Well, that's unique IPs based over a long time. I would assume most active desktop users probably run within one or two of the latest versions. So it is certainly not 35 mil. Not to mention that I probably count as about 5 or 10 of their users given all the places I've run yum at on my laptop.

Anyway, who cares. It is on the order of 10s of millions, and that was my point.

I could be a bit out of date (yeah I am getting older ...)

If I'm reading the statistics correctly (http://www.census.gov/hhes/computer/) then about 14 million US households had computers in 1989. Given Ubuntu says they have 20 mil users and fedora seems to have about 4-5 mil based on their yum stats, I guess opensuse about the same, and debian and mint also takes some. Some of those are servers, but I assume there is a lot fewer servers than desktops in general, and I bet most ubuntu computers are desktops (I wouldn't be so sure). I guess it would be then conservative to say 14 mil linux desktops or the same number as us households in '89 (a bit more than 20 years ago and only counting US).

Anyway, my point was that if it is 10 to 20 million, that is a lot. Just because percentage-wise it is not everyone, it doesn't mean that it is insignificant.

If you give a whole bunch of money to a charity and save a hundred thousand people from starvation, you'd just save 1% of those that die every year. But would you consider it insignificant?

For whatever reason, when talking about software and technology people always talk about percentage of the market, rather than any absolute numbers. In other contexts people often take the absolute numbers.

If we discover an alien race in another galaxy that we have no way of contacting uses linux and has 100 times more people, will suddenly Mac and Windows become insignificant?

Well, there's probably more people using Linux on the desktop now than there were people using computers 20 years ago. 1-2 percent is a LOT of people (millions). If I publish a piece of software and millions of people use it, I'd say it is successful. Who cares about what percentage of the entire market it is. In absolute terms, there is an assload of desktop users.

I put a less than sign between 365 and 2010, but slashdot ate it

Sun did even better with solaris ...

And now MS office went straight to 365 ... it'll be hard to find easily recognizable numbers that are higher ... Hmm wait ... they actually have gone down, they had 2010 just before. Isn't 365 2010?

This is "version masturbation." There seem to be lots of people who all they do is upgrade their computer all the time. Then they are somewhat unhappy when a new version comes out once every few years rather than once every few months or days. If you don't actually use your computer, then yes, upgrading every few months is fine. If all you use your computer for is browsing porn, then yes, it is fine. But if you really don't care what version you have, you just want it to keep working, then this is nonsense.

Did anyone notice that Win XP is still very popular, and not just among businesses. Still it is the most used system. Why? Because why change if it works. The problem with upgrading is that it will inevitably lead to something breaking. If the computer is a tool, you don't want it to update automatically all the time.

Why is TeX/LaTeX still heavily used (e.g. in Physics, CS and Math community, the community it was intended for). It is not because a new version comes out every year. In fact, a new major version has really not come out in 17 years. Nobody is complaining. There is a LaTeX3 project, but it has been around for more than 20 years. A new version is not likely to come out. OK, some new macros have appeared in the meantime, but you don't have to use them. In fact, I know someone who still uses plain TeX to write papers (and uses unix "mail"), and is very productive with them (I have gotten him to use LaTeX for some of our collaborations though, and he does sometimes start "pine" for mail to send/receive an attachment).

But TeX/LaTeX is used by the scientific community that uses it as a tool. If a new version comes out, it will be years before it would get adopted.

The upside of updating to new firefox is minimal. Actually one can't really see much of a difference over all the firefox versions really. There are very few websites that won't work even with very old firefox. The downside is work interruption.

Some countries do better, but the rest of the world is moving towards how the US does it (some are already worse). The US is not the worst in the world in primary education. I can see the trend in Czech when I go back. Under socialism, there was no incentive to water down curriculum. There was a protected group (kids of communist party cronies and friends) and there was a repressed class (those people that didn't agree with the system or happened to be
born into such families). But if you happened to have parents that didn't piss off the commies (majority of the population), then the system was rather fair to you (within that class of people).

With capitalism, there is a movement away from "schooling is a right" to "grades are a service to be paid for." If you have rich parents, then the teacher doesn't want them complaining in all the wrong places. So quality of education (on all levels) is slowly declining even there. When colleges start being run primarily from private funds, then those providing the funds might stop unless their kids get the degree. It seems that majority of people do not see the folly with this approach.

I would partly blame grade inflation. Nowdays it is possible to pass all your classes without doing much work at all. Given that certain counties (according to the census) have nearly universal college attendance (95% in orange county in CA if I remember correctly) means that college curriculum must be brought to a level such that only the dumbest 5% of the population cannot obtain a degree (unless you make a convincing argument that being born in a richer county automatically adds braincells).

Point is, that the half that didn't get it wasn't supposed to get further. This isn't going to fly if current political climate all over the western world treats college education as a service to be provided for a fee.

So only US state department employees won't know the leaked US state department cables but everyone else will? Sounds like a winning plan.