There's this interesting post by Paul Graham on distractions and time sinks (tv-then-internet):

"Something that used to be safe, using the Internet, gradually became more and more dangerous. Some days I'd wake up, get a cup of tea and check the news, then check email, then check the news again, then answer a few emails, then suddenly notice it was almost lunchtime and I hadn't gotten any real work done. And this started to happen more and more often. (...) The problem is a hard one to solve because most people still need the Internet for some things."

So what to do?

"At first I tried rules. For example, I'd tell myself I was only going to use the Internet twice a day. But these schemes never worked for long. Eventually something would come up that required me to use it more than that. And then I'd gradually slip back into my old ways. (...) The key seems to be visibility. The biggest ingredient in most bad habits is denial. So you have to make it so that you can't merely slip into doing the thing you're trying to avoid. It has to set off alarms.

Maybe in the long term the right answer for dealing with Internet distractions will be software that watches and controls them. But in the meantime I've found a more drastic solution that definitely works: to set up a separate computer for using the Internet. (...) If you try this trick, you'll probably be struck by how different it feels when your computer is disconnected from the Internet. "

Why do I blog this? interesting hint about work practices. It reminds me of a friend working in a big aerospace company where no personal computers are connected to the Internet (for security reasons) and where people have to go to an Internet computer (yes, in 2008).

Although I agree with the time-sink problem and suffer from it myself, I am still wondering about the definition of "work" Graham have. There are indeed different definition of work: - need to be connected to newsfeeds. - looking for intelligence, reports, material, hints, stats, etc. This requires first-hand sources or second-hand sources, search engines, tagging systems... - ...

(Picture taken from my temporary office in a swiss train)

And there are of course different recombination of work allowed by networks: - look things up on the google even for crazy things such as checking grammar (google fight), looking for a reference in a paper, etc. - need to access definition (wikipedia or urban dictionary!) - share and work on documents: Google Docs/Spreadsheet for example - use of certain websites to "compute" things, for example, I have few urls like that one to compute statistic things on-lines - communicate with people using Skype/Google Talk, etc. - update agenda - find picture on Flickr to illustrate a talk

Of course, all of this results from the choice I made (with colleagues) to use on-line tools. Personally I find more efficient to split my time between different moments/activities and places:

browsing/having a glance at stuff (daily read of my RSS feeds, websites, quick glance at magazine at news shops every day...), generally after breakfast. This is about news, people blogging about their activities or what they are doing. It's then a sort of ambient awareness of lots of things.
selecting few "signals" and turning them into something more concrete (a blogpost, in delicious, a note in a .txt file that corresponds to a specific project, and email to myself or friends), generally while/after browsing (morning)
reading "seriously" (on paper or on the computer), generally in a disconnected place (like trains!)
talking to people (whenever), eating with people (whenever), chatting (generally in the afternoon)
be on the field (observations) or in a work meetings (mid-morning/afternoon)
analyzing data coming from the field or writing seriously on a document (in a disconnected place sometimes) or with email/browser switched off.

All of this separated by breaks (walking, jogging, dreaming, taking weird pictures).