SNARF from Microsoft:

On the topic of people/contacts:

Another example of uses of information attached to people and filtering/sorting/triage:

<http://news.com.com/SNARFing+your+way+through+e-mail/2100-1032_3-5979217.html?part=rss&tag=5979217&subj=news>

This is interesting.

Excerpt from CNET article:

"The software maker [Microsoft] this week released a free utility that aims to sort e-mail in a new way: It can organize messages not just by how recent they are, but also by whether the recipient knows the sender well."

I think SNARF, "in spirit" is very similar to Ducky's methodology for dealing with email. http://www.amazon.com/gp/product/0970885172/102-6521947-3550501?v=glance&n=283155

A question of approach

I wonder though if sorting by "how well you know a person" is just adding yet another "imperfect" way to read your email. (I'm not saying that this isn't useful. In fact, it sounds to me like a step in the right direction. The kind of metadata SNARF is attempting to capture is an important step on the way to better user interactions with email . My discomfort is more with how the metadata is presented to the user and how the user is expected to act on it. (As in, the email is sorted by "how important the sender is to you".)

The idea is that any single "aspect" or "attribute" of an email is an over-simplistic way to decide "in what order" a user will want to read and deal with their email.

There are simply

1. too many factors to consider
2. those factors change over time (sometimes mid-stream as you go through your email) AND
3. many factors live in the murky realm of the user's subconscious or "gut" understanding of priorities.

In sum, there are too many variables to consider and many of them are beyond the access of both the software and the user.

Variables such as: How much time do I have right now. How fried is my brain right now? How do I feel about the sender right now? What are the high priority, high urgency issues at work right now?

What if instead of having the software proposing a particular "order" to the user, the UI simply did a better job at exposing important meta-data about the email so that users can exercise their "better" judgement on what "email to look at first" without having to first scan each individual email (which is what we do today).

What are some of the characteristics or metadata of email that in aggregate would help users do this?

• Is this email To: me, CC: me, BCC: me or To: some list I'm on?
• Is this the first email? or the 19th email in a thread that's been going on forever?
• Is this email part of a thread, which I have already marked as Status: Dealing with right NOW.

Armed with this metadata, metadata the client software has access to, the user can then mix in other information, as in information only the user has access to:

• How tired am I?
• How much time do I have right now?
• What fire alarms just went off in the last 15 minutes?

I think there is a subtle, but important difference in philosophy and approach here that's worth teasing out. I think of it as:

Software that helps people make decisions VERSUS software that makes decisions FOR people.

In more specific terms: How can we help people make smarter decisions about the order in which they tackle email, rather than finding yet another way to order their email FOR them?

The original SNARF from Thundercats :o)