This issue's highlights are from risks, dcsb, cypherpunks, risks, tbtf, and crypto-gram.
It's been stated that people would never tolerate their sneakers blowing up as regularly as their computer crashes. Well maybe not. From an NPR piece on high-tech toilets in Japan (which may have caused several fires): Reporter to woman shopping for toilet: "Are you concerned about the possibility of fires?" Woman (as translated): "No toilet is 100% safe. I am willing to accept some risk."
An Asian Technology Information Program (ATIP) report states: "A German court recently decided to hold a bank liable for losses in connection with a stolen Eurocheque card in part because the 56-bit encryption protecting the card was considered "out-of-date and not safe enough."
Two Canadian ISPs claimed they were subjected to denial of service attacks from the Chinese government because they host a web site for a religious group banned in China. Story
Services that provide anonymity unless faced with a subpoena may be relatively easy to crack open. An email article claims that to file a subpoena in a civil suit all you need is to get a lawyer (OK, that takes some money), file a lawsuit, and you have the authority to issue a subpoena (is it really that easy?). These subpoenas typically do not require a judge's OK. AOL and MSN has let users know about pending subpoena's, but it's not part of the terms of service. Yahoo has not. TrustE has been asked why it does not require this as part of its seal (which Yahoo has). TrustE made some standard mouth noises about looking into it.
Erik Parker put out a call for reviews of security products for a "security product review section". It could be he's looking for some free labor, or it could be a chance for someone to increase their visibility.
The Clinton administration plan for Federal Intrusion Detection Network (Fidnet) has caused some reactions. It calls for one system to monitor activities on nonmilitary Government networks (just which ones is currently unspecified) and a separate system to track networks used in crucial industries like banking, telecommunications and transportation. The plan is to be fully in place by 2003. Critics are concerned with misuse and the opening of new security holes. It's unclear what Fidnet does in the face of encrypted traffic and how information is protected (in fact much of the plan is still vague). The House Appropriations Committee approved a budget for the Justice Department that specifically forbids the FBI to spend any money on Fidnet. Story. In a wonderful attempt to coopt the privacy issue, White House national security adviser, Sandy Berger stated "Obviously we are very concerned about protecting privacy rights. But there is also a privacy right in not having hostile entities attack systems."
E*Trade states that recent CA legislation (SB 1124) allows them to open accounts with electronic signatures (instead of requiring paper forms), so it plans to do so.
Janet Reno has urged the German Justice Minister to ban crypto products on the Internet, including public domain items. Story.
A Risks contributor outlined in detail how he can use ActiveX controls shipped on his HP Pavilion computer that is running Windows 98 to break into it Story. They are marked safe for scripting even though they can launch programs and read and write the registry. ActiveX signature checking happens only on download of the control; there is no related checking on execution. The solution seems to be to use a new feature called HTML applications (or .HTA files). They are locally trusted so do not have to be marked safe for scripting.
The House armed services and intelligence committees are warning that widespread use of encryption technology would be "devastating" for US law enforcement and the national security establishment.
A year-long study by a senior panel of Defense Department officials recommended an expansion in the role the reserves play in national defense, including the formation of a virtual cyberdefense unit to protect the nation's critical infrastructure. The new reserve cyberdefense unit "would consist of individuals with information technology skills who could perform their duties from dispersed locations rather than working as a single consolidated unit at a specific training center". The suggestions on how to staff this unit seemed less innovative, suggesting exchanging high-tech DoD training for years of service. I'd bet you'd need a pretty good clearance for any really interesting training.
Paul L. Hutton ((425) 825-3450 - Voice, (425) 814-4254 - Fax) is looking for a "competent and trustworthy consultant" to reverse engineer and break a PBX encryption protocol as part of his product development efforts.
shows you how much information your web browser gives out about
you. It's a pretty thorough job.
[Biased editorial comment: So does Mike Reed's Snoop Server. You can find it and pointers to other privacy test sites, including at Onion Routing's List of Good Privacy Test Sites. --P.S.]
Here's a full excerpt from Keith Dawson's TBTF on the SAFE bill, as
Keith can always say it better than I can:
..Bill to relax crypto exports is gutted
Two years ago the most privacy-friendly attempt to relax US crypto export controls, Bob Goodlatte's SAFE bill, was savaged  in two House committees and was eventually withdrawn. I wrote at the time:
On Wednesday 21 July the House Armed Services Committee voted 47 to 6 to replace the text of the SAFE bill with one drafted by the law- enforcement community , and manifesting exactly the opposite ef- fect. Here is the text of the new "unSAFE" bill  (PDF format, 47K). Eventually the House Rules Committee will need to decide which version of SAFE, if any, reaches the floor for a vote.
Let's do an exercise in political accountability. 17 SAFE co-spon- sors sit on the Armed Services Committee, and 13 of them voted to gut the bill. (Note that even if all SAFE co-sponsors had voted to preserve the original bill, the outcome would have been the same, by a vote of 34 to 19.)
I have updated the Congressional Hypocrites page , which first spotlighted the 285 politicians who voted cockeyed on matters of Internet pornography, to cast a cold light on the 13 members who both sponsored the SAFE bill and voted last Wednesday to ream it out. They are:
It's a pity that we need to single out for special accolades that minority of Congress members who do what they say they are going to do; but such is our system. The four SAFE co-sponsors who voted against gutting the bill are:
Confinity's PayPal is an instant payment service that allows people to exchange money through their Palm organizers using infrared. The software is supposed to be ready for widespread use this fall. An analyst said: "If two people go out for dinner and decide to the split the tab, one person points their palm device at the other persons palm device and it's done. The money is passed. You're basically enabling the person to transmit money, and I think there's a niche for that." There was much speculation on the security challenges here, and Hellman is an investor.
Lucky Green's announcement of S/MIME experiments associated with cypherpunks meetings inspired Bob Hettinga to accuse all X.5nn variants of being spawn of the devil and doomed to failure for their assumption of a centralized authority. Surprisingly, the discussion actually went uphill from there, including the actual and current deployment differences between X.5nn (or X.BlahBlah as it was less affectionately called) and PGP. The most striking technical difference is X.509's emphasis on signed certificates vs. PGP's use of signed keys. Having recent experience with implementing PKIX freeware (see our paper on Jonah at Usenix), I agree that the lack of information about keys such as key lifetime can sometimes be problematic. There was also discussion of the difficulty of managing revocation information in the X.5nn framework, and the popularity of S/MIME and X.5nn implementations ("Every man and his hamster are working on X.509 cert handling"). Several people claimed they had found it easy to translate keys from one format to the other in response to the accusation that DNs cause problems when working with X.5nn.
The draft the Electronic Communications Bill in the UK would give electronic signatures legal status. Story. It would also force companies to register their encryption keys with a third party. The bill also makes it a crime for an employee who is aware of the "decryption notice" submitted by police to leak information about its existence. A Tory spokesman described the legislation as a "dog's breakfast" (because of its length).
Rumor has it that travelocity.com's "secure" web server only supports 40 bit encryption, and that they have not responded to the customers who have complained about this weak security.
A Federal Trade Commission workshop on the Federal Children's Online Privacy Protection Act caused a spate of privacy-related items. The Washington-based Center for Media Education did a study that involving a random sample of children's sites and the 80 most popular children's commercial Web sites. The random sample showed that while 95% of children's sites collect personally identifiable information, 73% of those post no privacy policies. Less than 6% attempt to get any permission from parents at all, and less than 3% use methods the center said it considers to be acceptable for obtaining verifiable and prior parental consent. 88% of the popular sites collect personal information from children, more than a quarter of those post no privacy policies. Less than 26% attempt to get any kind of parental permission, and not quite 13% use methods for obtaining verifiable, prior parental consent. Before the workshop, the FTC reported that privacy legislation was still unnecessary. A risks reader shared this excerpt from a privacy statement: Children 12 years of age and younger are not permitted to opt in for these future e-mailings because the opt-in software requires users to fill in their age and only users above 12 years of age are able to submit opt-in authorizations.
Schneier's Crypto-Gram discussed three sites whose use of SSL is not airtight (onsale, Verisign and Spree), and another which claims to encrypt passwords but doesn't seem to use SSL or any other security protocol at all (BizTravel).
The Secure Digital Music Initiative has released version 1.0 of their standard. This technology is supposed to end (or at least severely limit) music piracy. It lacks details but discusses the framework and goals. Systems that manipulate encrypted SDMI music are supposed to obey usage rules. Cryptography is also used for device authentication, which is supposed to stop non-SDMI devices from even downloading SDMI content. Software techniques are familiar from copy protection techniques. Some of the interesting hardware restrictions include degrading music that is played at high speed and limiting the bandwidth of recording music with a microphone and making it monaural. You can copy your music around to (ostensibly your) devices, but you can only have 3 copies "out" at any time. This seems to be an attempt to limit propagation to your two closest friends. Later they hope to implement a watermark that keeps music from being loaded if it has been compressed, as an attack on the MP3 format. This allows legacy CDs to be passed around at MP3 format and played on these devices forever.
Saflink has opened what it says is the information technology (IT) industry's first biometrically enabled Web site - http://www.safbank.com. It simulates a typical Internet banking/brokerage site and requires users to present a biometric credential. At a glance, I didn't see any information on how to revoke your credential.
There was much discussion about what IETF should do about DES. While it's being deprecated because of its weakness, TLS (the IETF version of SSL) is increasing security by upgrading from 40 bit crypto to DES. Many cypherpunks consider that to be an illusion of security, and worse than no security at all. Off lists, discussion continued at the July IETF meeting. 3DES is perceived as too slow for applications like IPSEC. Attendees were split between choosing an AES finalist, waiting for the AES winner[s], or moving to something else asap.
An inordinate amount of cypherpunks bandwidth has been taken up by a gent who has published a book claiming Emily Dickinson encyphered the letters SamB (for Samuel Bowles) into many of her poems. He basically strips the poems of all other letters so that you can see how often those letters are sprinkled across the page. Visual patterns are said to emerge.
Jim Gillogly, cypherpunk, seems to have been the first person outside the CIA to crack much of the cypher on a sculpture called Kryptos dedicated there in 1990. Story.