see how quickly after finding such research, I see confirmation that it is indeed necessary to have it.

[1] http://vjolt.student.virginia.edu/graphics/vol2/vol2_art2.html
[2] http://vjolt.student.virginia.edu/graphics/vol2/vol2_art2.html#IVD
[3] http://vjolt.student.virginia.edu/graphics/vol2/vol2_art2.html#IVD
[4] http://news.com.com/Minnesota+court+takes+dim+view+of+encryption/2100-1030_3-5718978.html

Okay, yes, I am vain, but this week’s Debian Weekly News email just made my day.

Debian-Volatile Strategy. The volatile team [29]pondered to create a second [30]volatile archive with less strict criteria. This would help packages such as [31]Gaim that need to be updated during the lifetime of sarge in order to support modified protocols. They are also looking for ways to announce updated packages.
29. http://lists.debian.org/debian-devel-announce/2005/05/msg00016.html
30. http://volatile.debian.net/
31. http://packages.debian.org/gaim

I talked to some of the #Debian-release guys about this late last week, when I learned that we would not be able to back port some rather trivial bug fixes to Debian’s 1.2.1 Gaim package with any significant chance of getting them accepted. If we cannot get trivial bug fixes in, how in the world would we ever get the kind of significant changes that a new authentication for a protocol would cause? I would rather avoid a situation like the current Debian stable, which has a 0.58 that will not work with many of the protocols it claims to support.

It is nice to be listened to.