I have been using Fedora since FC3. Used to use Mandrake before that. I'll have to check 21 out tonight, but my gut feeling is that it's not going to go so well. I believe the last version of Fedora that was rock-solid stable and had support for pretty much anything I threw at it was FC18. For the sake of diversity, I run Ubuntu (XFCE) on my desktop at home, FC20 (XFCE) at work, and CentOS5 and CentOS6 on all the servers I'm involved in.
One of the botches I believe FC team did was when they changed the interface for the hard drives during the installation. Yeah, I know, I switch to console and fdisk and parted everything the way I want it, but the GUI used to be really simple before they changed it.
I seriously doubt it'll be $10/month. Assuming a new O/S off-the-shelf costs about $80, and that a user is going to use it for 4 years, the cost is just under $2/month. MS practically gives those licenses away to OEMs, so the gross sales for MS would be about $1/month/PC. If they charge you $10/quarter they'll have tripled their revenue.
The question is, what will subscription get you. Does it get you the opportunity to upgrade to the latest OS from MS? Or does it merely get you hotfixes? If it's just hotfixes, then they're really shooting themselves in the foot, since many PCs will go unpatched and even more stolen identities, etc. will be attributed to Windows machines. I suspect they'll go to something like $x/year will get you free upgrades.
Sure, a set of anonymized, randomized set of users at the beginning and ensure they remain anonymous throughout the study, then do the study. The question is whether FB can truly anonymize the data they are studying. I would place a wager that they cannot. There is so much information creep in FB that anonymizing the data may not be possible.
Second solution, give the research projects to people who truly have no interest in the data or the results.
Looking at their example conversation, I had no choice but to face palm. Having never looked at Ruby code before, I was able to deduce perfectly well what the first iteration was doing. Do we really need to expand a function that can accomplish its task in one line into a function that may be a little more readable?
I wonder how today's programmers would make do with resources that were available in the early days of computing, or even when the IBM PC came out. Having to deal with small amounts of RAM caused programmers to be extremely creative in their programming. Granted that we do not have to go to such extremes today to write programs, reading about such practices is still very inspiring.
According to the article, out of 800 positions created under DOJ's COPS program, 629 MUST go to veterans who have served at least 180 days of active duty since 9/11. Although this does not provide statistics for the existing law enforcement population, it does provide some insight. There are also numerous articles on the web that talk about transitioning veterans to local police forces.
Happiness is twin floppies.