“Here’s a theory you hear a lot these days: “Microsoft is finished. As soon as Linux makes some inroads on the desktop and web applications replace desktop applications, the mighty empire will topple.” Although there is some truth to the fact that Linux is a huge threat to Microsoft, predictions of the Redmond company’s demise are, to say the least, premature.” Read Joel’s commentary here.
How Microsoft Lost the API War
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
“Yes, I have been watching Google Zeitgeist for almost four years now, and Linux seems to be consistent at one percent.”
thats because it was hyped very early. Microsoft has reasons to be worried. thats very true. remember how netscape was at 90% only a few years before?. where was IE then?. now look at the same google zeitgeist?. where is netscape?. so dont take the current stuff as something guaranteed.
longhorn is a big change. so much as a change that people will hesitate. thats the key problem joel is highlighting here.
Because Linux doesn’t need marketshare. It needs Love, and its user give it that. Better have a few that are happy, than many that complain.
People like Microsoft for some reason. Why would they switch? They would probably only switch if Linux was made by Microsoft.
Joel is spot on in this article. Basically he says Redmond are churning out tools faster than people need them.
Recently I have ‘switched’ to Python for personal projects. Far better than .Net and Java on my older spec Laptop but more importantly it is a stable platform who I can trust to be around in the same core form in years to come!
That was a very intersting article.
It mentioned things that I had never given much thought to, and put into better words some things I already think (In real life when you shout “shut up!” to two people arguing loudly you just create a louder three-way argument.). It’s as good an argument as any I’ve heard concerning things like competing standards, and any sort of attepmt at unification, or solution to precieved problems regardless of the system in question.
POSIX/LSB, CVS/Arch/Subversion/BitKeeper, Win32/.NET, GNOME/KDE/GNUstep, Debian GNU/FreeBSD etc.
“POSIX/LSB, CVS/Arch/Subversion/BitKeeper, Win32/.NET, GNOME/KDE/GNUstep, Debian GNU/FreeBSD etc.”
your examples arent very convincing.
lsb is a set of package specifications unlike posix which only defines the api. cvs and subversion are centralised systems while arch and bitkeeper are distributed stuff. both with different idealogies and usage scenarios.
win32 and .net is right.
debian is not just a linux distro. there are efforts to create a freebsd/netbsd/open bsd too including gentoo’s efforts.
gnu doesnt have a good kernel implementation yet. almost everyone uses linux with gnu.
freebsd is under the bsd license which has a very different idealogy from debian free software guidelines.
so i think what joel talks about is api compatibility in the same operating system. he is arguing that such a major leap is unprecedented and that MS has very little experience moving people from one platform to another and this is problematic to developers in a huge way.
for example asking them to move from gtk to qt or something like that
“your examples arent very convincing.
I’m not trying to convince anyone of anything here (least of all you, as you’re the most closed minded person I’ve encountered online), I’m merely stating my thoughts on a very intersting article. ?
how about saying something that makes more sense
I think the API war hasn’t begun.
Nobody is ready yet. No battle, no winner.
Microsoft will keep its monopole until at least 2 years. But when the 2 other platforms (linux and bsd/osx) will move on its land (desktop users) with compatible APIs, windows will start to become no more needed. Simply because Joe applications will continue to work just like before. Independently of the platform he uses.
There is no industry without standard.
I have 2 comments:
1. Web clients vs. rich user interfaces
I have long wondered why web interfaces aren’t much good. The technologies are there; Java applets, Flash, Python could do it, JavaScript could with a few extensions, XUL, heck, even C, compiled on the fly. All these stop just short of integrating well with the web and the client platform. Why? Why has nobody managed (or tried) to take the last step? Even .NET doesn’t go there; the GUI interface is very much tied to Windows.
2. API changes
Contrary to Microsoft’s, UNIX and POSIX APIs have been very stable. There have been extensions, but the bulk of what used to work still works. This makes the case for switching over to standard technologies, now that Microsoft is pushing you to switch anyway.
It’s an interesting article. And he’s right about Microsoft tossing out development tools for free, or nearly. I have a full copy of VisualBasic.Net Standard Edition (not for resale) right here that came with 5 CDs, and getting started manual, that I got just for the asking.
His main point is a good one. Microsoft has given up on compatability and has shattered and remade the API, building ever more complex development environments. And people are leaving it behind for the web.
I can see his point, but people have been predicting the rise of the webapp for years and we’ve yet to see it happen really. Yeah, some stuff has moved to the web but I don’t see it supplanting the desktop environment anytime soon (if ever), however much Microsoft screws up Windows.
I have to agree tht desktop Linux is further off that most would expect, 2004 is again the year of the Linux desktop and here were are more than half way through the year with little progress in terms of market share.
I remember telling people they were wrong back in 99 when they claimed Linux would dominate the desktop in 2000, 2001 etc. And not much has changed since then acceps Linux distros have better installers now, Windows is stable, and they both gained some bloat, more so Linux than Windows.
I remember the hype from when I first heard about Linux, about how it was faster and more stable.
I think in many situations this is no longer the case, if it ever was.
Joel has some good points, but I really don’t see the new API having much of a negitave impact on windows. To many people this is a welcome change.
:Why? Why has nobody managed (or tried) to take the last step? Even .NET doesn’t go there; the GUI interface is very much tied to Windows. ”
xul has been cross platform a long time before MS came up with avalon. there isnt much adoption outside mozilla. MS can push things better.
“have to agree tht desktop Linux is further off that most would expect, 2004 is again the year of the Linux desktop and here were are more than half way through the year with little progress in terms of market share.
”
yes. i have already told this. its was hyped very early. Desktop enrichment requires a lot of investment which distributions have started to do only now. linux has been much more established in the server and its likely to be stronger there for quite sometime than the desktop
“Joel has some good points, but I really don’t see the new API having much of a negitave impact on windows. To many people this is a welcome change.
”
MS has always valued backward compatibility even with ugly hacks than breaking things. this is a very big change that many developers dont like. this is likely to have a bigger impact due to delays and migration issues than you think
[i]… they could reinvent themselves as a shaved-ice company at the last minute.[i]
Could this reinvention be in the form of making Microsoft’s products (Windows, Office suite) retail for $20 (or less) for each license? I am asking this question because I am getting email offers (which I am not interested in) to buy Windows XP Pro for $40, and Office XP Pro for $50.
People like Microsoft for some reason. Why would they switch? They would probably only switch if Linux was made by Microsoft.
He’s not saying everyone’s going to switch. He’s just saying that Microsoft’s prospects for lock-in and growth into new fields are shrinking.
Personally I’m very glad to see a critique of the Longhorn/.NET whatever hype from someone who used to work as MS. Especially since we’re now getting it not only from MS marketing but also some of the Gnome team as well.
I also liked his point about managed/unmanaged code vs. object-oriented/procedural.
I think that the cross-platformed nature of web development has changed everything (and therefore agreed with Joel on many of his points).
The Win32 API sucks, but millions of developers learned the thing in annd out because it was relevant in the PC desktop era, it is no longer so. Even when .NET framework and platform is easier to develop for (than Win32), it seems that not many care. The reason MS changes their development methodologies so fast is because they need to create something that developers will use and then they could dominate the API war again, trying to create a hype (Hailstorm anyone???)
Longhorn or not, I don’t think anything MS has to offer on Windows matters anymore… Even if I have a rich client app, i would more likely use development models that are cross-platformed and have more “reach”, just as I would with a web-based app (something like Java, wxWindows, Qt, Python, Flash, …). MS is not dead, they just matter less.
I have been watching Google Zeitgeist for almost four years now, and Linux seems to be consistent at one percent.
Google Zeitgeist is not an accurate measure of market share. This has been discussed in detail before, but the real Linux market share is more likely to be two to three times higher than what Google indicates. So your entire argument seems to be based on a faulty assumption.
Those email offers for cheap XP Pro and Office are from bootleggers. Microsoft doesn’t compete with bootleggers. Nobody does. Microsoft needs to lower the cost of their prime products, but not because of illegal copies floating around the net.
“[snip critique]
Did you notice that that’s a quote, not his own statement, and that he disagrees with it?
”
essentially he says that billions of dollars of reserve means that they can continue doing mistakes and still retain themselves or reinvent themselves in a diff market like ibm did
“But there’s a price to pay in the smoothness of the user interface. Here are a few examples of things you can’t really do well in a web application:
2. Build a real-time spell checker with wavy red underlines”.
Joel has obviously not tried many of the web mailers, such as Fastmail that have a built-in spell checker. Additionally, if you are on Linux, konqueror the browser provides the wavy curled underlines that Joel likes so much. So at least, some clients would get this as a free benefit. It wouldn’t surprise if future versions of Mozilla and IE incorporated similar functinoality.
“So at least, some clients would get this as a free benefit. It wouldn’t surprise if future versions of Mozilla and IE incorporated similar functinoality.”
actually their implementation is pretty different from the MS word thingy.
btw mozilla has spell checker plugins
http://ftp.mozilla.org/pub/mozilla.org/mozilla/releases/mozilla1.6/…
http://spellchecker.mozdev.org/
“This has been discussed in detail before, but the real Linux market share is more likely to be two to three times higher than what Google indicates”
I think I missed that discussion, could you please enlighten me as to what is wrong with looking at Zeitgeist statistics?
I know from my personal experience, only about 2% or 3% of the IT people I know use Linux as a desktop. (I am one of them). Windows is a standard on the desktop for every business, school, and library around here. I have done Linux support for several IT people “taking the plunge” and Linux is hardly without several major hurdles ahead.
I think 1% is probably pretty accurate, I also thing that out of that 1% most of them are “IT professionals” and many of them are very very wrongly informed.
I am proud to say that I belong in the .0000001% of people using Linux that does not have a political agenda against MS or a false understanding of what the Windows operating system is capable of.
I might even be the ONLY Linux user without an anti-MS political agenda.
“Nobody (by which, again, I mean “fewer than 10,000,000 people”) wants to develop for the Windows API any more. Venture Capitalists won’t invest in Windows applications because they’re so afraid of competition from Microsoft. And most users don’t seem to care about crappy Web UIs as much as I do.”
“Much as I hate to say it, a huge chunk of developers have long since moved to the web and refuse to move back. Most .NET developers are ASP.NET developers, developing for Microsoft’s web server. ASP.NET is brilliant; I’ve been working with web development for ten years and it’s really just a generation ahead of everything out there. But it’s a server technology, so clients can use any kind of desktop they want. And it runs pretty well under Linux using Mono.”
These two quotes from a former Microsoft employee are incredibly telling. He is basically telling you that developing for windows isn’t economically sustainable as it is nearly impossible to compete with the Microsoft Behemoth when you have to compete on their terms and run applications on their platform.
When joel says that apple has a 2.3% marketshare, he’s refering to the sales of apple hardware for this year (or the last one) but even if everybody knwo that mac people don’t change their machine every 2 years, they’re always keep telling 2% but I have a lot of friends of mine who bought their imac in 1999 or 2000 and they’re only consider switching or they’re just want to buy some ram or whatever…And I suppose they’re not alone doing that…
In fact I do know how they compute the selling rate but the market share ?
for the console market they add the totality of the ps2 sale since its launchs but not for the mac ones ?
and for the pc, how do they know wich one is an upgrade, which is a new one, which is for linux, for whatever ???
it sounds like crystal ball to me
The reason that google underestimates the number of LInux clients is because it only counts hits that come from different unique public IPs. Some of my clients have as many as 30 Linux desktops behind a cheap DSL router with one public IP.
All of these Linux clients are going to be counted as one. The success of consumer-grade routers for home-networking is huge. You cannot tell me that you are not aware of this. Well, if you have a few people running Linux in your household, google only counts it as one single user.
Many people disguise their browser identification string in order to fool silly sites that block access to you if you are not running IE on Windows.
Additionally, believe it or not, not everyone uses google. While google is well known around the world, it doesn’t have the same degree of penetration in Asia and Latin America,two of the areas where a lot of the huge Linux deployments are taking place.
Finally, many Linux clients installations are within intranets for data-entry or in places where internet access is not desirable. There simply isn’t a way for google to count these.
I think I missed that discussion, could you please enlighten me as to what is wrong with looking at Zeitgeist statistics?
Sure, though I’m getting tired of repeating it…
There are two main factors that contribute to a lower Linux count on Google Zeitgeist: browser ID strings, which has a minimal impact, and static vs. dynamic IPs, which has a much larger impact.
You’re probably already aware of the web browser ID string issue: most Linux browsers can masquerade as other browsers/OS combinations in order to fool certain sites that will only provide content to supported browsers (even though other browsers may handle the information just fine). Personally, I don’t think this accounts for a lot. Maybe 10 to 20% of Linux users, at the most. The issue of static vs. dynamic IPs is much more significant, in my view.
As you may know, about 35% of North American Internet users have broadband, while 65% still use modems. However, the use of broadband among Linux users is much higher – it’s hard to get exact figures, but the Linux counter indicates that as much as 85% of Linux users have access to Broadband. This makes sense, as a lot of Linux users will download iso images of their favorite distributions and/or update their Linux installation through the internet, which can be a gruelling experience on a 56K modem…
What’s that got to do with the Google Zeitgeist figure, you may ask? Well, what Google counts are visits from individual IP adresses and logs their browser/OS ID strings for statistical purposes (you can do that for your home web server with a variety of tools – I use a great perl script called Awstats).
This means that narrowband users, who frequently disconnect from their ISP, are more likely to count as multiple users when connecting to Google, since they will get a new IP address every time they connect. Broadband users, on the other hand, are more likely to keep the same IP address each time they visit within a single month (especially cable users – my IP addresses usually last for 6 months or so…). Since Linux users are primarily broadband users, they are much more likely to be counted less than average Windows users, who are more likely to be narrowband users.
So there you go. This is why no one – except the odd pro-MS advocate – will take the Google Zeitgeist figure seriously when it comes to representing market share.
I know from my personal experience, only about 2% or 3% of the IT people I know use Linux as a desktop. (I am one of them).
Nothing personal, but “personal experience” is never a good source for statistical figures. The sample is too small, and the data collection is likely to be biased.
I am proud to say that I belong in the .0000001% of people using Linux that does not have a political agenda against MS or a false understanding of what the Windows operating system is capable of.
Please kindly stop pulling statistical figures out of your *ss. This is equivalent to me saying that I belong in the 50% of people using Windows who do have a political agenda against MS (or at least dislike them a lot)…
Seriously, let’s try to keep it real, m’kay?
“I can see his point, but people have been predicting the rise of the webapp for years and we’ve yet to see it happen really.”
The thing you need to ask is “Are the pieces in place, or soon will be?” instead of going “same old, same old”. That’s how people get broadsided by technology.
“Yeah, some stuff has moved to the web but I don’t see it supplanting the desktop environment anytime soon (if ever), however much Microsoft screws up Windows.”
Mainly the reason is that “web apps” are happening on Intranets, not the Internet. Change doesn’t always happen right in front of us.
[A nun, he moos (IP: 67.71.241.—) ]
“Sure, though I’m getting tired of repeating it… ”
Save a copy on your HD, and C&P when necessary.
“Although there is some truth to the fact that Linux is a huge threat to Microsoft, predictions of the Redmond company’s demise are, to say the least, premature.”
yes, premature but it will surely be sure.
Save a copy on your HD, and C&P when necessary.
Yes, good idea, I will. Thanks.
>I think I missed that discussion, could you please >enlighten me as to what is wrong with looking at Zeitgeist >statistics?
I would look at it differently. Google Zeitgeist is giving client stats. The places where linux/BSD is making the most impact at the moment is the server. The server is where the major players including microsoft want to be since it gives them a nice revenue stream instead of a one off payment.
Linux on the desktop is coming I expect but it is the server market that has microsoft scared the most.
I can see his point, but people have been predicting the rise of the webapp for years and we’ve yet to see it happen really.
Oh, it absolutely has happened. Not for word processors or ides or spreadsheets and that stuff – that probably wouldn’t be a good idea anyways; plus the market is already saturated.
But it’s already in place for, say, tax software[1], email[2], collaboration software[3], not to mention a whole host of custom database-driven apps.
Also, if the web app hasn’t happened, why are most .NET programmers working with ASP.NET?
[1] http://www.turbotax.intuit.com/
[2] yahoo, gmail, hotmail, most college campus email systems; even outlook now has a web client.
[3] e.g. http://www.collab.net
The server is where the major players including microsoft want to be since it gives them a nice revenue stream instead of a one off payment.
I would not be so sure. Look at Google: it runs 100,000 Linux servers as somebody speculates. Yet, I very much doubt it pays a dime to any Linux vendor in subscription, entitlement, service (or any other doublespeak) fees.
Look at http://www.akamai.com: almost 15,000 Linux servers. Please point a Linux vendor Akamai provides with revenue stream.
No, not at all. Server market is as good as dead.
Business desktop market is a source of a revenue stream.
Home users- hard to say, if some kind of truly free of charge Linux makes to home desktop- may be. If not, then Microsoft will keep its low cost ($45/desktop) offering and get nice revenue stream from every new computer sold.
Take away desktop from Microsoft- and it’s over for big guy.
Which means, 1% year after year, even if it corresponds to true 5% of Linux desktop users year after year,- good news for Microsoft.
Google Zeitgeist may not be giving correct absolute numbers, but it gives a trend. Ups and downs.
I am getting sick of this. Linux has is not only ready for the desktop but it has already surpased it! I’m running SuSE 8.2 pro without any av or constantly huge patches that in the end also need to be patched.
I’m ALSO using FC1 on 24 mb ram with only a 2gig HD and that s WITH KDE 3.1 with the standard 2.4.22 kernel!
The only arguments against Linux on the Desktop come from m$ zealots that use the same old arguments.
Linux doesn’t use any standards
Of course it does! Thats why, no matter what disto you use, the same software is available for it.
There is no support
Of course there is. Infact there tons of books avilable as well as tons of support from the web!
I could go on and on.
If someone promotes linux then that person is automaticaly casted as someone with ‘anti-m$’ rhetoric.
Its not the fact that its some sort of ‘anti-m$’ properganda but maby its the fact that Linux makes a better operating system overall!
http://www.linuxelectrons.com/article.php?story=20040115100015615
http://www.google.ca/linux?num=100&hl=en&ie=UTF-8&safe=off&q=%2…
Yes, on server market money comes from hardware sales and selling warranty for hardware and hardware related services – plus free of charge OS named Linux on top of hardware.
Thank you for helping me with my point of lost server market for software vendors who charge money for their OSes in the form of license fees and service fees.
“Thank you for helping me with my point of lost server market for software vendors who charge money for their OSes in the form of license fees and service fees.”
Evolution = Adapt or Die! Go Free Market!
These two quotes from a former Microsoft employee are incredibly telling. He is basically telling you that developing for windows isn’t economically sustainable as it is nearly impossible to compete with the Microsoft Behemoth when you have to compete on their terms and run applications on their platform.
If they can’t compete against MS, who depend upon a paid staff to produce software, then in what universe do you think they can compete in an OSS environment where opposing projects have unpaid/voluntary staff?
It is not worth my time to argue with you but,
Of course there is. Infact there tons of books avilable as well as tons of support from the web!
Have you been locked in the basement for too long to realize that most people are not going to buy and read Linux technical books and post to IRC and newsgroups for help?
This is something even few IT people do, I was once called to someone’s house to install a wireless router. Sure this task might be more than most users are ready for, but this person held a BA in Computer Science and works in a store that sells computers (and wireless routers).
I have found with some cases, even if I can replace everything people used in Windows with a Linux equiv, they are annoyed by many of the little things like “Gaim does not flash in the task bar to let me know I have a new message”
I never hear about software installation problems because I usually handle all of that stuff, few people are ready to move off windows, even if they have very limited computing needs.
not trying to troll, but NT4 is still more suited to today’s Joe user than today’s Linux distributions and I dare say but quite a margin.
I can’t say that I really want everyone on Linux, I would prefer it be a light fast solution aimed at power users, but it seems that battle has already been lost. At least windows is arguably easier to turn off all the annoying defaults aimed at joe users.
“As you may know, about 35% of North American Internet users have broadband, while 65% still use modems. However, the use of broadband among Linux users is much higher – it’s hard to get exact figures, but the Linux counter indicates that as much as 85% of Linux users have access to Broadband. This makes sense, as a lot of Linux users will download iso images of their favorite distributions and/or update their Linux installation through the internet, which can be a gruelling experience on a 56K modem…”
Here’s a few more arguments which describe it’s inaccuracy:
1) It only calculates desktops/workstations which have webbrowsers and visit Google. Far from every workstation/desktop in this world do. Yes Google’s marketshare is high, but not very near 100%.
2) Thin client set-ups distort. Minor influence i think.
3) Even Netcraft isn’t accurate. Ever heard of an internal webserver? Netcraft knows nothing about these. Neither does Google know anything about internal clients.
4) And possibly it calculates per IP (if i understood the above correct) which means a company with 50 clients behind NAT only gets calculated 1 time. Ouch!
5) One can run a CLI browser on a server. Nice if you’re maintaining and need to look up something fast, ie. search for a patch.
I’m not saying either of these push up the “Linux” usage. It is hard to say.
Anyway i find 1% slightly impressive. What tops it? No proprietary Unix except perhaps MacOSX. No BeOS or any other of the GUI’s and there are about 100 times less Microsoft Windows computers. Imagine how many Microsoft Windows computers there are on this world. Divide it through 100. That’s still a damn lot right?
>I would not be so sure. Look at Google: it runs 100,000
>Linux servers as somebody speculates. Yet, I very much
>doubt it pays a dime to any Linux vendor in subscription,
>entitlement, service (or any other doublespeak) fees.
That is not the point I am making.
Microsoft makes money from the server market by selling a server OS with ongoing maintainence contracts and support fees. Linux does the same job but does not incur a licence fee.
Support fees and maintenance is an area that many linux vendors (such as Red Hat) are happy to play in.
I don’t know the specifics of Google and their support. They may run it all themselves but most companies do not. They will gladly pay a yearly fee to a company to support them. This part of the market is HUGE and there is lots of money to be made.
Linux directly threatens microsofts cut of this market.
Linux on the desktop will eventually threaten microsofts market as well but it is the server space that linux is gaining ground.
Cheers
David
A browser hit on Google tells Google _nothing_ about the person who’s behind it. A family with one computer, father, mother, 3 kids using 1 computer is likely while a geek with 3 computers with which he alone is surfing is also possible. Similar situations are possible in ie. companies. Zeitgeist nor marketshare say anything about that.
So to analyze, one could use Zeitgeist + a way to debunk above problems to get an accurate state of the world’s usage on operating systems. It won’t be 100% accurate, but it can as near to 100%, and this one has a few, huge flaws which have been mentioned throughout this thread, and others.
“I would not be so sure. Look at Google: it runs 100,000 Linux servers as somebody speculates. Yet, I very much doubt it pays a dime to any Linux vendor in subscription, entitlement, service (or any other doublespeak) fees.
Look at http://www.akamai.com: almost 15,000 Linux servers. Please point a Linux vendor Akamai provides with revenue stream. ”
both these people have paid redhat before. talk about stuff you know
“If they can’t compete against MS, who depend upon a paid staff to produce software, then in what universe do you think they can compete in an OSS environment where opposing projects have unpaid/voluntary staff?”
1) money doesnt automatically create best software. if so whatever MS produces must be the best which is clearly not the case
2)OSS is not entirely done by voluntary people any more. novell,sun,ibm,redhat et all pay people to develop a good amount of OSS stuff
Google Zeitgeist is not an accurate measure of market share. This has been discussed in detail before, but the real Linux market share is more likely to be two to three times higher than what Google indicates. So your entire argument seems to be based on a faulty assumption.
The point isn’t whether or not Linux as 1%, 2%, 3%, 5% or whatever – the point is that number has remained static for the ~4 years he’s been watching.
There are two main factors that contribute to a lower Linux count on Google Zeitgeist: browser ID strings, which has a minimal impact, and static vs. dynamic IPs, which has a much larger impact.
I think you’ve got that back to front. The browser ID strings are going to skew the results much more than NATs and static vs dynamic IPs.
You’ll have to work pretty hard to convince me there are a greater proportion of Linux machines hiding behind NATs, or attached to broadband, than Windows machines at _all_, let alone enough to be statistically relevant.
The number of Windows machines also behind NATs, or attached to broadband, should more than balance out the number of Linux machines in the same situation.
In short, I can’t see why these two aspects of the marketshare don’t balance out, from the perspective of statistical relevance.
Till few days ago, i used to believe that it is hard to make rich web interface, but few days ago i came across a jabber IM client “rhymbox”, their interface is partially web-based and its so sleek, much better than most IM clients available, including yahoo and MSN.
After looking at their UI design, i must say that i am looking forward to web or XAML based clients. They are much more rich than traditional clients, you can easily embed pictures, audio, video in them. They are just cool, i would say.
Pankaj Garg
http://www.intellectualheaven.com – Home of E_OatBot – A useful chatbot
People are getting fed up of windoze and .Not issues. It seems like no one wants to use .Net and windows any more. Linux and Java is preferred way to go.
Till few days ago, i used to believe that it is hard to make rich web interface, but few days ago i came across a jabber IM client “rhymbox”, their interface is partially web-based and its so sleek, much better than most IM clients available, including yahoo and MSN.
After looking at their UI design, i must say that i am looking forward to web or XAML based clients. They are much more rich than traditional clients, you can easily embed pictures, audio, video in them. They are just cool, i would say.
Pankaj Garg
http://www.intellectualheaven.com – Home of E_OatBot – A useful chatbot
Here’s a few more arguments which describe it’s inaccuracy:
1) It only calculates desktops/workstations which have webbrowsers and visit Google. Far from every workstation/desktop in this world do. Yes Google’s marketshare is high, but not very near 100%.
2) Thin client set-ups distort. Minor influence i think.
3) Even Netcraft isn’t accurate. Ever heard of an internal webserver? Netcraft knows nothing about these. Neither does Google know anything about internal clients.
4) And possibly it calculates per IP (if i understood the above correct) which means a company with 50 clients behind NAT only gets calculated 1 time. Ouch!
5) One can run a CLI browser on a server. Nice if you’re maintaining and need to look up something fast, ie. search for a patch.
These things don’t matter is because there is no reason to believe they apply to only a single platform. In other words, there’s no reason to think there are (proportionally) more Linux machines behind NATs than there are Windows machines, etc.
The zeitgeist is good for getting a general “feel” for how the market is divvied up. It’s not, by any stretch, completely accurate, but there’s no reason to think it’s unrepresentative by any statistically significant amount. Moreoever, even if it is misrepresenting marketshare (eg: Linux is really at 5%, MacOS is at 10%, etc) it should do so consistently – so using it to observe trends should be fairly valid (ie: Linux marketshare doesn’t appear to have gotten any bigger).
I’m not saying either of these push up the “Linux” usage. It is hard to say.
I’d say it all balances out. For example, a larger proportion of Linux users might be on broadband, but a smaller proportion of Windows users are probably using google (as opposed to Yahoo, MSN, etc), since it isn’t a default setting.
“The point isn’t whether or not Linux as 1%, 2%, 3%, 5% or whatever – the point is that number has remained static for the ~4 years he’s been watching.”
(On the desktop.) And that seems slightly unlikely to me given the commercial Linux distributions sell solutions for the desktop as well.
Also don’t forget you cannot say to every point “doesn’t influence the conclusion much” and then conclude it is accurate because all the points “do not influence the conclusion much”. They hav to be added up since this is a statistic; when calculating statistics it is bad practice to round up when you’re gonna use that number again. You don’t make an int or round up when you’re gonna use it again. If you want to calculate Z, and use A + B + C = D and E + F + G = H then use D + H = Z while D and H were rounded up/down you end up with a flawed end result!
How about my first point regarding the lack of an Internet connection or lack of usage of a browser?
Hmm… Oops? Here’s a version more friendly for the eyes.
“These things don’t matter is ecause there is no reason to believe they apply to only a single platform.”
There is no reason to believe that, but there is a problem with the “because”. If it ie. favors X over Y, then it matters with the additional question “how much”.
“In other words, there’s no reason to think there are (proportionally) more Linux machines behind NATs than there are Windows machines, etc.”
As pointed out above proportionally (or slightly) more X than Y is something different than no X or Y at all on any of the earlier said examples.
My point is: Zeitgeist does not research that at all. Several effects on the statistics are ignored. Which does mean the result is more inaccurate than when that would have been taken into the analysis.
What we (will) disagree on is the relevance of the inaccuracy while i’m pretty much open-minded to ie. offline computers, having several cool examples in my mind. Why the heck have Internet on your computer which is only serving you the ability to use a CAD program?
Yet another thing which remotely matters is the relevance and usage of the computer…
Thank you for helping me with my point of lost server market for software vendors who charge money for their OSes in the form of license fees and service fees.
So MS and Sun lose money. Boo hoo. The first one could survive for years on its cash reserves alone, while the second one would lose to the first one anyway.
Red Hat, Novell, HP and others have profited from Linux in the server market. Microsoft hasn’t. Them’s the breaks in a market economy.
(As for Sun…well, Sun has a problem finding its place in the emerging server market.)
As usual, you try to paint Linux as a catastrophe for the IT economy, yet there are in fact only a few companies losing out because of it, one of them being an abusive monopoly who has 40G$ stockpiled in cash.
I hope you’ll forgive me if I’m thoroughly unimpressed with your doomsaying…
This is even more important to your point: “because there is no reason to believe they apply to only a single platform.”
This is actually an argument against yours
here’s why: if 99% of the desktop computers which are not hooked on the internet (an example part of my earlier argument) and are not using Linux -which is almost the same as “they apply to only a single platform”- then it does not affect the statistic. If more than 1% of the whole world does, it does have an effect on the statistic.
Here’s a question for you: how many Amiga’s and C64’s on this world, according to you, are running a version of their original OS AND have an internet connection?
(On the desktop.)
Well, the desktop – specifically the business desktop – is where it matters. Everything else is either influenced from there or an insignificant percentage of the market.
And that seems slightly unlikely to me given the commercial Linux distributions sell solutions for the desktop as well.
Seems perfectly reasonable to me. The market as a whole *is* still growing. Same reason Apple’s sales are strong but their overall marketshare is, at best, static (and probably declining).
Also don’t forget you cannot say to every point “doesn’t influence the conclusion much” and then conclude it is accurate because all the points “do not influence the conclusion much”. They hav to be added up since this is a statistic; when calculating statistics it is bad practice to round up when you’re gonna use that number again. You don’t make an int or round up when you’re gonna use it again. If you want to calculate Z, and use A + B + C = D and E + F + G = H then use D + H = Z while D and H were rounded up/down you end up with a flawed end result!
The point isn’t that, individually, they “don’t influence the conclusion much” – the point is that there’s no reason to suggest any of them favour one platform over the other. There might be errors throughout the whole process, but they’re *indiscriminate* errors.
The flaws inherent in the way the zeitgeist collect data affect *every* platform, so if there is any misrepresentation, it’s working against *every* platform.
The zeitgeist shouldn’t be used to say whether or not a platform has 2% marketshare or 3%, it should be used to see whether a platform has 5% marketshare vs 30% marketshare, and to follow the trends – and none of the arguments against it have offered a decent reason why that is unreasonable. It may well be that Linux has 3% instead of 1%, but the significant part is that it’s still more than an order of magnitude less than Windows.
That’s what I’m trying to get across – none of the reasons given (save, perhaps, the browser ID issue) – have described anything that won’t affect *all* platforms and how they are reported, equally, to any statistical significance.
Imagine how many Microsoft Windows computers there are on this world. Divide it through 100. That’s still a damn lot right?
Yes, it is. The Computer Industry Almanac estimates at 800 million the number of active computers in the world. One percent of this is 8 million.
Since Linux’s market share is probably similar to Mac’s (around 2.5%), we could estimate the number of Linux computer at about 20 million. The Linux Counter estimates it at 18 million, which is in the same range.
That may be a small percentage, but that’s a lot of computers.
You’ll have to work pretty hard to convince me there are a greater proportion of Linux machines hiding behind NATs, or attached to broadband, than Windows machines at _all_, let alone enough to be statistically relevant.
It makes sense that the ratio of broadband/narrowband users is higher for Linux than Windows. First, as many anti-Linux advocates remarked in past OSNews comments section, Linux had for considerable time problems with winmodems (some would argue that winmodems are still not fully supported – I wouldn’t know, I’m a broadband user myself). This favors broadband.
Also, Linux users on the whole tend to be more computer-savvy, which means they are more likely to use their computers more and therefore prefer broadband over narrowband.
Finally, as I’ve mentioned earlier, Linux users download nearly all of their software – there’s very little boxed software products. Do you know how long it is to download a 100Meg package over narrowband? How about upgrading your distro – 3 CDs!!! It’s enough for any Linux head to fork out the cash and switch to broadband.
I’ve provided figures, which you can verify with a one minute search on Google. Moreovers, these stats make perfect sense. Perhaps you’d like to provide your own figures, or at least a counter-argument, instead of simply stating your doubts?
The number of Windows machines also behind NATs, or attached to broadband, should more than balance out the number of Linux machines in the same situation.
The “number” of Windows machines is irrelevant. We’re talking about proportions here. 35% of North Americans have broadband. 85% of Linux users do. In other words, the Linux users’ ratio is almost three times that of users at large, it makes sense that they would underepresented in a similar proportion.
There is no reason to believe that, but there is a problem with the “because”. If it ie. favors X over Y, then it matters with the additional question “how much”.
True, but the whole issue hinges on whether or not X is favoured over Y – and that’s the part none of the criticisms have justified.
My point is: Zeitgeist does not research that at all. Several effects on the statistics are ignored. Which does mean the result is more inaccurate than when that would have been taken into the analysis.
And my point is, for what reason should we think that these inaccuracies *only* affect (or don’t affect) a single platform. Why don’t they affect _all_ the platforms ? It’s the crux of the argument against the zeitgeist, and one which has not been justified.
What we (will) disagree on is the relevance of the inaccuracy while i’m pretty much open-minded to ie. offline computers, having several cool examples in my mind. Why the heck have Internet on your computer which is only serving you the ability to use a CAD program?
The disgreement seems to not be the relevance of the inaccuracy, but whether or not it is biased for or against a particular platform. I can see no reason, nor compelling argument to believe it should be.
This is actually an argument against yours
here’s why: if 99% of the desktop computers which are not hooked on the internet (an example part of my earlier argument) and are not using Linux -which is almost the same as “they apply to only a single platform”- then it does not affect the statistic. If more than 1% of the whole world does, it does have an effect on the statistic.
I’m sorry, but I don’t understand what you’re trying to say.
The question you should be asking is not “of all the unconnected machines in the world, if X% are running OS $A and y% are running OS $B, then…”. The question you should be asking is “if I take all the machines running OS $A in the world and all the machines running OS $B (where OS $A and OS $B are relevant to the discussion at hand), is there any reason to think the proportion of A that are unconnected/on broadband/behind a NAT/not using google/etc is significantly different to the proportion of B that are unconnected/on broadband/behind a NAT/not using google/etc ?”
I contend the answer to that, is “no”.
The zeitgeist isn’t there to say “this is a breakdown of all the computers in the world”. It’s there to say “of the computers that access google, this is what they look like”. The significant issue here is whether or not there’s any reason to believe the computers accessing google are, on the whole, significantly different to desktop computers in general.
Again, I contend the answer to that, is “no”. I’ve not seen any reasonable evidence or argument to suggest otherwise.
Here’s a question for you: how many Amiga’s and C64’s on this world, according to you, are running a version of their original OS AND have an internet connection?
And the answer is “it doesn’t matter because those platforms are irrelevant”. It’s like asking what proportion of global OS marketshare the software running my microwave has – it’s just not germane to the discussion. It would be sort of interesting, in a pure intellectual wankery kind of way, but it’s not really significant.
it should do so consistently – so using it to observe trends should be fairly valid (ie: Linux marketshare doesn’t appear to have gotten any bigger).
Well, if broadband users are underepresented, then the marketshare wouldn’t appear to significantly bigger until there was a marked increase.
Look at it this way. If the zeitgeist underestimates users by a factor of 2.5 (as I’ve argued it does, while you have simply stated that you didn’t believe it was the case without trying to adress my arguments), then a 1% market share would appear as 0.4% – not quite enough to appear as 1% on the zeitgeist. Linux wouldn’t appear as such.
With a real market share of 1.3%, Google gives 0.52% – enough to register as 1% on the zeitgeist. Let’s say it rises to 1.9% – almost 50% increase in the number of Linux computers. Google gives 0.76%, for a zeitgeist score of 1%. Now let’s say it has a real market share of 2.6% – twice the amount it had when it appeared on the zeitgeist. Google would still count it as no more than 1.04% – still 1%.
In fact, before Google would show it at 2% (which it briefly did, I believe I hear someone mention in this thread), it would require a market share of 3.8 percent, almost three times the amount necessary to appear as 1%.
Evoking “trends” when talking about a figure of 1% shows how little you know about statistics.
It makes sense that the ratio of broadband/narrowband users is higher for Linux than Windows. First, as many anti-Linux advocates remarked in past OSNews comments section, Linux had for considerable time problems with winmodems (some would argue that winmodems are still not fully supported – I wouldn’t know, I’m a broadband user myself). This favors broadband.
I agree. I’d expect Linux users more likely to have broadband purely because they’re more likely to have a technical bent.
That same argument is likely to apply to Windows 2000, Windows NT and “Other” users as well, so it’s just as likely they are similarly under-represented.
I’ve provided figures, which you can verify with a one minute search on Google. Moreovers, these stats make perfect sense. Perhaps you’d like to provide your own figures, or at least a counter-argument, instead of simply stating your doubts?
I’m not really arguing with your assertion more Linux users have broadband. I’m disgreeing with your conclusion that it has any remotely significant affect on any resultant conclusions.
The “number” of Windows machines is irrelevant. We’re talking about proportions here.
Sorry, I meant proportions.
35% of North Americans have broadband. 85% of Linux users do. In other words, the Linux users’ ratio is almost three times that of users at large, it makes sense that they would underepresented in a similar proportion.
So, absolute best case scenario is Linux is at 3% instead of 1% (more likely 1.5 = 2%). In a vacuum, a 3x difference is not insignificant. In the grand scheme of things, 3% is very marginally less irrelevant than 1%. Particularly when taking into account that the “3%” has apparently been stagnant for the last ~4 years.
The significant issue here is whether or not there’s any reason to believe the computers accessing google are, on the whole, significantly different to desktop computers in general.
I agree with this. To me, the proportion of Linux and Windows computers accessing Google is probably representative of desktop market shares. The problem is that Google counts IP addresses, not individual computers. Figures tend to indicate that Linux users are more likely to use broadband, and therefore will change IP addresses much less often than Windows users.
If you’ve got figures of your own, please give them to us. Preferably ones that can be verified quickly with a Google search…
Again, I contend the answer to that, is “no”. I’ve not seen any reasonable evidence or argument to suggest otherwise.
I’ve given you such arguments and evidence. If you don’t deem them “reasonable”, then it’s up to you to tell us why, and preferably provide your own. Otherwise you’re just wasting our time.
I agree. I’d expect Linux users more likely to have broadband purely because they’re more likely to have a technical bent.
That same argument is likely to apply to Windows 2000, Windows NT and “Other” users as well, so it’s just as likely they are similarly under-represented.
Uh, no. Windows is the OS of the masses. When you buy a PC at virtually any computer story, it now comes installed with Windows XP. There’s nothing indicating a particular technical bent in that. Therefore Windows users, representing the vast majority of users, is proportionately less technically-oriented than Linux users. Come on, you guys are the first one to whine that Linux is “for geeks only”, and now you’re trying to argue the contrary?
You can’t have it both ways. Which one is it?
I’m disgreeing with your conclusion that it has any remotely significant affect on any resultant conclusions.
…and not giving any counter-arguments in return.
Personally, I will argue that the ratio of broadband:narrowband users being nearly three times higher for Linux users than users at large will have a significant statistical impact.
In the grand scheme of things, 3% is very marginally less irrelevant than 1%. Particularly when taking into account that the “3%” has apparently been stagnant for the last ~4 years.
Well, at least you’re admitting that the figure is closer to 3% than 1%. I’ve already addressed why saying that the 3% has been stagnant is a statistical mistake in another post.
In the grand scheme of things, 3% is very marginally less irrelevant than 1%.
I’d just like to add that, in the real world, that 2% represents 16 million computers. That’s a lot of computers, any way you cut it…
I would agree with drsmithy and say that the ratio of Linux machines that are behind NAT or offline are not necessarily higher than Windows and OSX.
Also, not many desktop Linux machines are “offline”, even you agree mention that probably 75% or 85% of Linux desktops are on broadband.
Just becasue they have a semi static IP address does not mean they don’t get counted. How many dialup users constantly disconnect, quickly reconnect, and then search google again?
The browser ID thing is also partly flawed, I don’y have an exact string at the molment, but opera for instance can be changed to send something like
MSIE 5.5 blah blah on Linux
Although it ID’s itself as IE, it is also still different. I am sure google and other page statistics people know this too.
Linux users are more likely to use Google than windows users, as a matter of fact, I think every Linux user I ever met uses Google, I know many windows users that use Yahoo, MSN, and Lycose.
I think the latter point would almost suggest that the 1% statistic is inflated.
The point isn’t that, individually, they “don’t influence the conclusion much” – the point is that there’s no reason to suggest any of them favour one platform over the other. There might be errors throughout the whole process, but they’re *indiscriminate* errors.
The flaws inherent in the way the zeitgeist collect data affect *every* platform, so if there is any misrepresentation, it’s working against *every* platform.
First of all -which i said already actually*- is that i disagree that it matters wether an error is indiscriminate or not. An analysis can still be not accurate while it doesn’t discriminate, or doesn’t seem to. Second, i disagree they’re indiscriminate to platforms which weren’t made with networks/Internet/WWW in mind.
* My point has been for about 4 posts that the statistic is INACCURATE not that it is PREJUDICE.
“The flaws inherent in the way the zeitgeist collect data affect *every* platform, so if there is any misrepresentation, it’s working against *every* platform. ”
SIGH. My point was that they do not necessarily in the same way AS THE CURRENT PERCENTAGES. Just because it -according to you- doesn’t discriminate, does not mean that all those problems addressed would actually back up the current numbers. And, as i argued earlier, you have to calculate them all up not address them individually. The more i do that, the more i see the statistic _leaves_room_for_ inaccuracy.
You can’t simply say “doesn’t discriminate”, “doesn’t discriminate”, “doesn’t discriminate” x 10 times and then conclude at whole that because the statistic doesn’t discriminate, it is accurate. It says NOTHING about the usage which is not calculated. NOTHING. It says NOTHING about CAD workstations not hooked on the Internet being used in this world. NOTHING! And you assert the C64/Amiga is not relevant to the statistic; you do not know wether it is relevant to the market — for which this statistic is used (i know examples where Amiga and SGI computers are used for professional usage which are NOT hooked on the Internet (example of the “relevance” point i asserted earlier)).
“and none of the arguments against it have offered a decent reason why that is unreasonable.”
Strawmen.
“It may well be that Linux has 3% instead of 1%, but the significant part is that it’s still more than an order of magnitude less than Windows.”
3% is 300% of 1%. The significant part to me is the 300% rise which is relatively huge. This implies there’s a market for it while stable on 1% (minus overal growth) implies this much less.
“True, but the whole issue hinges on whether or not X is favoured over Y – and that’s the part none of the criticisms have justified.”
For you; not for me.
“[…]
I contend the answer to that, is “no”.
[…]”
I have already said i have an open mind on that, and i there’s various interesting, non-WWW things one is able to do with a computer with a non-Windows OS; and those are all NOT calculated. Wether we can assume the statistic discriminates or not doesn’t change that.
We disagree on that, and you assert you assume it doesn’t discriminate after which you conclude it’s change is not significant (see the 1 vs 3% for an example of why i think it is). We will not agree because you do not find the premise of _inaccuracy_ relevant while i do not agree because i do not find the premise of _indiscrimination_ relevant.
“Linux users are more likely to use Google than windows users, as a matter of fact, I think every Linux user I ever met uses Google, I know many windows users that use Yahoo, MSN, and Lycose.
I think the latter point would almost suggest that the 1% statistic is inflated.”
Good point. Another reason why a statistic like this is flawed (which i also already pointed out with “not everyone uses Google” on which you (drsmithy) will most likely say it isn’t discriminating. I cannot proof, and i cannot proof Linux users use Google more than Windows users do.
Such kind of uncertain points leave me to doubt the accuracy of this statistic.
(Oh, and as a matter of fact, i have similar points regarding Netcraft and Uptimes.)
PS: in your humble opinion, is a GSM a desktop? I’ve used one with Opera on SymbianOS to surf on the Internet. On Google also. GPRS, WiFi and the way i surfed on 9k6 (can’t remember how that is called but it was via dialup and normal GSM way, i think) all blur the statistic. Yet another example…
I would tend to agree with the author of this article except in one place. That is where he says that so called “web apps” are going to take over completely.
Web apps are fine for forums (like this one), on line shopping, web based e.mail and multi player on line gaming in the “slow” areas of gaming like cards, chess and monopoly. However I can think of several reasons that “web apps” SHOULD NOT be used for office and productivity apps.
Probably the worst problem with such software is not the software itself or the server carying the software or your own computer running the software. The problem is that little “on and off” crashy server that may lie BETWEEN the word processor, or spreadsheet program’s server and your computer, that NEITHER you or the web Office suite vendor have any control over, and DON’T Think it can’t happen. It is happening to me now with one of the political websites I visit. The crashy server between the web app’s home and your your computer would simply be DEVISTATING to computer dependent businesses if “web apps were to take over completely.
This leads to the second reason that I don’t think a “one size fits all” web app system is going to go over. CONTROL. I just can’t see people in mass trusting their personal and often copyrightable work to computers that they do not CONTROL!!! (This is particularly true amongst linux users most of which are no longer FOSS fanatics but look to linux as simply the best way of breaking Microsoft’s CONTROL over their computers.) I somehow doubt these people are going to be corralled into a system where their software is fed to them by a system like the Telephone company on a RENTAL basis. The Microsoft desktop monopoly was bad enough for them to make the move to Linux.
However I do agree that there is a strong future for mamory managed Software and the languages and development systems that produce it. (I’m migrating from C++ and PowerBASIC to Java and Python now myself though I still intend to stay with desktop software.) But that future is ON THE DESKTOP. Even Sun had to give up on the web app idea and re create Java with the ability to incorporate a main method for DESKTOP APPLICATIONS. (You remember that sun invented this web app idea with their “the network IS the computer” concept.)
Actually I think the best solution to software monopoly and developer boredom problems is NOT moving it from one monopoly platform (Microsoft Windows) to another (Some monopoly corporation’s server as a “web app”) but in using the interpreter/virtual machine with bytecode “binding” appriach with with multiple interpreters/virtual machines
to produce desktop EXECUTABLE FILES that really can be run on any OS or device. Don’t tell us desktop programmers that we have no innovation left, I may even get one of those infamous software patents on this idea of multiple interpreter/bytecode binding;c)
Hmm, i think i’ve found a way out of this discussion. Drsmithy what do you think is the % of _client_systems_ which have both a WWw browser which is used once in the year to access Google? Google has.. what.. 70% market share?
Figures tend to indicate that Linux users are more likely to use broadband, and therefore will change IP addresses much less often than Windows users.
Yes, but as a proportion of the *overall market*, those Linux users who are on broadband, and not changing IPs, are not significant. 1% or 3%, it’s still SFA.
As an aside, it’s entertaining that this argument – “the problem is that Google counts IP addresses, not individual computers” – never seems to be an issue when Netcraft’s webserver marketshare results are being quoted
.
If you’ve got figures of your own, please give them to us.
I haven’t made any arguments that need figures. I’ve been questioning your reasoning and conclusions, not your numbers.
I’ve given you such arguments and evidence. If you don’t deem them “reasonable”, then it’s up to you to tell us why, and preferably provide your own. Otherwise you’re just wasting our time.
You’ve given reasoning and figures that suggests the zeitgeist may be innacurate. I’m trying to say I don’t think the reasoning and figures – even if they are correct – make any significant difference to the overall conclusions.
Uh, no. Windows is the OS of the masses. When you buy a PC at virtually any computer story, it now comes installed with Windows XP. There’s nothing indicating a particular technical bent in that.
Windows NT and 2000 have never been sold as an “OS of the masses”. Machines running NT or 2000 are almost certainly either corporate desktops, or technically-inclined home users. Either way, the same logic regarding static IPs applies.
…and not giving any counter-arguments in return.
Ignoring them doesn’t mean they don’t exist.
Personally, I will argue that the ratio of broadband:narrowband users being nearly three times higher for Linux users than users at large will have a significant statistical impact.
And I say it won’t because at the end of the day it’s only going to mean the difference between ~1% and ~3%.
If I glance into a room full of 100 people and see someone who looks like Axl Rose, but then on closer inspection there are actually 3 people with long hair in tight leather pants drinking Jack Daniels from the bottle, it doesn’t mean I need to break out my Bon Jovi and Midnight Oil CDs.
Well, at least you’re admitting that the figure is closer to 3% than 1%.
No, I’m saying that even if it is closer to 3% than 1% it doesn’t make any realistic difference.
Added to that, in my *personal* opinion, based on evidence you wouldn’t consider valid, other aspects of OS usage patterns will make up for any under-representation Linux may have from fixed IPs.
I’ve already addressed why saying that the 3% has been stagnant is a statistical mistake in another post.
Which is still based on the same assumption that Linux machines are being significantly under represented due to static IPs.
I’d just like to add that, in the real world, that 2% represents 16 million computers. That’s a lot of computers, any way you cut it…
It certainly is, but it’s a lot less than 700 million. Or even the 64 million people who aren’t using Windows *or* Linux.
Just becasue they have a semi static IP address does not mean they don’t get counted. How many dialup users constantly disconnect, quickly reconnect, and then search google again?
I don’t know about you, but I use cable and my IP can last for months. Within a single month, my IP won’t change while that of a narrowband user is bound to change at least a dozen time.
The browser ID thing is also partly flawed, I don’y have an exact string at the molment, but opera for instance can be changed to send something like
MSIE 5.5 blah blah on Linux
Although it ID’s itself as IE, it is also still different. I am sure google and other page statistics people know this too.
Konqueror can masquerade the OS as well as the Browser. I don’t think Google goes to the trouble of doing advance filtering on their logs (which must be monstrous) in order to separate the “fake” browser/OS ID strings from the real one.
Linux users are more likely to use Google than windows users, as a matter of fact, I think every Linux user I ever met uses Google, I know many windows users that use Yahoo, MSN, and Lycose.
As I’ve said before, personal experience is not a valid measure of overall trends. Unless you’ve got some credible statistics to validate your assertion, then we cannot conclude that Linux users are more likely to use Google in a significant enough proportion to offset the underevaluation due to static vs. dynamic IP adresses.
As an aside, it’s entertaining that this argument – “the problem is that Google counts IP addresses, not individual computers” – never seems to be an issue when Netcraft’s webserver marketshare results are being quoted
.
That’s because Netcraft counts the number of servers, not actual machines. You can have multiple web servers on a single machine.
Also, web servers tend to change IP addresses at a much lower frequency than narrowband users.
I haven’t made any arguments that need figures. I’ve been questioning your reasoning and conclusions, not your numbers.
You’ve been questioning them without using counter-arguments. Or figures. In other words, all you’re saying is that you disagree. In other words, you’re wasting my time.
I’m trying to say I don’t think the reasoning and figures – even if they are correct – make any significant difference to the overall conclusions.
But not offering counter-arguments. Although you did admit the figure was more likely 3%, immediately trying to rationalize this by saying that 3% is still a small percentage. Well, it may be, but it’s three times bigger than 1%!
No, I’m saying that even if it is closer to 3% than 1% it doesn’t make any realistic difference.
16 million machines is a realistic difference.
Ignoring them doesn’t mean they don’t exist.
Well, if they exist, let’s hear ’em. Otherwise you’re just bluffing (badly).
Windows NT and 2000 have never been sold as an “OS of the masses”. Machines running NT or 2000 are almost certainly either corporate desktops, or technically-inclined home users. Either way, the same logic regarding static IPs applies.
Windows NT+2K represents, what, 25% of all Windows installations? That’s not enough to counter the underevaluation of Linux users. Not to mention that Windows 2000 was actually used by quite a few home users looking for stability after the WinME fiasco.
Added to that, in my *personal* opinion, based on evidence you wouldn’t consider valid, other aspects of OS usage patterns will make up for any under-representation Linux may have from fixed IPs.
“Yeah, yeah, I’ve got proof of what I say, but I won’t show it to you because you wouldn’t believe.”
Yeah, right. What a load of crap. You’ve wasted enough of my time – I forgot you’re one of those “runaround” troll who simply repeat their position without actually answering any of the arguments you present to them. Keep on believing that the zeitgeist is an accurate measure – that still won’t make it true!
Which is still based on the same assumption that Linux machines are being significantly under represented due to static IPs.
It’s more of a logical conclusion gained from published figures than an “assumption” – in any case, it’s better than your non-arguments and “secret” information!
I’m out of here.
First of all -which i said already actually*- is that i disagree that it matters wether an error is indiscriminate or not. An analysis can still be not accurate while it doesn’t discriminate, or doesn’t seem to. Second, i disagree they’re indiscriminate to platforms which weren’t made with networks/Internet/WWW in mind.
* My point has been for about 4 posts that the statistic is INACCURATE not that it is PREJUDICE.
Yes, and my point has been that it doesn’t really matter if the raw numbers are inaccurate, because they should be equally inaccurate for *every platform*.
Your position seems to be that the statistics may be innaccurate because, somewhere, somehow, in the data collection process, one platform is being favoured over the other.
My question is, *why* do you think that ?
It says NOTHING about the usage which is not calculated. NOTHING. It says NOTHING about CAD workstations not hooked on the Internet being used in this world. NOTHING!
Again, the important question (I think) is, why do you think the proportions of machines in these situations, per platform, are going to be significantly different ?
And you assert the C64/Amiga is not relevant to the statistic; you do not know wether it is relevant to the market — for which this statistic is used (i know examples where Amiga and SGI computers are used for professional usage which are NOT hooked on the Internet (example of the “relevance” point i asserted earlier)).
There are most certainly machines out these in those situations. I’m sure a lot of business are still running their accounting system on XTs running DOS 2.0, as well.
What I’m trying to say is that I think machines in these situations – that are relevant to the platforms being talked about – are simply so small in *overall percentage* that they can be safely ignored (or simply grouped in “other”).
3% is 300% of 1%. The significant part to me is the 300% rise which is relatively huge. This implies there’s a market for it while stable on 1% (minus overal growth) implies this much less.
But it hasn’t gone from 1% to 3%. It’s remained at 1% (or 3% if you prefer). That’s the thing – whether or not the zeitgeist is caclulating the wrong numbers isn’t important as long as its *consistently* doing so.
I agree that, on its own, a 300% increase is significant. But these things don’t operate in a vacuum. 300% of not very much is still not very much, overall. A marketshare increase of 1% to 3% is hardly sounding the death knell for Microsoft.
We disagree on that, and you assert you assume it doesn’t discriminate after which you conclude it’s change is not significant (see the 1 vs 3% for an example of why i think it is).
Not really. I agree that there are probably quite a few machines out there that never go near google. Where we differ, I think, is that I think the proportions of machines that google doesn’t record either apply to irrelevant platforms, or apply equally to all platforms.
We will not agree because you do not find the premise of _inaccuracy_ relevant while i do not agree because i do not find the premise of _indiscrimination_ relevant.
That probably sums it up reasonably well. But I still think I’m right
.
PS
Hmm, i think i’ve found a way out of this discussion. Drsmithy what do you think is the % of _client_systems_ which have both a WWw browser which is used once in the year to access Google? Google has.. what.. 70% market share?
I’m not quite sure why you think this would matter. My basic assumption basically says it wouldn’t matter, because there’s no reason to think any one platform would be disproportionately represented in this hypothetical subset of machines that only browse google once a year.
Anybody notice that it has WinXP at 50%. From articles I’ve read most businesses are on Win98 or Win2k and not WinXP. Even Microsoft is complaining that people aren’t upgrading to Windows XP.
I think it’s flawed in all directions. Also there is that 5% of other
.
I just don’t see people who start really using Linux, meaning they are on it for a few months, switching back. I think there are a lot of people who toy with it and then forget about it. It really depends, it’s certainly not everyone’s cup of tea.
I honestly believe that the biggest thing going for all users right now is choice. I work in a MS shop where in the past there was no choice in platform for os or development. Now I can run a dual boot machine and perform most of my work on linux (vnc helps to fill the few voids)
My point is that it’s great to finally have a choice! I don’t have to waste time patching & rebooting my machine with the latest hotfixes anymore (not to mention I can bypass corporate startup scripts constantly reconfiguring my system) I have a choice of a more secure system. I have a choice between photoshop & gimp, I have a choice between visualstudio & kdevelop.
I’m going to use the best tool for the job, independent of platform. Some things are easier for ME to do in windows and thats where I will do them-others are easier for ME in linux.
Whatver the linux usage is claimed to be, I find it hard to believe that fortune 500 companies are going to migrate any time soon. However, they are not saying no to linux-which is nice, because the choice is now up to me and not the CIO.
That’s because Netcraft counts the number of servers, not actual machines. You can have multiple web servers on a single machine.
The same principle applies, however. The webserver that tends to be popular for virtual hosting is over-represented.
You’ve been questioning them without using counter-arguments. Or figures. In other words, all you’re saying is that you disagree. In other words, you’re wasting my time.
Actually I’m saying why I disagree and why I think you’re wrong. However, you seem to be ignoring that.
But not offering counter-arguments. Although you did admit the figure was more likely 3%, immediately trying to rationalize this by saying that 3% is still a small percentage.
No, I made a comment saying that “if it were 3% […]”. I do not agree it is closer to 3% than 1%. I am, however, quite happy to consider an additional scenario with that as an assumption.
16 million machines is a realistic difference.
Not in the shadow of 700 million.
Well, if they exist, let’s hear ’em. Otherwise you’re just bluffing (badly).
I’ve stated my position and arguments numerous times. The crux of it, however, is that there’s no compelling reason to believe any platform is unfairly assessed.
Windows NT+2K represents, what, 25% of all Windows installations? That’s not enough to counter the underevaluation of Linux users.
Let me get this straight. Windows 2000 and Windows NT represent ~20% of the machines out there, but for some reason aren’t enough to “make up” for Linux, which at the most optimistic, accounts for ~4% of the machines ?
Not to mention that Windows 2000 was actually used by quite a few home users looking for stability after the WinME fiasco.
Not enough to matter.
“Yeah, yeah, I’ve got proof of what I say, but I won’t show it to you because you wouldn’t believe.”
Yeah, right. What a load of crap.
You wouldn’t believe it because you can’t find it on a google search, it’s based on my job experiences over the last ~5 years and it only deals with numbers of machines and users measured in tens of thousands.
You’ve wasted enough of my time – I forgot you’re one of those “runaround” troll who simply repeat their position without actually answering any of the arguments you present to them.
I’ve answered all your arguments, numerous times.
Keep on believing that the zeitgeist is an accurate measure – that still won’t make it true!
I don’t think its an accurate measure. I do, however, think it’s good enough to paint a broad overview of the industry. It certainly agrees with everyone environment I’ve ever worked in.
Web apps are fine for forums (like this one), on line shopping, web based e.mail and multi player on line gaming in the “slow” areas of gaming like cards, chess and monopoly.
I’ve got to disagree with some of this. I think web apps are *atrocious* for forms and email. Usenet and real email programs are far superior.
You can say the burden of proof the statistics are discriminating lies at me, but you know just as well that such is incredibly hard to proof. Futhermore i do not know wether and which extend they’re discriminating because of several details missing which is what i’ve asserted this whole thread but because i (and you, and many people) do not know this, i find these statistics not usable. There are some aspects in which i find it possibly discriminating which i’ll point out hereunder.
“because they should be equally inaccurate for *every platform*.”
If they want to correlate with current statistics they should however i am in no way of stating this in certainity.
“Your position seems to be that the statistics may be innaccurate because, somewhere, somehow, in the data collection process, one platform is being favoured over the other.”
Favored might imply it is done with the intention to do so; that is not what i’m implying. The option that -with or without intention (doesn’t matter to me) one is in advantage over another is indeed one i find possible and because of the lack of the details i think i -and you- are not able to finnish the whole puzzle. Because of the lack of several stacks which are missing (which are becoming more and more as i see it) i find the possibility the statistic is not accurate more and more likely.
“Again, the important question (I think) is, why do you think the proportions of machines in these situations, per platform, are going to be significantly different ?”
* Because of rounding errors. You even pointed this out with stated “other”.
* Because of several niche markets, usages are ignored in this statistic.
* Because of the several unknown aspects which all add up.
“But it hasn’t gone from 1% to 3%. It’s remained at 1% (or 3% if you prefer). That’s the thing”
I agree that that (minus the overal rise of the computers) is important to state Linux market growth on the desktop market is stagnant — when the statistic were accurate, that is.
“doesn’t record either apply to irrelevant platforms, or apply equally to all platforms.”
Irrelevant platforms? Irrelevent to what?
“I’m not quite sure why you think this would matter. My basic assumption basically says it wouldn’t matter, because there’s no reason to think any one platform would be disproportionately represented in this hypothetical subset of machines that only browse google once a year.”
Yes — an assumption. I do not want to assume such, because i’m not able in any way to know details on this. I’m able to state examples of relevant, niche platforms which generally are never using a WWW browser. Does that apply to your terms of discrimination?
I think your whole argument is based on the above assumption. We could go in depth, thinking of how many % of the world’s computers access Google every year (this leaves the IP offset aside) but in all honesty, i am afaik not able to know that and because of that wish not to base myself on a statistic which leaves aside several details of possible importance.
PS: Yeah, i know the way you’re using is used in many statistics. For example regarding people && voting they use only a small % of people from what they claim “all layers of the population”. I do not believe such is accurate, they do not influence my way of voting, and it tends to be always inaccurate regarding the smaller parties. Logic, because if you ask 500 people what they’re gonna vote and 10 say X then that’s a huge difference in contrast to 600 people which points out 20 say X. You have the advantage that the numbers are higher and a much higher % is asked however because of what i’ve earlier asserted i it questionable and therefore not accurate enough.
“Not to mention that Windows 2000 was actually used by quite a few home users looking for stability after the WinME fiasco.
Not enough to matter.”
drsmithy, you see? That’s the kind of thing i’m always worried about when it comes to statistics. “not enough to matter”, “irrelevant”. But that doesn’t apply on one individual inaccuracy. It applies on the whole. Only on the whole, if all the inaccuracies have pointed out are taken, one is able to state how the inaccuracies are not relevant enough to influence the statistic in such way that one would define it as “inaccurate”. And even then they should be pointed out instead of ignored imo because the relevancy one sees in it, could be very well different than the relevancy one other has.
Therefore i think it is a good thing to debunk possible flaws.
Here is also exactly where we differ, methinks. I find inaccuracies very fast relevant, and the danger with inaccurate statistics (a lot are) is when the inaccuracies are not pointed out; ignored in order to gain trust and credibility, sacrificing honesty and clearity. When the statistic is then used to draw other conclusions for example with a few other statistics we get this problem of so-called “inaccurate conclusions”.
How do you know that Google attempts to group unique IPs as one user? I think it’s equally likely that they just look at each request. It doesn’t say “people” accessing google in the stats, it just says “Operating systems accessing google.” So if person A using Linux accesses Google more than person B using Windows, I would suppose that Linux I would get extra counts. Do you have any proof that the counting is done in the highly illogical way that you describe? That way introduces complexity and problems, which you have elaborately over the last 80 posts described, and would all be solved by looking a the raw requests.
Also, Google, being filled with PhD level geniuses probably has decent solution that would hold up to some random OSNews whinger’s attempts to minimalize it.
Can we please stop arguing about platform browser statistics? This is a cool article that warrants discussion. Stop haggling over meaningless crap.
There’s also another item to consider. How many are dual-booters and how does that affect the statistics?
‘google’s zeitgeist says….’
Now repeat after me:
As Moses climbed up on the mountain the Google appeared and handed Moses the ZEITGEIST on tablets of clay.
And thus it was-
Linux would forever have a value of 1% on this tablet of clay.
and Windows would forevermore always unitl the end of eternity have %90 something percent.
The be all, end all, alpha and omega of FACTS-the deeply sublime cosmic metaphyiscal truth of numbers accompanied by percent signs has revealed the TRUTH in it’s awe inspiring and humbling presence.
The WORD has been spoken. Thus spoke Google’s Zeitgeist.
And the marketplace of Google’s Will defined the the world.
more seriously:
Google does not performs any kind of statistical research. They merely aggregate what is recorded by their servers. This has no real statistical value-scientifically it is meaningless. Without knowing all of the variables which influence these statistics we can deduce verly little about the viability of such numbers and any inferences to be made based thereon. The mere fact that statistical research is not being done- and Google does not pretend to be performing research-renders such discussions about the supposed ‘truth’ of such numbers or the supposed ‘reality’ which they represent moot.
“There are lies and there are damned lies, statistics are damned lies”, Samuel Clemens, a.k.a Mark Twain.
Unfortunately much of modern American discourse has been debased to the point that the citation of numbers is held to be “the presence of truth” legitimized by the assumption of empirical facticity. In most such “debates” the argument then revolves around whether the persona has accurately cited the numbers and whether their inferences are logical based on such numbers. At every step in such a “discussion” the truth-charachter, ie. it’s facticity, is presumed-this presumption itself is abosulte, even if this presumption itself cannot be verified independently-in fact it’s unverifiability is the key to it’s truth-charachter. If one calls this presumption into question one must offer an alternative statistic to procede in the argumentation. If there is no such alternative statistic-either nonexistant or not readily available at our disposal- the abesence thereof will be interpreted as the truth, ie. the presence of, the statistic being argued about.
Actually I’m saying why I disagree and why I think you’re wrong. However, you seem to be ignoring that.
You haven’t given good arguments as to why my arguments would be wrong, and no figures to contradict mine.
I’ve stated my position and arguments numerous times. The crux of it, however, is that there’s no compelling reason to believe any platform is unfairly assessed.
Your position, yes. Arguments, no.
Let me get this straight. Windows 2000 and Windows NT represent ~20% of the machines out there, but for some reason aren’t enough to “make up” for Linux, which at the most optimistic, accounts for ~4% of the machines ?
Right. You’re not very good at math, are you? What matters is not the absolute numbers, but the proportions. If 20% of Windows users would be underepresented, then it would balance out with 20% of the Linux underepresented (20% of 3%, if you will).
But anyway that doesn’t matter. You’re trying to confuse the issues (as I predicted you would). The important figure is that 35% of computer users have broadband: that means about 35% of Windows users, whether or not they use Win2K, Win98, etc.
Meanwhile, 85% of Linux users have broadband. That’s where the discrepancy lies, and that’s why Linux users are underepresented. When I mentioned the fact that Linux users were more computer-savvy, it wasn’t to support the fact that Linux users are underepresented, but rather to justify why the proportion of broadband users was higher.
But since you obviously don’t care about this argument, you’re predictably trying to distort it instead of answering it.
You wouldn’t believe it because you can’t find it on a google search, it’s based on my job experiences over the last ~5 years and it only deals with numbers of machines and users measured in tens of thousands.
Right. Sorry, but personal experience is worthless to figure out statistics. Anyway, since you’re obviously biased against Linux, this would unconsciously taint your estimates.
I’ve answered all your arguments, numerous times.
No, you haven’t. You’ve stated that you disagree. Different thing.
Anonymous (IP: —.client.comcast.net)
So if person A using Linux accesses Google more than person B using Windows, I would suppose that Linux I would get extra counts. Do you have any proof that the counting is done in the highly illogical way that you describe?
Counting is done by IP. Otherwise, it’d be easy to stack the numbers and write a script that would have a computer connect every second, all the time. Even if it was as you describe, then it would still not be an accurate reflection of market share, since it wouldn’t even try to count computers.
Also, Google, being filled with PhD level geniuses probably has decent solution that would hold up to some random OSNews whinger’s attempts to minimalize it.
Why, you sneaky little troll. You were just setting up a way to insult me, were you…ah, just seen your IP. Figures.
Agreed. Though I can guarantee you drsmithy won’t be able to refrain from posting one more message about it!
frankly, i love asp.net.. its a big improvement from classic asp.. im still stuck (i am in the philippines) using visual basic 6.. some are bothering to study .net..
Getting away from the pointless browser debate. Back to getting away from the Windows API as MS drop backward compatability. In the corporate area the new IBM Rich Client Platform based on the Eclipse IDE fits in with Joel’s general thesis. Its Jave based so it is both managed code and cross platform.
This platform provides native widgets via SWT it combines the advantages of a web portal with a rich client but it is not dependant on the Windows API. It provides a stable open source platform for developers to code for programs that will still work no matter what the changes in the Win API are – Winforms, Longhorn/Avalon etc. and that these programs will run natively cross platform on Windows, Linux and OSX I think that it stands a chance of gaining widespread use not just by IBM alone.
Getting away from the pointless browser debate. Back to getting away from the Windows API as MS drop backward compatability. In the corporate area the new IBM Rich Client Platform based on the Eclipse IDE fits in with Joel’s general thesis. It is Jave based so it is both managed code and cross platform.
This platform provides native widgets via SWT and it also combines the advantages of a web portal with a rich client but it is not dependant on the Windows API. It provides a stable open source platform for developers to code for programs that will still work no matter what the changes in the Win API are – Winforms, Longhorn/Avalon etc. and that these programs will run natively cross platform on Windows, Linux and OSX.
I think that it stands a chance of gaining widespread use not just by IBM alone.
we can talk about statistics for days -it’s very hard to get real stats for a freely downloadable OS with so many distros. Google zeitgeist is an interesting piece of data and let’s just leave it at that.
That same argument is likely to apply to Windows 2000, Windows NT and “Other” users as well, so it’s just as likely they are similarly under-represented.
Yep. I have 6 PCs running behind a router on a broadband connection.
5 of them run Windows (98,NT,2000,XP and another XP box) and one runs Linux (Debian)
So is google counting 1 windows box and 1 linux box from my place ? Cause if so its missing 4 windows machines.
//Why, you sneaky little troll. You were just setting up a way to insult me, were you…ah, just seen your IP. Figures.//
Ok, i just about laughed myself out of my chair when I read that.
Props.
🙂
I’ve given you such arguments and evidence. If you don’t deem them “reasonable”, then it’s up to you to tell us why, and preferably provide your own. Otherwise you’re just wasting our time.
I’m going to have to go with Google on these stats.
Googles stats indicate a general trend of computer usage.
I’m far more inclined to believe the numbers from a company with a reported 10,000+ servers that take more hits than almost any other internet site over a bunch of wannabe statistics experts on an forum.
Like the guy claiming “85% of linux users have broadband” – Where did he get this figure ? I’m sure he collects more meaningfull data sitting in his mother’s basement than Google does.
pfft
I just got a G4 Mac and installed OS X, even got MySQL running. Safari seems OK, but I couldn’t print a page text and graphics. Firefox handled that perfectly in Linux.
What I want to say I have Win98 dual booted with Slackware, I spend 98% in Slack and resort to Win98 for the Canon CP-200 to print digital pictures, because I have only drivers for Win98 and OSX (it also prints nicely in OS X). The pace of change seems to be even worse with MAC OSX I only just bought Appleworks and hear about iworks to replace Appleworks!
Is it not just a simple case that somethings are better done in each of the 3 O/S’s (Win/Linux/OSX). I am unable to switch TOTALLY to just one – although I am really happy with Slack, OSX is really tantalising……….
OSX by the way was a very easy install…….
OSX and Linux “feel” in normal use much better for daily use than Windows – especially when installing or uninstalling software. That’s my experience.
Just becasue your IP address only changes once a month, do you honestly believe that the IP address is only counted one time per month?
The free page statistics I used back in 2000 on homesteads free web space was far more intuitive than that.
Google’s logs are no doubt huge, but I doubt they purge 99% of them to filter out any matching IP addresses over a months time.
Like many page statistics I have seen, they are likely session based. That means that behing my static IP and NAT box, if I connect to google from Linux today and Windows tomorrow, both of them would be counted.
I would say that the higher percent of Google using Linux users would more than offset your arguement that “Linux desktops don’t change IP’s as often as Windows desktops”
Konq has the option to change the the browser ID, but from what I see Mozilla, Firefox, and Epiphany do not. My browser ID’s are:
Mozilla/5.0 (compatible; Konqueror/3.2; Linux) (KHTML, like Gecko)
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040207 Firefox/0.8
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040115 Epiphany/1.0.7
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040115
I have been trying to tie together all the loose ends of information about the direction Microsoft and programming are heading and this article was able to do that. The levels of complexity that people new to the tech culture have to wade through to be able to visualize the skills they will need for such a moving target is tough. It has all the makings of a strategy game, make the wrong move and you lose.
“If they can’t compete against MS, who depend upon a paid staff to produce software, then in what universe do you think they can compete in an OSS environment where opposing projects have unpaid/voluntary staff?”
1) money doesnt automatically create best software. if so whatever MS produces must be the best which is clearly not the case
2)OSS is not entirely done by voluntary people any more. novell,sun,ibm,redhat et all pay people to develop a good amount of OSS stuff
1. The reason I mention the paid staff of MS was not an attempt to say that paid programmers produce better code. It was to point out that MS actually has similar kinds of overheads (Eg Staff to pay) to its competitors. Overheads that don’t neccessarily exist when competing against project staffed by volunteers.
Even reading it again this position seems clear in my original comment.
2. Yes, they do. However can you reasonably expect to produce a media player for Linux and sell it (MPlayer)? How about a word processor (Open Office)? A graphics app (Gimp)?
The number of areas on Linux where you could hope to produce an app that isn’t going to be in competition with a free (Beer), community developed, OSS project isn’t that great.
So I stand by my original assertion. If you have insufficient brand recognition/originality/niche exploitation to make your app a success when competing against MS then your product will have no better chance of gaining a market when competing against OSS projects.
“Konq has the option to change the the browser ID, but from what I see Mozilla, Firefox, and Epiphany do not. My browser ID’s are: ”
every browser either supports it by default orthrough extensions
I’m going to have to go with Google on these stats.
Of course, since it confirms your preconceived notions.
Googles stats indicate a general trend of computer usage.
And yet they do not represent an accurate account of market share. Otherwise, companies wouldn’t be paying thousands of dollars to research firms to get accurate market share numbers.
I’m far more inclined to believe the numbers from a company with a reported 10,000+ servers that take more hits than almost any other internet site
No one’s questioning the numbers. It’s what those numbers mean that are in question. The methodology, if you will.
over a bunch of wannabe statistics experts on an forum.
Yes, another insult from a wintroll! I’m on a roll!
Meanwhile, I have yet to see anyone seriously challenging my arguments…
Like the guy claiming “85% of linux users have broadband” – Where did he get this figure ?
The “guy” runs the Linux Counter web site. The figure represents registered Linux users (more than 140,000), and so should give a good ballpark of the proportion of broadband Linux users. I’ve also given pretty good hypothesis as why Linux users would favor broadband. I notice you haven’t tried to address these as well.
My, the wintrolls are aggressive today. I guess I struck a nerve…
I don’t think the author took into account the coming migration to 64 bit computers.
Think about it, in the next three years, it’s not unreasonable to think that all new machines will be 64 bit, and due to people’s familiarity with Windows, they’re going to go right on buying Windows machines (that was the obvious part).
Now, if Microsoft only provides compilers that only use the .NET/WinFX APIs, would that not serve to help them establish those as the new standard for Windows programming, allowing them to retire Win32 pretty much altogether?
This of course doesn’t have anything to do with web services in general (I’ve got little to say on that topic).
Please don’t give me any guff about people who don’t want to upgrade. People have kids, and kids grow up, and quite a few of them buy computers, new ones, with Windows preinstalled.
”
Please don’t give me any guff about people who don’t want to upgrade. People have kids, and kids grow up, and quite a few of them buy computers, new ones, with Windows preinstalled.”
ignorance is bliss. joel is talking about developer pains not about users with kids. did you read the article btw?
Yes! Let’s talk about IBM’s strategy. The last thing IBM wants is to use rich client software on the desktop. Everything they do now has something to do with applications in the browser. They declared manny rich clients as legacy (think about Lotus Notes) and they are pushing their web stuff everywhere (think about IBM Workplace).
That Eclipse is an great application is out of question. And Eclipse is definatly not a IBM mass product, where they earn massive money with it.
“did you read the article btw?”
Yup. Sure did. But there was more than one issue in that article. I guess I just picked up on something that you conveniently missed!
“ignorance is bliss.”
“Yup. Sure did. But there was more than one issue in that article. I guess I just picked up on something that you conveniently missed! ”
no. you dont see the point about developers. he was not talking about end users at all.
“no. you dont see the point”
Sure I do. You don’t see mine. The fact that I mentioned end users at all is to say that they will be buying Windows computers, meaning that developers will still have more of an interest in developing for Windows rather than some other OS, and that the move to the 64 bit chips will give MS an excelant opportunity to try to make another API lock-in scenario, which has everything to do with developers.
You apparently have no ability to think things through. I repeat, “ignorance is bliss.”
“Sure I do. You don’t see mine. The fact that I mentioned end users at all is to say that they will be buying Windows computers, meaning that developers will still have more of an interest in developing for Windows rather than some other OS, and that the move to the 64 bit chips will give MS an excelant opportunity to try to make another API lock-in scenario, which has everything to do with developers.
”
64 bit is 32 bit compatible and MS cannot afford to drop compatibility. thats nonsense.
i am pretty sure joel is aware of that more than you. anyway i saw your point and i dont agree with your assumptions.