MozillaZine

Google Zeitgeist Browser Stats Now Recognise 'Netscape 5.x'

Thursday February 20th, 2003

FriedEgg writes: "'Netscape 5.x' (Mozilla based browsers) are now listed separately in the Google Zeitgeist, and no longer grouped into 'Other'."

Targeted at members of the press and the curious, Google Zeitgeist offers regular statistics on the trends, patterns and surprises found by analysing Google search data. Following on from a discussion in our forums, MozillaZine reader jgraham has been in touch with Google, who say that 'Netscape 5.x' represents "all versions of Netscape 5.0 and newer and anything that uses Gecko."

Update! Zak adds: "Onestat.com's latest stats (released at beginning of Feb) show that for the first time, Mozilla has a higher share than NS4 (compare with the previous month's stats)."


#1 Safari?

by DavidGerard <fun@thingy.apana.org.au>

Thursday February 20th, 2003 12:22 PM

Reply to this message

Are they putting Safari in 'other' (which is getting quite sizable) or just picking the word 'Gecko'?

#2 Other

by tepples <tepples@spamcop.net>

Thursday February 20th, 2003 1:01 PM

Reply to this message

Safari and Konqueror use the KHTML engine and would probably fit within Other.

#4 Re: Moz can already active plugins without restart

by grahams

Thursday February 20th, 2003 1:09 PM

Reply to this message

But Safari has the word "Gecko" in it's useragent string so if a website is delivering speciality versions of the page it gets the one optimized for Gecko (since it will work better than the one formatted for IE).

#5 i'm sure google sniff's correctly.

by joschi

Thursday February 20th, 2003 1:39 PM

Reply to this message

its not that hard to do correct user agent sniffing if you keep up with the browsers, and i'm positive that google does.

#3 Finally eclipsed Netscape 4

by tepples <tepples@spamcop.net>

Thursday February 20th, 2003 1:06 PM

Reply to this message

It looks like the purple line (Gecko based browsers) has risen above the dark green line (Netscape 4.x) and the cyan line (IE 4) just before the first of 2003. This means web developers may soon not have to wrestle with hacks for compatibility with 4.x browsers anymore.

#6 Analysis

by jvm

Thursday February 20th, 2003 3:39 PM

Reply to this message

When the Jan 2003 Zeitgeist was posted, I did some analysis of it, and I had two conclusions that might interest folks.

1) Netscape 5 is counted twice on the Jan 2003 graph for all data points prior to the Jan 2003 data point. This can be verified by going back and analyzing the graphs from previous Zeitgeists. In addition to the data on the Jan 2003 graph being out of sync with previous Zeigeist graphs in this regard, the drop in the "Other" category from Dec 2002 to Jan 2003 is almost entirely accounted for by the percentage that Netscape 5 represents in the Jan 2003 data. (The only caveat here is if they were not including Netscape 5 in the "Other" category prior to Jan 2003 and they were not representing it on the graph (i.e. the percentages would not add up to 100% of the hits on Google). In that case, all previous data is suspect since Netscape 5 would not have been included.)

2) Taking the above adjustment into account (or analyzing the Dec 2003 Zeigeist by itself) the share of Google hits by IE (all versions) dropped for the first time since May 2002. The drop is about 4%, from 89% in Nov 2002 to 85% in Dec 2002. This does not necessarily mean that users switched to another browser; it could be that Google was getting hit by some web services that are not actualy human browsers. Without knowing what the jump in "Other" in Dec 2002 is specifically, we cannot explain it easily.

#7 Further analysis

by Kommet

Thursday February 20th, 2003 4:54 PM

Reply to this message

Based on a steady hand and some pixel counting, it looks like the end of the graph (mid January 2003?) shows the following values (give or take 0.6%): IE 6.0 - 49.9% IE 5.5 - 18.4% IE 5.0 - 17.8% Other - 8.9% "NS 5" - 2.9% NS 4.x - 1.8% IE 4.0 - 0.8%

This puts all version of IE at about 87.0%, compared to 88.1% a year ago (not much change), but more heartening is the fact that Netscape 4 and IE 4.0 have slipped from about an 11.7% combined share of the market to about 2.6% combined. Percentage share of browsers other than a version of IE or Netscape 4 ("Other" browsers plus Mozilla-ish browsers) has trippled in the last year, albeit mainly at the expense of Netscape 4.

Take my numbers with the grain of salt I used to measure them with...

#8 Netscape 4

by PaulB <pbergsag@home.com>

Thursday February 20th, 2003 6:37 PM

Reply to this message

My take on Netscape 4's decline in user share relates to Mac OS X. For years Nescape was used by Mac users for several reasons. They used Netscape 4x when IE failed them by not properly displaying pages (though Netscape 4.x often did not display the sites any better), because their ISP promoted Netscape 4.x (and not supporting Gecko based browsers) for the Mac and IE for the Windows users and some form of Anti-Microsoft reason. (My reson for not using IE on the Mac has been that it caused several hard to define instabilities with the operating system OS X, there in system 8 but not as noticable. Removing IE quickly improved OS stability.)

I will explain my point here. Recently OS X has reached a level of maturity and apps written natively for OS X users have been waiting for have finially been written for OS X. Users are migrating from System 9 to OS X now. Mozilla on OS 9 is very slow and users kept using either NS 4x or IE. Its been a while since I used the OS 9 Mozilla verson, but as I remember it is slow, sometimes even slower than NS 4. Thus on OS 9 most users continued with IE, mainly for speed unless they were a NS 4 user. With the migration to OS X the browser choice is much different. NS 4x does not have a native OS X version and going back to system 9 is fustrating just to use NS 4. IE on OS Xis a native app, but IE Mac does not equal IE Windows. IE is slow and has quirks with rendering, though I am not an expert since I have not used it all that much. My info come from users experiences posted at Mac web sites. Mozilla has been a viable option on OS X ever since version 1 for most users. This includes the CFM Mozilla as well as NS 7x. With the switch to the Mach-O Mozilla, Mozilla zooms on OS X and is a very good OS X browser. Thus as users adopt OS X as their primary OS on the Mac they are leaving behind NS 4x and in many cases IE and have been to Mozilla or Chimera. This may explain some of the increase usage of Gecko based browsers recently reported.

With the introduction recently of Safari by Apple we will have to wait and see if Mac users jump ship and move to Safari. My take is that some will prefer Safari over Chimera. Most users who prefer Mozilla over Chimera, as I do, will stay with Mozilla rather than move to Safari. Many users of OS X are still at 10.1.5 and not on 10.2 which is required to run Safari. It cost 129.US$ to upgrade to 10.2 so Geck based browsers may have a chance to catch on on the Mac. Will the Gecko based browsers develop a loyalty so that by the time most users have upgraded to 10.2 will want to stay with a Gecko browser rather than switching to Safari?

This is my take on some of the upsurge in usage of Gecko based browsers on the Mac. My take may be accurate or it may be way off base.:)

#10 Re: Netscape 4

by bzbarsky

Friday February 21st, 2003 3:24 AM

Reply to this message

> IE is slow and has quirks with rendering

Just so we're clear, most of IE/Mac's "quirks with rendering" are actually "correct rendering of CSS, unlike IE/Windows and NS4". It's just that Mozilla's quirks mode is more extensive that IE/Mac's..... (not that this has any bearing on the substance of your post, mind you ;) )

#12 Re: Re: Netscape 4

by TheK <kl@3dots.de>

Friday February 21st, 2003 7:04 AM

Reply to this message

problem is, that around 90% of the users out there call the W3C a Quairk of the MS-Standard :(

#15 Re: MacIE

by WillyWonka

Friday February 21st, 2003 8:20 AM

Reply to this message

Mac IE also ignores the standards in places to favour the IE5 way of doing things. I believe the box model is one and I've had problems with float: right; not working the way it was supposed to... but it matched IE 6 perfectly.

#23 Re: Re: MacIE

by MozSaidAloha

Friday February 21st, 2003 11:24 PM

Reply to this message

I found that although both MS browsers have the same name, they use different engines (MSHTML for Win32 MSIE and Tasman for Mac MSIE), I've found that they render differently and still suck...

#25 Re: Re: MacIE

by bzbarsky

Saturday February 22nd, 2003 6:21 PM

Reply to this message

No, MacIE has a correct CSS2 implementation of the box model, unlike IE5/Win.

#14 Analysis

by jvm

Friday February 21st, 2003 8:12 AM

Reply to this message

My numbers are as follows, for people to compare. Each is calculated from that month's results (i.e. Jan from the Jan graph, Dec from Dec graph, etc.). I do not know how to compute a margin of error for them.

Nov 2002 IE6.0: 46.0% IE5.5: 20.0% IE5.0: 20.0% IE4.0: 3.0% NS4.0: 3.0% Other: 8.0% Summary: IE accounts for 89%, everyone else is 11%

Dec 2002 IE6.0: 47.3% IE5.5: 16.9% IE5.0: 17.8% IE4.0: 3.0% NS4.0: 3.0% Other: 11.9% Summary: IE accounts for 85.07%, everyone else is 14.93%

Jan 2003 IE6.0: 48.4% IE5.5: 17.9% IE5.0: 17.9% IE4.0: 1.2% NS5.0: 3.7% NS4.0: 2.0% Other: 8.9% Summary: IE accounts for 85.36%, everyone else is 14.64%

For what it's worth, there was once a program I saw on freshmeat that could analyze graphs from the image files and compute data from them, but I can't find it. I think this would help get better data out of Google's graphs, so if someone can find it (I couldn't) let me know.

#16 Better data out of Google's graphs

by schapel

Friday February 21st, 2003 9:37 AM

Reply to this message

Google's data is not nearly as good as data that is collected from many different web sites. The larger and more random the sample, the better the data will be. I would recommend sticking with thecounter.com, onestat.com, and Chuck Upsdell's Browser stats for the highest quality data on overall browser trends.

#9 TheCounter.com

by thelem

Friday February 21st, 2003 2:04 AM

Reply to this message

TheCounter.com showed an impressive 16% growth rate for Jan03 compared to Dec02 (average is slightly below 10% growth), taking market share to 2.1%. Unfortunatly stats for Feb so far show a very slight drop in market share, down just over 1%. For some reason only about half the normal number of browsers have been recorded for feb so far.

#11 positive?

by TheK <kl@3dots.de>

Friday February 21st, 2003 7:00 AM

Reply to this message

is this a try to see everything positive? Fact is, that Netscape+Mozilla have _lost_ 0.2% als MS has won 0.3%. Also Gecko is only every 4th of the so called "alternative browsers"! As I see the google-Numbers are: 45% MSIE6, each 18% for MSIE5.0 and 5.5, around 10% for the "others", maybe 5% for Gecko, and each 2 for the "oldtimers". Also Mozilla does not even brake MS' in taking over the whole market, it only grabs users away from other alternatives as Opera or konqueror!

#13 Re: Positive!

by schapel

Friday February 21st, 2003 7:21 AM

Reply to this message

Gecko-based browsers are continuing to gain in popularity. They're now second only to IE. Additionally, non-IE standards-based browsers (Mozilla, Opera, Safari & Konqueror) are continuing to gain in popularity while old browsers like Netscape 4 and IE 4 are going away. Isn't that positive?

#17 Re: Re: Positive!

by TheK <kl@3dots.de>

Friday February 21st, 2003 10:08 AM

Reply to this message

nice, if there are only 2 candidates left, its not that difficult to be second. The ONLY positive point is, that the oldtimers are more of less gone - at least there.

The dissission users make, is still "MSIE 5.x or MSIE 6.0?", Gecko is far far far away from being visible. A positive message would be "Gecko has left one of the current MSIE versions behind" or "Gecko has 20%".

#19 Re: Re: Re: Positive!

by schapel

Friday February 21st, 2003 12:36 PM

Reply to this message

I have to completely disagree that Gecko has to reach 20% to be positive. Even if Gecko had only 10% of the market, how many web designers would make a page that doesn't render properly with Gecko-based browsers? Probably much fewer than today. Isn't that positive?

#20 Abandoning Netscape 4 hacks

by tepples <tepples@spamcop.net>

Friday February 21st, 2003 12:48 PM

Reply to this message

The important part is that the demise of Netscape 4.x frees web developers from the complexities of table layout and the contortions to workaround its bugs. Instead, they can put their effort into following standards more closely. So instead of designing for "IE and Netscape 4", they can design for "IE and standards" and stop DENYing standards-based user agents.

#21 What!!??

by PaulB <pbergsag@home.com>

Friday February 21st, 2003 3:03 PM

Reply to this message

"they can design for "IE and standards" and stop DENYing standards-based user agents."

Why wouldn't developers save themselves some work and simply develop one page which is standards based. Are web developers paid to wast their time with unnecessary extra work (pages specifically designed for IE). Developer are not worth what they are being paid if once the 4.x browsers are gone they are still writing two pages and not simply writing one stadards based page all browsers can access-->do employers enjoy paying for unneeded extra work?

#22 Designing for IE

by tepples <tepples@spamcop.net>

Friday February 21st, 2003 5:24 PM

Reply to this message

What about the fact that IE and Mozilla can't use the same plug-ins? Mozilla uses np*.dll and XUL jars, while IE uses ActiveX controls.

What about IE's inability to view MNG animations out of the box?

What about the errors in IE's interpretation of CSS, especially relating to whether or not the border and padding are included in the box dimensions? W3C says they MUST NOT but IE includes them anyway.

What about the fact that instead of interpreting XHTML sent as application/xhtml+xml as XHTML, IE 5 and 6 both display a source tree view?

#24 Re: Designing for IE

by asa <asa@mozilla.org>

Saturday February 22nd, 2003 11:11 AM

Reply to this message

"What about the fact that IE and Mozilla can't use the same plug-ins? Mozilla uses np*.dll and XUL jars, while IE uses ActiveX controls."

A site can be designed to use Flash, Real, Java, etc. and work fine in both IE and Gecko-based browsers. Besides, none of that falls under the "develop for the standards" suggestion that your post is in response to.

"What about IE's inability to view MNG animations out of the box?"

Nobody's perfect. Neither Gecko nor IE support every standard fully. Developing for the large and growing subset of the standards that both IE and Gecko supports seems like a prudent and time saving approach.

"What about the errors in IE's interpretation of CSS, especially relating to whether or not the border and padding are included in the box dimensions? W3C says they MUST NOT but IE includes them anyway."

Like I said, there's a large and growing subset of standards that modern browsers do support. Developing with that subset seems to be a good approach.

"What about the fact that instead of interpreting XHTML sent as application/xhtml+xml as XHTML, IE 5 and 6 both display a source tree view?"

Is that true even if you reference an appropriate style sheet with the content? I'm no web developer but it seems like there are pretty trivial ways to work around most browser quirks without forking the whole web page/site. I see your point, that developing for the standards might leave IE users with a less than ideal experience but there's certainly a useful subset of the agreed apon standards that web developers can use and successfully target Gecko, IE, Opera and others (excluding the generation 4 and older browsers).

--Asa

#26 Re: Re: Designing for IE

by bzbarsky

Saturday February 22nd, 2003 6:23 PM

Reply to this message

> Is that true even if you reference an appropriate style sheet with the content?

Yes. IE/Windows has absolutely and utterly no XHTML support. It only understands tag soup.

#27 Stylesheet?

by tepples <tepples@spamcop.net>

Sunday February 23rd, 2003 1:05 PM

Reply to this message

"Developing for the large and growing subset of the standards that both IE and Gecko supports seems like a prudent and time saving approach."

According to this page <http://www.w3.org/People/…xhtml/media-types/MSIE6.0> , XHTML doesn't seem to be one of them; look at the big red column in the table under "application/xhtml+xml".

It's true that the difference between IE and W3C box models can be worked around with a box-within-box hack, but what part of the subset of the standards that both IE and Gecko support covers patent-free lossless animations?

"Is that true even if you reference an appropriate style sheet with the content?"

What do you mean, exactly? Do you mean the same CSS stylesheet that I already use for HTML and which works with a <link /> element in XHTML in Mozilla but does not get read in IE, or something else that I don't know about? Would somebody please give me a link that explains step-by-step how to get XHTML working in IE?

"I'm no web developer but it seems like there are pretty trivial ways to work around most browser quirks without forking the whole web page/site."

I've seen many such workarounds on Netscape DevEdge, but I haven't seen 1. a good workaround for IE's lack of support for MNG animations, or 2. an appropriate stylesheet to use with IE to force it to render XHTML as XHTML rather than a source tree, along with instructions to get IE to use it.

#28 Re: Stylesheet?

by asa <asa@mozilla.org>

Monday February 24th, 2003 2:46 PM

Reply to this message

"I've seen many such workarounds on Netscape DevEdge, but I haven't seen 1. a good workaround for IE's lack of support for MNG animations, or 2. an appropriate stylesheet to use with IE to force it to render XHTML as XHTML rather than a source tree, along with instructions to get IE to use it."

So don't use MNGs or XHTML. Use HTML and gifs.

--Asa

#18 Re: positive?

by admbws

Friday February 21st, 2003 10:50 AM

Reply to this message

That's because Joe User uses IE, and will continue to do so as long as it's the most accessible browser availiable to him. Of course, the fact it's the most accessible browser availiable for Joe User has absolutely _nothing_ to do with the fact it's forced^Wbundled^Wan integral part of Windows, and it would be inaccurate and misleading to suggest otherwise, right guys?