Wilfred

Funny you should mention that. There was a survey of favoured web frameworks[1] in the Lojban community, and Haskell with snap won[2].

1: http://www.surveymonkey.com/s/T6B69LK

2: http://teddyb.org/~rlpowell/public_media/lojban_survey.PNG

Python's docs are reasonable, but could be better. I don't like the left-hand contents, I find it breaks the flow. Wikipedia's inline contents seem far more intuitive. I'd like to be able to jump to specific functions.

(Moving to bpython solved most of these problem though, as it shows the docstring for a given function inline.)

Also, you end up spending a lot of time reading the docs and so a little typographical polish wouldn't hurt. The page is simply too wide.

I've built a little morphology parser in my spare time, and I'd love your thoughts.

http://metaoptimize.com/qa/questions/3103/morphology-parsing...

I'm fixing the sorry state of Esperanto online dictionaries. My system allows the user to use any of the major writing systems, tolerates spelling errors and even does some stemming and morphology analysis. It's been interesting because this has never been done before (AFAIK).

Fish is also an excellent alternative to the Perl one-liners proposed. Typing in part of a previously entered command then pressing up performs a substring search on the history. It's easy and adapted to your current usage habits.

That would seem to be consistent with the statistics of users with less followers: http://twitterfacts.blogspot.com/2008/06/tweetburner-clickth...

However I would consider clickthrough rates to be a poorly suited metric for Twitter. If you're not using Twitter to just follow your friends then your typical use case is just dipping in and reading some tweets when you have the time (the 'fire hose' model). Having one million followers does not mean that one million people will read a given tweet.

I would argue this already being taken advantage of (though 'exploited' seems a bit strong). There are pages such as this one: http://www.facebook.com/pages/96-percent-of-people-cant-figu... that require the user to 'like' it before they can see whatever it is.

It would be interesting to see what sectors are targeting this new spam avenue. Facebook's anti-spam approach is largely effective though -- as a centralised service they have a lot of behaviour metrics to characterise misbehaviour.

The author's intent is to write enough of the surrounding context that you recover it first time.

Still, it raises two questions:

What about fragmentation?

Why don't have a GNU safe-rm yet that moves files to the (freedesktop.org specified) trash location to avoid this?

I have used NFS for two years now without any complaints. If you're doing something small-scale, it's an excellent service at great pay-as-you pay scheme.

I assumed it was a tenacity thing -- note the dog has something in its teeth and is presumably very reluctant to let go.

Twitter, be being public by default means it doesn't have that intimacy that Facebook does. I'll follow someone if they're interesting, not because they're a friend.

But to answer your question, I find it fantastic for foreign language learning. Real, native speakers without travelling long distances.

Still, at $20/month I use a small $100 plug computer (such as http://en.wikipedia.org/wiki/SheevaPlug) as a home server instead. Your use case may be different to mine but that is more than I'm willing to pay within a student budget.

Ctrl or alt foot pedals are the way to go here -- using caps-lock is a workaround but in proper typing I wish to use the modifier key on the other hand to the letter.

The LinuxMCE is no thing of beauty yet, though. Here's hoping this changes soon.