Filtering, email, and differences from Usenet

March 2, 2011

Yesterday I told the story of how and why I started filtering Usenet and suggested that the same evolution applied (or would apply) to email. Well, maybe. Today it's time for a counter-argument:

Email is fundamentally different from Usenet in one core way. For the most part (and definitely if you were not a frequent poster), Usenet was simply free-floating information going by. Some of it could be useful and some of it could be interesting, but none of it was essential and none of it was specifically for you.

Email is not like that. Or rather, some email is not like that. Some email certainly is like that; it is the same sort of general free-floating information and notification and whatnot, where it doesn't really matter whether or not you read it. But unlike Usenet, some of the email you get is specifically intended for you and is in fact actively important. It is not just useful to read that email, you need to read it.

You may or may not care if your spam filters take out some of the interesting free-floating information email (it depends on how far along the Usenet evolution you are). But you definitely will care if your spam filters take out email that you need to read, because reading only some of that email is not good enough. This means that you can never have the sort of aggressive, throw away almost everything filtering on email that people wound up with on Usenet.

Or to summarize pithily: on Usenet it was acceptable to throw away some of the baby along with the bathwater because it wasn't your baby. In email, it is your baby and throwing any bit of it is not acceptable.

(Well, okay, you could see such aggressive filtering on email if email became Usenet-like, ie if people stopped getting this sort of important stuff by email, leaving only the Usenet-like stuff. I tend to think that this would be a net loss.)


Comments on this page:

From 71.116.128.176 at 2011-03-07 04:01:04:

I use nmh and filter my list mail into separate folders. For a long time I would keep everything, but at some point I realized that (1) most of these lists are archived, and that I can probably find things in the archives more easily using a search engine than I can searching through my local copies, and (2) I was using lots of space to store local copies of these lists.

On top of that I read some of these lists carefully and frequently, and some I look at when I'm having an issue and I want to see if others are also having it, or when I'm bored and willing to take the time to see if there's anything interesting being discussed.

So I wrote a script that runs every night and expires old and read mail on a per-folder basis, taking advantage of nmh's commands to populate an expire sequence (with pick searches) and rmm that mail. The script loads a configuration file that specifies which folders to look at and sets expiration periods for read and unread messages, so I can keep everything, keep unread messages but rapidly purge read messages (for high-volume but important lists), keep a month's worth of messages (for lists I don't read often but might want to check in on), and so on.

I am also occasionally tempted by the idea of switching to using Gnus to read my MH format mail, which would give me access to all the killfile tools I could ever want, but thus far have resisted. (Switching to Gnus could also give me a path towards moving away from nmh to IMAP, which is something I have decidedly mixed feelings about.)

Written on 02 March 2011.
« The evolution of filtering (a story from the Usenet era)
The POSIX shell and the three sorts of Unixes »

Page tools: View Source, View Normal, Add Comment.
Search:
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Wed Mar 2 00:58:46 2011
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.