Thursday, February 06, 2003

This essay focuses on the use of email. You also want to read my recent comments on spam.

I was reminded of how far we are from being able to use email after meeting with my son's teacher. After the meeting I asked for a copy of the report via email but I was told that was not possible since the information was confidential and that with their email system everyone at the school can read it.

In fact, all of our email is sent in the clear through intermediate systems with no formal rules about who can look at the messages and files. There is also no good way to determine if the mail was actually delivered and it is very easy to send the message to the wrong address, especially due to the ease of false matches in address books.

There seems to be a disconnect between the importance of email and the reality. We fully trust postal mail even though the only protection we have is laws against opening first class envelopes and the social conventions against doing so. The difficulty of accessing paper mail provides some degree of protection. What makes this even stranger is that I'm beginning to see the retronym, postal mail as people are beginning to assume email is the norm while our institutions seem to view it as a novelty.

We tend to treat email as if it had the same protection as postal mail except when we have a legal or explicit obligation to be careful. Then we have a problem. Actually we have a number of problems. One problem is that few organizations seem capable of effectively managing access to shared information but that's a separate topic.

The question is why can't a student evaluation be sent via current email systems?

Some of the reasons are technical such as the lack of a level of security model corresponding to first class mail. Simple, and anonymous over-the-wire encryption of email would allow us to assume the kind of confidentiality we associate with first class mail (in the US). This is a standard far below the high security systems which try to keep mail encrypted even when stored locally and which also attempt to provide authentication. We simply want to keep the messages from prying eyes. We also need a delivery confirmation mechanism which doesn't make the recipient feel spied-upon. While the post office does offer delivery confirmation it is not used for most mail since delivery is assumed. Email has been sufficiently reliable, at least relative to its normal use, so we have tended to presume that the mail got through. But with the increasing focus on spam and the likelihood that messages will get lost amid the junk and with the tendency of people to abandon old mail accounts we do need to be sure the messages got through. Yet I, among others, turn off receipts since they are too intrusive. A compromise would be to better integrate the receipt mechanism into the delivery and allow one to associate it with relationships. Knowing the mail got past your gatekeepers may be sufficient even if it doesn't prove that you actually read the message.

We are still in the teething stage of email and view it as just a faster version of postal mail. One example I keep pointing out is that email addresses are treated like street addresses or phone numbers rather than what they really are--tokens that can be used to get our attention and manage conversations.

The reality of email falls short of what we should expect and that is just one more reminder of how little we take advantage of the potential of these technologies even as they become fundamental to society. The good news is that this means we have barely tapped into the potential for productivity and economic value. Of course, that's also the bad news since the inability to take advantage of the technology creates a weak market for the technology itself and slows the virtuous cycle.

I got another reminder of the problem this week. I'll go into the travails of rebuilding my systems in a separate, much longer essay. One side-effect of the transition process is that one of the entries in my mail delivery list (in Geekspeak -- the MX records) resolved to zero instead of the correct IP address. The mail specifications says that if you can't deliver via one path, try another but apparently some mail delivery programs don't try very hard and take a spurious "cannot relay" as a reason to assume that the mail cannot ever be delivered. This has the cascading effect of sending bounce messages back to mailing list managers who then delist the user and there is no good way to trace back and undo the damage. To be fair, some list managers allow policies such as requiring the delivery fail for a week though one still can lose the messages sent during the outage. But that's yet another issue--the limited ability to treat posting and email as complementary delivery mechanisms thought Lotus Notes did that more than ten years ago.

When I first started using email in the 1960's it was viewed as strange. It took twenty-five years for it to become normal to use email. Please forgive me for my impatience during this long learning period and my concern that the original tools that worked within a single timesharing system don't work well outside that community. Perhaps it takes a techie to be intolerant of bad software. Most people simply assume that they must put up with the problems and not demand something better. After all, it's so much better than having to lick a stamp that it must be the best possible design. It isn't.

Sunday, February 02, 2003