Linked by Thom Holwerda on Sat 2nd May 2009 10:23 UTC, submitted by Luis
Linux Yes boys and girls, it's Net Applications time. Sure, their figures are flawed, and sure, they're misused all over the non-scientifically educated media, but that doesn't mean they do not indicate trends. One of those trends was a slowly rising popularity of Linux, which hit 0.93% market share in August 2008, only to sink back again during the following months. Well, it's April May 2009 now, and Linux has finally crossed the 1% market share line!
Thread beginning with comment 361414
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[2]: Not really
by silix on Sat 2nd May 2009 13:34 UTC in reply to "RE: Not really"
silix
Member since:
2006-03-01

Apps matter, not the OS.

Not quite. Without a good foundation it's tough (not impossible, but tough) to write good apps, and also the without a good OS the security of all your stuff is negligible. I'd say they both matter. Without good OS it doesn't matter what you run in it, and without apps even the Holy Grail of OS-development is useless.

not quite.
on one hand, for the vast majority of people, the OS is just a mean to run applications that allow them to leverage the workflow that suits them best

without good applications no END USER will be interested in the OS, for as good as it may be -
except maybe as a technology showcase or a toy, or for those who don't care about industry - recognized, professional grade applications, and /or take pleasure in contemplating the exquisite technicalities of operating systems, or fiddling with alternative operating systems for the sake of it...

OTOH, application development is a classic example of a chicken and egg problem: a platform can hardly grow without catering to the widest possible audience (ie without a supply of good and diverse applications)
but development of new applications would require a large and vital user base for the target platform, in order to be a worthy investment, and not a risk...
or, ISVs could take the initial risk, if given enough incentive - in the form of good libraries, development tools and so on, or maybe assistance in writing / porting their app, or actual funding, or even backward and forward compatibility in the platform (an often overlooked key factor in preserving one's investment) ...

catering to third party developers in every possible way can often contribute to the overall success of the platform, for as poorly engineered, implemented and performing it may be, more than allegedly superior kernel implementation features - as it removes one part of the above riddle

providing *good tools*, a "good enough" but *consistent* platform and the assurance the platform worked well for microsoft in the DOS and win 3.1x times, maybe FOSS could learn something...

Edited 2009-05-02 13:38 UTC

Reply Parent Score: 1

RE[3]: Not really
by vivainio on Sat 2nd May 2009 14:04 in reply to "RE[2]: Not really"
vivainio Member since:
2008-12-26


OTOH, application development is a classic example of a chicken and egg problem: a platform can hardly grow without catering to the widest possible audience (ie without a supply of good and diverse applications)
but development of new applications would require a large and vital user base for the target platform, in order to be a worthy investment, and not a risk...


This chicken and egg problem has been surmounted already.

We have Qt under LGPL. ISV can choose win32 as their main sales platform "right now", and provide a Linux version as an almost free bonus. Admittedly, we've had that option for a long time with Gtk+/wx, but they really couldn't cut the cross platform thing as well as Qt (and the pricey commercial license was reasonable option *for some*, but out of reach for many).

If you are a company that is paying for custom software, you'd do well to require that it be cross-platform (to avoid excluding part of your desktops, now or in the future).

Note that LGPL'ing Qt only happenend in 2009.

KDE4 started becoming realistic option in 2009.

We are seeing lots of stuff happening right now. It may be "cheap shot" to ridicule the market share, but those with eyes and ears for what's happening in the industry can't escape sensing the momentum that is piling up.

Reply Parent Score: 4

RE[4]: Not really
by silix on Sat 2nd May 2009 16:55 in reply to "RE[3]: Not really"
silix Member since:
2006-03-01

This chicken and egg problem has been surmounted already.

on windows, yes, it has been solved a long time ago, but on linux, i suspect it is still there ...
otherwise, we'd have plenty of those user-required, industry-recognized applications i mentioned earlier (plus games) on linux, too

We have Qt under LGPL.

for C++ gui development, QT is a wonder, but there's much more to developing and deploying a third party application, than the gui framework, especially on linux
when the few promiment closed source applications that desktop linux users actually install (iirc Opera uses Qt, btw) , are individually packaged not just for every major distribution, but in some cases for *releases* of distributions (to select from long lists), then it means those distributions and sub-versions account for different deployment targets in their own right, with slight differences and incompatibilities - rightly so, since for all intents and purposes of SW development and testing, every variation in the system image (runtime configuration, kernel version, libraries present and their versions, etc) essentially creates a new case to be individually supported (especially inpractical for commercial SW that strives to address the widest audience but cannot afford the user tampering with the program's source or makefiles and recompiling, to make it work on the unsupported distribution of his choice...)
thus the collective "linux" market share becomes less relevant, and ubuntu's / opensuse / fedora /... individual marketshare is what counts

Edited 2009-05-02 17:05 UTC

Reply Parent Score: 1