Slashdot Mirror
Linux Counter Hits 120,000
meni writes "The Linux Counter Project now has over 120,000 Linux users listed worldwide. And with their brand new AlphaServer machine, they're ready to get slashdotted. If you haven't registered yet, please go over there and do." Okay, I just tested it out. W-a-y faster than it was in the past. I'm registered, and you should be, too.
When you install a smtp package one of the final options is 'send a local message from postmaster to root, send a test message to the linux counter project, dont do a test'
-Yarn - Rio Karma: Excellent
People are still quoting Red Hat's 7M figure from over a year ago. The number of users will have *at least* doubled since then, and my finger-in-the-air estimate would be more like tripled or quadrupled. Red Hat, or someone else with the money, would be well advised to sample again and see how many we can find.
--
Xenu loves you!
=)
--
It took some digging around, but I finally found some price information on the AlphaServer 800 (I really wish companies wouldn't be so coy about their prices... (no, I do not want to fill in a form so that you're sales people can call me back). Anyway, it looks like it's around $7000 (as of January). But the specs are 500Mhz, 64Mb, and a 4.3 Gb drive. If you want it configured with some real memory I would guess it's going to be an extra $1500 for 256Mb.
Press release, prices of AlphaServers
Maybe I'm confused, but isn't this still kind-of pricey? Compare that to these guys, who've got a 256Mb alpha machine with a 9Gb drive listed for less that $3000:
SWT Digital Alpha Linux System
Or course, this is technically a "workstation" rather than a "server". But I don't understand what exactly it is that you get when you ask for server hardware, except a ten grand price tag.
Oh, I would never suggest forcing participation.
I was thinking of things along the lines of:
- A message saying that it exists (This is one of the emails to root that Slackware automatically installs)
- A client program to make it easy to register which could be integrated into the system's install/configuration tools (eg: a menu in YaST or Linuxconf)
I agree that forcing it would be a very bad thing.
Or even better, when Redhat sets the default homepage for netscape (I don't know if this happens) to redhat.com, they could probably have a form to sign into the counter on their page.
Actually, this is a very good idea.
It's the easiest to implement for sure.
SuSE sets the default homepage to www.suse.de and I would imagine Red Hat does the same.
It's really just a plot to find out which of us actually use linux. You don't know what they'll do with this information. When a fascist totalitarian regime meant to oppress our resistance (read: Microsoft Fascism v2.0) comes about, will those who have registered be forced to wear penguin armbands and relocate to getto camps with only 9600 baud access?
You really should ask yourselves these things.
(ps, it's a joke)
Michael Chisari
I think you are way off base.
I've been using Linux continually since the 0.99pl7 kernel was state of the art. It has been my primary computing platform for nearly as long (since 0.99pl14).
My guess is that there are more than 120,000 Linux boxes continuously on the net acting as web servers. NetCraft counts over four million web servers running Apache. Estimates I've seen say that at least 1/3 of the Apache based servers on the net are running Linux. Personally I believe that number to be conservative, but even if only 10% of the Apache servers on the net are Linux based, its over 400,000. And that isn't even considering that some Linux based web servers are running Stronghold, Zeus, Roxen, AOLServer, Netscape Fastrack, or something else.
My guess is that only about 1 in 5 Linux boxes are actually used to run a web server on the publicly accessable Internet. If that is true, then I'd say that 2 million would be a bottom number for the number of hard-core Linux users.
I don't think the 10-12 million numbers are at all out of line if they include casual, occasional and part-time users.
I guesstimate that between 0.2% and 5% of all Linux users have registered with the Linux Counter. So the total number of Linux users is probably between 2,396,400 and 59,910,000 people.
Apart from the use of the appalling word 'guesstimate', my only reservation is the calculation of percentages; I'd suspect that the percentage of linux users is slightly higher than the given ceiling of 5%. Linux users are a different breed to Windows users; the majority have at least some inclinations towards zealotry/active advocacy. I'd say that it's more likely that 10% have registered. The number of linux users worldwide is left as an exercise for the reader.
It just occurred to me to check out my own registration; I was number 109251 when I registered on 24th February this year. According to the pretty graph, 110000 was reached in (approximately) July/august.
And if linux is taking off geometically as the graph here would seem to suggest, then there should have been considerably more than 10,000 in the last six months.
IANA statistician, but something looks dodgy. I'm not implying that microsoftian tactics were used, but it does seem a little sloppy; elucidation would be appreciated from anyone who's involved.
I had first found it while going through www.linux.org. But you're right. Hardly anyone advertises it, and it doesn't have a really easy to guess at URL. They need a URL like www.linuxcounter.com or something similar.
Netcraft has information on how their numbers are generated on their pages if you are curious.
There are undoubtedly some Linux users doing virtual hosting, but my guess is that most of the really large virtual servers are being run on big iron like Solaris. Linux is obviously more popular with smaller hosting sites that likely have fewer domains per machine. I'd guess that the average number of IPs for Linux web servers is closer to 1 than 2.
Most of the large hosting sites that use an Apache derivative (like CnG and Rapidsite) have changed the identity string so that they are no longer counted on Netcraft as 'Apache', which means that the numbers are not likely to be off by as much as you seem to be implying.
There is really no way to know for sure how much that multiple IP's per machine might affect the netcraft numbers, but if it is more than a few percent when spread across four million Apache servers checked by Netcraft, I'd be surprised.
Chorizo
dreamchaser asked:
;), there would be little incentive to code for Linux. But as you add zeros, it gets more and more attractive. 30! 300! 300,000,000!
Does it really matter how many people are running Linux? I can see -some- value in it, but people seem far too concerned with 'market share'. A free OS shouldn't care about market share, as long as developers are coding for the OS (which they are).
I think it is significant, because the idea of a "critical mass" enters heavily into software development and public acceptance. If there were, say, 3 Linux users world wide, all in Linus' family
That might be more of an issue to closed- than open-source programmers, but it still applies.
As far as public acceptance, well the past 12-18 months have seen huge leaps in public apprehension and understanding of Linux, even if it's overheard as "You heard 'bout this LYE-nucks doohickey?" Again, if it was just the Torvalds family sitting around the living room sending messages to each other with PINE, no story. But USA Today and Newsweek can cite geometric growth, or (soon) "a million registered users," or "the thousands of small businesses using Linux instead of (whatever operating system)."
That's why I think counting is a good idea. The numbers aren't everything (and as the other threads on here are mostly about) they're not easy to extrapolate. But they give a starting point!
timothy
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
I know maybe a couple of hundred people that use Linux as their primary system, including first and second-hand friends, acquaintances, and contacts. Of those, I am aware of only two that have registered with the Linux Counter. That doesn't mean that none of the rest have done so, but it does mean that it's not the done thing to register and let it be known that you have done so.
Extrapolating from this personal experience, I'd say that while the error margins on any estimate of total Linux population must necessarily be very large, the ballpark figure is probably massively underestimated.
Under a million and over 50 million would be unreasonable, but anywhere in between is not unreasonable at all. And remember that estimates of Microsoft usage count the number of licensed copies, so to be even-handed one ought to count all the released distros that each Linux user has obtained, ie. often a whole lot of them.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
For unrelated reasons I looked at the counter yesterday. It's very interesting; does anyone know who the single Linux user in the Wallis and Futuna Islands is? (Down at the bottom of the country list...) I can tell that person signed up earlier this year, and that they've elected to be private.
(I'm not picking on that person except that I happened to notice a 'one person per island listing' and that island sure looks nice right about now! I've yet to visit French Polynesia. Why not say hello to fellow Linux users in exotic locales?)
You can find out a scarey amount of info about people there. I'm glad that my listing (happy user since 1994!) is private. Most folk are letting a scarey amount of info about themselves be visible. Look at your town ...
PING counter.li.org (195.139.236.73): 56 data bytes
--- counter.li.org ping statistics ---
31 packets transmitted, 0 packets received, 100% packet loss........i'll hafta check later
AFAIK, Slackware is the only major distrib that encourages people to participate in the counter project.
I've installed RH, SuSE, Debian, Mandrake, and Caldera and have gotten no such message.
Considering the number of Slackware installations compared to the bigger guys, and the fact that not all Slackware users have registered (I registered years ago when it first started) I think I know of one reason that number is so ridiculously low.
Imagine if Red Hat had a message in there at the end of the install.
Hmmm, what about a "counter server" that lets a client program to add/change information about a user? How about integrating that into the installer?
The participation rate would greatly increase.