Valid XHTML 1.0!


(last update 15. Jul 2006 numbers up to April 2006 included and formerly rewritten tendency remarks)

Opera statistics
New Netscape 6–8 statistics
(Original) Mozilla statistics
(Mozilla) Firefox and -bird statistics

Special Remarks about this Page

Opposed to the general and OS specific statistics, where all browsers on one OS are compared, this one is about the few browsers available on the biggest, noticable desktop systems. Because out of my statistics those are M§ Windoze, the old Mac OS <= 9.22, the new UNIX based Mac OS X, GNU/Linux and Solaris, I request availability of these browsers on all of those five platforms. And there are only three browsers at now, which fulfill this condition: the new Netscape browsers since version 6, Mozilla 1.n and Opera (6, version 7 we are still waiting for on the Mac platforms, though a public preview is meanwhile available).

Remark: as of February 2004 I decided to add the (Mozilla) Firefox (formerly Firebird, originally Phoenix) browser to these portable browsers, because first it is rapidly approaching the 1% share overall and second because the not supported classic Mac OS is now rapidly dying, no more serving as exclusion reason for a so-called portable browser.

Current remark: obviously the Firefox browser has meanwhile begun to raid an attack against the illegal Redmond browser monopoly — stay tuned and watch it below and in total; it's usage is still exponential growing(!) as since the browser stepped in (then as Phoenix).

Of course the monopoly of a certain OS and certain browser has to be kept in mind, therefore I compare the shares of the browsers on those platforms relatively to all of the certain browser model over all platforms together. If there are differences, those shares will be different too among the different browsers. But changes may be due too changes in shares of operating systems, another note I have to make.

Due to the virtual impossibility to seperate the two Mac OSs correctly and completely I list both together, because the mentioned browsers are available on both, this doesn't affect the statistics generally (but masks possible differences between them). For the interested here is a little tarball with the used simple shell scripts, with which I determine(d) the numbers below.


time interval\OS shareM§ WinMac OS *GNU/LinuxSolarisFreeBSD, OpenBSDothers
200105–20011088 (7260)1 (124)9 (773)000 (6)
200111–20020492 (12205)1 (140)6 (900)00 (8)0 (9)
200205–20021090 (15018)1 (267)7 (1220)0 (2)0 (26)0 (9)
200211–20030491 (22281)1 (259)7 (1851)0 (14)0 (19)0 (8)
200305–20031093 (18863)0 (111)5 (1062)0 (2)0 (5)0 (27)
200311–20040493 (17957)0 (69)5 (1015)0 (18)0 (9)0 (31)
200405–20041094 (29291)0 (114)4 (1395)0 (50)0 (75)0 (57)
200411–20050494 (26509)0 (113)4 (1260)0 (32)0 (48)0 (34)
200505–20051092 (29650)1 (324)6 (1947)0 (96)0 (23)0 (71)
200511–20060490 (28763)1 (389)6 (2191)0 (123)0 (65)0 (310)

This table shows a slow, but steady rise in Opera usage (only partially caused by general traffic increase) on all platforms, but mere fluctuations in the shares regarding the different OSs. Therefore general remarks apply: the Windoze share of Opera browsers is roughly in agreement with its desktop monopoly, therefore this translates into a relative average Opera usage. Opposed to this, the GNU/Linux share is much higher than the Mac share, while the OS shares are probably in favor of the Mac platforms (but not for sure). And the GNU/Linux share is definitely lower, than the Opera usage share, while for the Mac OSs the opposite holds true. Altogether, the monopoly OS is Operas strongest platform (hardly avoidable), but GNU/Linux is by far the second most used, making it strategical for the small Norwegian company. The Mac share is (yet) rather low, leaving space for improvement; all others are neglectible (remark: the former BSD numbers are result of running the GNU/Linux version in compatibility mode, the native FreeBSD version became available as last OS in autumn 2002).

For the especially Opera interested I have to offer two scripts for GNU/Linux resp. Windoze, the only two major Opera hosting OSs so far. You can use them to determine the browsers major and minor version distribution there, if you like, from your (Apache) webserver logs.

Netscape since Version 6

time interval\OS shareMž Winall Mac OS *Mac OS X onlyGNU/LinuxSolarisothers
200105–20011093 (7061)5 (402)01 (76)00 (1)
200111–20020489 (11964)8 (1154)1 (198)2 (279)0 (12)0 (6)
200205–20021086 (14057)9 (1624)3 (532)3 (579)0 (53)0
200211–20030487 (22177)8 (2164)2 (566)2 (716)0 (189)0 (65)
200305–20031089 (24285)7 (1917)2 (730)2 (626)0 (161)0 (28)
200311–20040491 (27269)6 (1909)3 (1026)1 (395)0 (82)0 (1)
200405–20041091 (37738)5 (2448)3 (1309)1 (691)0 (201)0 (256)
200411–20050490 (29080)7 (2444)5 (1736)1 (491)0 (81)0 (7)
200505–20051091 (25627)7 (2090)5 (1413)<1 (239)0 (45)0 (22)
200511–20060488 (16531)9 (1707)7 (1407)1 (274)0 (97)0 (74)

This one shows a different picture: while the Windoze Netscape share is roughly in accordance with its desktop monopoly, this share fluctuates, despite total usage is rising slightly above the traffic. The increase in OS X share, part of the total Mac OS share, is mainly due to the increased share of OS X from all Mac OS systems, otherwise it seemed at least at times, that Mozilla is more popular than new Netscape on OS X, while the opposite holds true for its predecessor OS (compare below). On Solaris the old Netscape is more and more substituted by its successor, which makes the Solaris share visible here too… The apparent termination of Netscape development will decrease all new Netscape numbers in the future, by the way. And it is already losing share to Firefox it seems.


time interval\OS shareM§ Winall Mac OS *Mac OS X onlyGNU/LinuxSolarisFreeBSD, NetBSD, OpenBSDothers
200105–20011060 (853)4 (62)034 (494)00 (6)0
200111–20020441 (943)8 (193)2 (56)48 (1084)0 (2)0 (9)0 (22)
200205–20021049 (3980)11 (936)8 (718)37 (3031)0 (30)1 (100)0 (2)
200211–20030456 (10056)5 (891)4 (744)36 (6489)0 (104)0 (147)0 (115)
200305–20031062 (9349)3 (454)2 (352)32 (4785)0 (89)0 (89)1 (159)
200311–20040465 (12094)2 (529)2 (409)29 (5522)0 (110)0 (102)0 (53)
200405–20041073 (21823)3 (955)2 (755)22 (6621)0 (186)0 (79)0 (165)
200411–20050471 (17207)2 (691)2 (488)22 (5408)1 (429)0 (61)1 (371)
200505–20051057 (13929)3 (791)2 (506)25 (6189)6 (1670)0 (76)5 (1403) *
200511–20060450 (9914)3 (597)2 (438)22 (4366)12 (2499)2 (578)7 (1541) *

* mainly due to a tremendous number of Mozilla browsers (nearly only used alongside with some rather dated Netscape 4.x and earlier there) running on HP-UX, lured on my site by the BOINC client download offers — I'm sorry for this (though little) artefact deviation from a normal browser share; usually HP-UX is virtually not visible in my access numbers.

Another completely different situation: with fluctuating Mozilla shares of M§ Win around 40 to 60 % and of GNU/Linux around 30 to 50 % we have a sensational near-split between those two rival OSs Mozilla shares! This is clearly opposed to the reigning desktop OS monopoly… And the Mac OS, with dramatically rise of OS X share especially, has systematically increased in this browser share, despite recently it didn't kept up the pace of the Win and Linux explosions; generally speaking, all values are rising very fast, including especially FreeBSD and Solaris. Vice versa, while Mozilla is the by far most used browser at now on GNU/Linux (see the special Linux browser statistics), its share on M§ Win is relatively weak, unusually considerable weaker than the total Mozilla browser share summed up over all OSs.

Firefox/-bird (Phoenix)

time interval\OS shareM§ WinMac OS XGNU/LinuxSolarisFreeBSD, NetBSD, OpenBSDothers
200205–20021080 (48)020 (12)000
200211–20030471 (1639)0 (8)27 (637)00 (10)0
200305–20031081 (3391)0 (40)15 (656)0 (19)0 (28)0 (2)
200311–20040486 (11471)3 (466)9 (1264)0 (28)0 (36)0
200405–20041091 (56050)2 (1454)5 (3636)0 (11)0 (51)0 (32)
200411–20050491 (145492)2 (4542)5 (7957)0 (228)0 (159)0 (32)
200505–20051088 (222689)2 (6711)8 (21523)0 (556)0 (214)0 (111)
200511–20060486 (228234)3 (7914)9 (24215)0 (1415)0 (1048)0 (304)

This browser was never offered for the classical Mac OS, therefore only the Mac OS X numbers are listed of course (also for the Mac only 0.5 or at most 0.4 of the then Phoenix browser were available, at first only M§ Windoze and shortly after GNU/Linux were supported; but meanwhile even Solaris, FreeBSD and the no longer vendor supported OS/2 get regular, though inofficially contributed builds). It is just a slightly modified and vastly simplified, but close standalone browser relative of the original Mozilla suite, the above uses the same GUI XUL toolkit implementation since version 1.6 as this one always did, by the way. You can easily witness the breathtaking fast ascend of this browser model, which was the primary reason to add it here relatively early. Keep in mind, that the statistics begin later here than for the others, because earlier I didn't see even a single hit of the then Phoenix called browser.

The name confusion: the first versions from 0.1 to 0.5 were named Phoenix, then an IMHO not legitimate legal threat of the evil Phoenix BIOS company (which threatens freedom of computer customers by enforcing vendor control via big brother BIOS versions) made respective the then still somewhat independent developers of that project to change the name to (Mozilla) Firebird. That one stuck just from versions 0.6 to 0.7.1, when a name collision with a free database project made the developers of it unhappy with that choice, so changed again to the name (Mozilla) Firefox. Sadly enough this caused not only confusion among users, but gave and still gives serious webserver log analyzers a hard time, because there is no common key in these different named browser versions apart from the common Gecko one, which is shared by the Mozilla suite, the new Netscape and lots of other Gecko based browsers. I can assure you, that I have carefully collected those data appropriately though. Many thanks to the developers at for that also inconsistent user agent naming, by the way ;-) (the configuration subdirectory was always and is still named Phoenix, for example, when at last with Firefox going from 0.8 to 0.9 it was also renamed and rearranged into the mozilla subdirectory)

General Conclusions and Perspectives

While this overview is limited in the mentioned manner, it has several interesting tendencies, most already discussed above. Of these three portable browsers with a slight advantage Opera is more popular on Windoze than the Netscape browsers since version 6, and both are much more popular than the original Mozilla — maybe partly due to the commercial nature of the OS and the first two browsers, with Mozilla being the only free software of those three. On Mac OS * the Opera hasn't found broad acceptance so far, maybe the ill-fated version 5 and late coming out of the still immature (compared to the Windoze and GNU/Linux versions) versions 6 and 7 have to do something with it — the new Netscapes and Mozilla are nearly equally used. On GNU/Linux we have a rather weak new (commercial) Netscape, a strong Opera (also commercial!) and a ruling Mozilla (counting other free Gecko-engine based browsers like Galeon and Phoenix there makes this advantage even more impressive). The Firefox/OS situation is somewhat similar to the Mozilla one, but read below. Finally the situation on Solaris is Netscape dominated with recently fairly amounts of Mozilla usage too; the next half-year will show this more clearly, I guess.

Firefox and the Monopoly: What will happen?

Nobody can grant what will happen in future, and browser market shares are no different. But there is striking evidence, that the illegally gained (and not adequately juristical countered) monopoly of Redmond will topple in near(!) future though: the (Mozilla) Firefox browser is up to the challange, the numbers above, there and elsewhere are witness of this tendency.

Firefox has managed to increase its market share every year by an order of magnitude — that means, that it will be about as strong as the current monopoly browser in 2006 or so, well before the Redmond Empire can strike back with the user-controlling Vista/Longhorn and expected new integrated browser — and it is doubtful, that their panical announcement of a premature own version 7 browser can stop the tide against them.

But you have to be aware, that the Redmond mob will experience a stop in Frontpage sells (an utility, whichs primary goal is obviously to block web access for other browsers than their own), because nobody will still create pages, which can't be viewed by a considerable share of visitors, and they will lose any grip on web technique and development, because the Apache webserver prohibits the server way of domination. Therefore it is unlikely, that this will happen without hostile action from Redmond.

Luckily the far too easy way they got away nearly unpunished for their crimes can't be used, to block out Firefox in the manner, they used to do so with Opera: the new Netscape browser, also Gecko based, is still in use and the Gecko engine is nearly identical to that of Firefox. That means, that an attack on Firefox will break Netscape for sure too, which would be in clear violation of the judgement.

There is a certain danger for the Norwegian browser company Opera, since it went public: because they are not able (nor willing, because of their monopoly) to create a clearly improved browser fast enough, to catch up to the Gecko and Operas Presto engines, the Redmond mob will likely buy Opera in the known manner to annihilate a competitor and to (ab)use that ones technique for their own (monopoly) purposes. At now Presto is the only rendering engine up to Gecko, and this would be a typical way of getting back without doing anything by themselves (there was simply never any kind of innovation by them, always only buyed and/or theft ideas and techniques). But all of this will take time, and dropping other OS support than that for the monopoly one with Presto — which is to be expected — will not benefit them so much, as they think, because on the Mac and GNU/Linux platforms the competition will even gain from that and they are expected to gain in OS share at least together too. By the way, Operas browser dominance in the mobile device market makes it an even more attractive aim for the Redmond empire — in fact, this was the main motivation to block Opera from its M§N pages.

Either way we can now seriously hope for a better future of the web, when innovation obstacles like M§IE will lose influence and a more progressive and standard conformant kind of web can emerge.


browser statistics  browser resources  back to main

remarks etc. to: