Time for another email rant. This morning while playing #1 videogame in my life -- that of zapping emails -- a quote from Bart Simpson in Screaming Yellow Honkers came to mind,
I didn't think it was physically possible, but this both sucks and blows.
I've said before that
email sucks. It's not a huge leap to declare that email blows as well.
Despite decentralized filesharing leading to global Internet traffic analysis in June 2004 revealing that in the United States peer-to-peer represents roughly two-thirds of traffic volumes, and in Asia peer-to-peer represents more than four-fifths of traffic volumes, I believe that the wasted time and resources due to the flaws in email's basic protocol (SMTP) in 2004 cannot be denied.
I've ranted before that email must evolve. For this post, let's ignore that even if email were spam-free and virus-free, that I cannot keep up with email -- my coworkers have noted that a good portion of my day is spent replying to, forwarding, archiving, and deleting email. (This is what it means to be a knowledge worker in 2004??)
Instead, let's do a reality check by reminding ourselves how bad spam and viruses really are at this moment in time. Postini's stats page, which I believe represents a good population sample, has some gems:
- 72.5% of email is spam
- 1 in 30 emails is virus-infected
- 1 in 2 SMTP connections is wasted
- An average of 30% of an email server's capacity is hijacked by spammers trying to steal proprietary email addresses and other information stored in the corporate directory
MXLogic's Industry Stats (
pdf) also has several gems:
- Employee activities surrounding spam management (such as deleting spam, reviewing quarantine areas and asking administrators for filter preferences) takes 80 minutes per 1,000 emails with spam blocking technology; without spam blocking technology, these activities take 200 minutes per 1,000 emails (Osterman Research, 2003)
- Organizations with 5,000 e-mail users waste 10 minutes per end user per day on spam, and 43 minutes per day per IT staff member; the ensuing drain claims nearly $4.2 million in lost productivity per year (IDC Research, 4/04)
- Out of the 31 billion emails sent per day, 2.5 billion are pornographic (totaling four pornographic emails per day, per user) (Spam Filter Review, Q1, 2003)
- As of July 2004, 3% of Internet email complied with the federal CAN-SPAM Law (MX Logic
Research, 07/04)
- More than 30 percent of spam is sent from computers that have been infected with a worm or Trojan horse and so are used as proxies for sending spam, completely unbeknownst to their owners
(Osterman Research, 2004)
- Over 95 percent of viruses are spread via "junk" email. (The Gartner Group, 6/04)
Viruses and email feed off each other; email provides a conduit for viruses to spread, and many viruses as part of their actions do something to their victims' email (usually spamming and/or using it to spread the virus). Fred Vogelstein's article
"Why Hackers Are A Giant Threat To Microsoft's Future" (
Fortune, Oct 18 2004, pp 263-272) cites reasons why viruses like Microsoft (sources: IDC, Microsoft, Symantec):
- There are an estimated 676 million PCs in the world
- 84% of those PCs are online (over half have a broadband connection)
- 90% of those PCs run Microsoft Windows (Mac and Linux each have 3% market share)
- Only 120 million PCs (less than 20%) have Symantec antivirus software, which accounts for 60% of the antivirus market
I believe that part of the challenge is the
social problem of teaching users to take responsibility for their machines -- that you cannot blindly trust things that come from elsewhere, and that if even one person in 6 billion responds to just one piece of spam, there will be incentives for spammers to up the ante. But there is also a
technical problem: part of the challenge is the protocols need to support users by not blindly trusting things that come from elsewhere. This is a key challenge in decentralization software research today --
How can decentralized software systems be built in which the interacting parts owned by different people do not blindly trust each other?
-- and breakthroughs in this area will inform better protocols in the future. Only then will we come closer to the dream vision that
Software should work on behalf of its owner
rather than on behalf of the software writer, the software seller, or other people who can hijack the software for their own purposes.
Recent Comments