Slashdot videos: Now with more Slashdot!
It is purdent to note the paragraph describing the exact data before the age statistic:
The Los Angeles Times recently did a story detailing all of the NHTSA reports of Toyota “sudden acceleration” fatalities, and, though the Times did not mention it, the ages of the drivers involved were striking.
In the 24 cases where driver age was reported...
The key here is that the statistic is about fatalities, not occurances. When put into dangerous situations requiring strength (to push hard on a break pedal in this case), and good reaction timing (to avoid people on the road at 90mph or more), the elderly are always going to perform worse as age generally reduces both strength and reaction time, thus increasing the likelihood of crashes and thus fatalities.
This statistic isn't as glaring as the defense lawyer author wants us to believe.
RHEL5, released March 14, 2007, uses Python 2.4.3, which was released March 29, 2006. Given a reasonable package-freeze/testing/bugfix cycle, using this version seems about right. Also, Python 2.5.0 was released September 19, 2006 -- I know I wouldn't want to make a potentially major jump for all my system tools before publishing a major distro release.
Perhaps you should rethink the presentation of your point next time -- given what you've said already concerning RHEL5 and Python2.4, you should also be saying "RHEL5 uses Linux 2.6! That was released back in 2003!!!! ZOMG!!!"
In re: Python 3 migration, moving to the Python 3 series presents FAR bigger issues than addon-distribution, namely the changing and/or removal of some particularly widely-used items from Python 2.
I will agree with you that distribution of third-party modules can be annoying in Python, but that's not necessarily the Python developers' problem. Why should they be implicitly responsible for something that is third-party? Just because another platform is doing it? C'mon, that's a flimsy argument at best.
About a year ago, when the whole capped internet debate surfaced, I got curious about my own usage. I was using a Linux box as my router and so it was easy to gather statistics with vnstat.
After a few weeks of letting it gather statistics on my outward-facing ethernet port, I got an average of 6MB/hour (about 3MB up and down) that my roommate and I weren't home. This is before I started torrenting anything and only 4 devices (my computer, my roommate's computer, a Wii, and a Tivo) on the network.
If you do the math, that comes to about 2160MB per month without doing anything significant. Consider email, Windows updates, and light web-browsing for 2 users in one month, and I could see hitting the 5GB mark. If you add in Youtube, Hulu, some video news feeds, and iTunes, I think 10GB wouldn't be unreasonable.
I wouldn't say that my results are bullet proof, but given the above, I agree that the majority of email-checking, light-websurfing users would exceed a 5GB cap fairly easily. I would encourage other users to try this sort of bandwidth tracking for themselves.
Other questions come to mind. What about bot scans for vulnerable computers? Byte-wise, those can add up quickly. What happens if I get infected with a spam-sending or network-scanning bot and I don't know it? If I do 1TB of data in a month because of such a bot, am I liable for it?
I could also see this having Net Neutrality implications: "Oh, Google counts towards your 10GB quota, but Yahoo doesn't because we have a special 'content provider' agreement with them."