Talk:Usage share of web browsers/Archive 1

Archive 1

Archive 2

Archive 3

→

Archive 5

Usefulness for web developers (version segregation)

If Possible, could we break "Internet Explorer" into Internet Explorer 6 and Internet Explorer 7? The two are very different from one another; and I think that this page would be much more useful if the data showed these trends. —Preceding unsigned comment added by 207.67.48.2 (talk) 19:57, 3 March 2008 (UTC)

The information about the current usage share of browsers by version is in each browser's article, at least for the more popular browsers. The Internet Explorer and Firefox articles both contain this information, for example. Historical information is also available by following the source links in this article. I don't think there's any need to break down browsers by version in this article's tables, as the information is readily available already and useful for only a minority of readers (web developers). -- Schapel (talk) 23:29, 7 March 2008 (UTC)

I don't think it's useful just for web developers. Besides, much of the readers of such a page would actually be web developers. An argument like "it's in each browser's individual article" doesn't count since one of the purposes of such statistics articles is to present pieces of information, that are commonly looked up together, together. Checking multiple articles for statistics that should be evaluated/compared together is tiresome. Or perhaps you propose an article "Usage share of web browsers for web developers". Well, it could be organized as a section of the article i guess.

But it's mainly IE that matters. Browsers like Opera & Firefox are usually updated regularly by the users (update notification is on by default) and users that took the effort to download a browser usually keep it up-to-date. Version information is, however, much more useful for IE since it comes preinstalled on so many machines and much of the users don't bother (or don't know that it's important) to update it i.e. version segragation in IE is more significant. I find the information as presented in [1] very useful.

IE6 vs IE7 is of interest to everyone as a partial proxy for WinVista uptake. —Preceding unsigned comment added by 75.36.144.70 (talk) 08:26, 4 October 2008 (UTC)

If you want to see Vista uptake, you can go to NetApplications stats and select the OS version. -- Schapel (talk) 12:04, 4 October 2008 (UTC)

I'm going to add my voice to the chorus that adding information about IE6 usage is very useful to have in this article; User:Schapel appears to be the only editor opposed to adding this data. I don't think a single outspoken editor should stop us from getting consensus; I will edit the article to add this information in a few days unless other people speak up or Schapel's gives us a reasonable explanation why adding this information to the article harms it. Samboy (talk) 01:43, 26 February 2009 (UTC)

I agree with Schapel's opinion - it is not symmetrical to have IE share broken by versions but not the others. I suggest, as a compromise, put something along these lines: For market share information of IE by version see Internet Explorer#Market adoption and usage share section. Wikiolap (talk) 05:57, 26 February 2009 (UTC)

I have added a link in light of the consensus above. If anyone has any suggestions for making this link interfere less with the flow of the article while remaining just as visible for people just glancing at the page, let me know. IE6 market share is an important statistic. Samboy (talk) 17:02, 10 March 2009 (UTC)

I have to disagree. I don't feel the symmetry of the page, and the data presented, should be given equal weight. The fact is that IE6 and IE7 are vastly different in both ability and rendering. Their initial releases were 5 years apart, and IE6 still maintains a market share far above almost any browser other than Firefox (barely depending on what particular statistics are cited) and IE7 itself. To display multiple versions of, say, Safari, or Opera, makes little sense because not only are the differences between versions relatively small by comparison, but their usage share is small as well. With IE this is a completely different story. Not only do both IE6 and 7 hold a tremendous amount of the market, but their capabilities and rendering are tremendously different. I myself, as a web developer, came here looking for the ratio of adoption between Safari, IE 6, IE7, and Firefox, and was very surprised to find that the distinction between versions of IE not readily available.--Cintax (talk) 09:13, 24 March 2009 (UTC)

The information is readily available. Go to many of the sources of the data in this article, and you can find information on the usage of the versions of all the browsers. Within a minute I could find several sources with this information. You can also easily find the information on current usage of IE versions in the Internet Explorer article. It took longer for you to complain that the information isn't in this article than it took me to find the information. -- Schapel (talk) 11:34, 24 March 2009 (UTC)

By your logic this page should barely exist, as all of the information it contains can be found in other sources and other wiki articles. The point, however, is that this page collects that information to a single point. The point is not that I "could not find it" but rather that it has a place here, in this article. I fail to see why some of the editors are being so obstinate in spite of the fact that reasonable arguments for the inclusion of this information have been made, and no solid arguments against it have been made. To argue that the information is available elsewhere is to argue that wikipedia itself serves no purpose, as almost all of the information is contains can be found elsewhere. --Cintax (talk) 08:36, 25 March 2009 (UTC)

This article takes the data available and simplifies it to an easily digestible form, just like the rest of Wikipedia. Just like the rest of Wikipedia, if you want more in-depth information, you are free to go to the original sources, and other external links. Actually, the single point where information on the current use of each separate version of the popular browsers is in the articles on the browsers. That information is already collected to a single point, in Internet Explorer and Mozilla Firefox. This article, on the other hand, collects the usage information about web browsers, without regard for version. You seem intent on simply complaining and arguing, rather than simply easily finding the information you desire. -- Schapel (talk) 13:24, 25 March 2009 (UTC)

You seem to misunderstand. If I wanted the information myself I could've gathered it and thrown it in an Excel sheet long ago. This isn't about my personal quest for the info, it's about whether or not it has a place in this particular article. This is the discussion section is it not? Frankly, it's rather conceited of you to say that I'm simply complaining when I'm simply trying to state my case for why I feel this information should be present here. The fact remains that by ignoring the distinction between IE6 and IE7, the information presented in this article is substantially less useful. --Cintax (talk) 08:19, 26 March 2009 (UTC)

I have to agree. I do not understand your point at all. Could you explain how this article would be substantially more useful with extra columns for IE6 and IE7 (and, presumably, IE8, and IE5, and IE4, and perhaps even Firefox 3, Firefox 2, etc.)? It seems to me that the effect would be mostly to add more columns to tables that are already unwieldy. It would also be more difficult to see whether the usage share of IE was rising or falling at any particular time from looking at the tables (because one would have to mentally add to get the total for all versions of IE), something that is immediately obvious now. What is important to web developers is which versions of IE are popular now, and this information is already in the Internet Explorer article. Are you saying it would be important for web developers to know which versions of IE were popular one year ago, two years ago, etc.? If so, why can't they simply go to the original data to get this information? -- Schapel (talk) 13:04, 26 March 2009 (UTC)

New idea: why not showing the table as it is and a uncollapsed edition of the same table with the splitted information by the different version? that serves both: normal user, webdevelopers and others! @schapel this article is for this purpose! nobody should go to any different article to get the information he should find here. that is as you have to go to the Comparison of layout engines (graphics) to get the info which browser support which file format (which should ALSO be found at comparison of web browsers) --> this is the ame "problem" data should be on both articles! web develops need the usage of ie6/7/8 splitted, because they totally rendering the pages different. in this versions were made such big steps that you can't compare to - say - ff2 & ff3 or safari 1 and safari 3... *edit link correction* mabdul 0=* 13:51, 26 March 2009 (UTC)

First, it isn't what this article is for. This article is for the usage share of web browsers. It is not for the usage share of different versions of browsers. Second, what, specifically, is the problem with going to another article for the information? What is the problem of going to the Internet Explorer article for information on the usage share of different versions of Internet Explorer? Third, what's important to web developers is the current usage of different versions of browsers. If you're going to insist on putting information on the different versions of Internet Explorer in this article, why not do it in the graph at the top of the page, instead of adding many columns to many tables, or adding a whole bunch of new tables? That would be a far less intrusive change, and I think would serve the purpose better. -- Schapel (talk) 14:42, 26 March 2009 (UTC)

This article is usage share of web browsers, meant as a first stop for the general public. Not a narrowly defined special interest article like Usage share of web browsers AND different versions of IE are grouped together. As a general rule, article should be adjusted for notability. That means we have an article that highlights statistically significant shares, while cutting minority browser stats (apparently the current cutting point is Opera). Clearly, IE6 has a large share. As a browser with profound influence on the Internet, its statistics should not be needlessly removed. —Tokek (talk) 11:18, 19 April 2009 (UTC)

It doesn't make sense to group all IE browsers together. There is a huge huge difference between IE6 and all other browsers on the chart. A chart that showed IE6 vs. all other browsers would make sense. This article is about browsers, not organizations. If the focus was on the organizations and their competing browsers, grouping all of IE6 might make sense. However, this page is about *browsers* and while IE6 and IE7 are made by the same company and related to eachother, they are very very different browsers. Safari, WebKit and Chrome all have more in common than IE6 and IE7 do. What is the reason these three aren't grouped? Again, it seems that the rational for the browser grouping is based upon the organizations that created them, not the browsers, which doesn't make sense at all. I agree with what Cintax said above: it makes sense to group versions of Firefox, Safari and Opera because the difference between the versions is small. Plus, they're all part of the "not-IE6" category. IE6 is the only "non-modern" browser that is still being used by a large number of people, and it doesn't make sense not to separate it out. Lets choose a method of grouping other than the organization that created the software, please, that doesn't make sense. Brentonboy (talk) 18:39, 30 April 2009 (UTC)

Here is how I think about it - all the web sites that we reference here either track browsers "by organizations" (in your terms), i.e. they group together all versions of IE, all versions of Firefox, all versions of Opera etc, or if they go by versions - they break every browser by version. So why would we do it differently from them ? They are the experts in browser tracking (this is why we cite them), we are not experts (no OR in wikipedia). Wikiolap (talk) 03:29, 1 May 2009 (UTC)

So we can track the historical usage of different browsers over time. For example, a reader can see at a glance when the usage of IE was growing and when it was falling. If IE were split into different columns, a reader would have to mentally calculate. This would make the tables harder to read. We've been over this all before. I asked some specific questions, and never got any answers. -- Schapel (talk) 11:32, 1 May 2009 (UTC)

The websites we reference have done the research to collect the data. Presenting that data in a way that is appropriate to an online encyclopedia is hardly original research. It's the same data, and by presenting it in a different format isn't by any stretch of the imagination original research. Otherwise all of Wikipedia would be original research unless composed of nothing but direct quotes. Most of the other arguments here made sense, but that is so ridiculous I am beginning to suspect that you aren't sincere in this discussion and have some ulterior motive.

Wikipedia should give useful and accurate information based on reliable sources. The chart as it is now is only useful for people who want to follow the success and failure of Microsoft, Mozilla Inc., Opera Inc, etc. Which is fine, but limited. I think that has it's place on Wikipedia, but most people are going to be more interested in seeing the really old and outdated browser versions separated out. The point is that a per-version distinction would be more useful to a larger group of people, so why is it excluded from Wikipedia? There are ways to do it without interfering with the current data. I don't think that adding a similarly colored sub-pie slice to the chart for IE would be that distracting or confusing. Nor do I think it would be original research. Brentonboy (talk) 21:23, 7 May 2009 (UTC)

TheCounter.com data

January 2005 data was not included due to irregularities in reporting (Mozilla usage in Jan 2005 was not 1%). I also went to a lot of trouble to get accurate data rather than the rounded-to-one-percent that is easy to read off the table. I'm going to revert the changes unless the bogus data is replaced with good, accurate data. -- Schapel

That's why I didn't include TheCounter.com previously. It is just too troublesome. Anyway, I changed the interval to annual since the change rate was just too slow. --minghong 1 July 2005 13:43 (UTC)

Could we settle on some way of handling the TheCounter data without you completely wiping out all of my work and presenting bogus data (for example, Firefox usage in 2005 being just 1%)? -- Schapel 03:24, 15 August 2005 (UTC)

Hmm, this is a very old discussion, but I also noticed quite a bit of bogus data on that site. Especially when looking at the OS usage of last month: only 2 people with Vista? Come on, there's no way that's possible. Perhaps there is another - more accurate site - available to get this information from? Glodenox (talk) 18:02, 6 May 2008 (UTC)

What about stats from w3schools.com (http://www.w3schools.com/browsers/browsers_stats.asp)? They are very different from your stats here. Opera does not have 0.71 but 2.3%, Firefox 44.2 instead of 20.78! —Preceding unsigned comment added by 91.11.227.82 (talk) 21:43, 1 January 2009 (UTC)

See the information about W3Schools stats in the article and on this talk page. -- Schapel (talk) 01:41, 2 January 2009 (UTC)

NetApplications.com data

The data listed in the NetApplications.com data table is not what NetApplications.com is reporting as overall browser usage. Those data are for one site, designergolfgifts.com, only. For example, compare May and June 2005 numbers listed in the table to the numbers given in this recent PC World article. We should take out the numbers currently in the table and replace them with official NetApplications.com stats -- Schapel 08:23, 15 July 2005 (UTC)

Thanks for the correction. ;-) --minghong 00:22, 18 July 2005 (UTC)

The NetApplications table is rapidly becoming the largest table in the article, even though it covers one of the smallest time periods. Additionally, because the data from each month is shown, it can be hard to see the overall trend of the data in the noise of the monthly variations. I propose that before the March 2006 data is added to the table, we switch to using NetApplications' quarterly data. -- Schapel 15:01, 24 February 2006 (UTC)

NetApplications.com is the most relied on source, and the source for the chart, So I moved it to the top of the section -Misty Willows (talk) 16:03, 2 October 2008 (UTC)

Huh, what makes NetApplications.com "the most relied on source?" Because they are at the top of the page in wikipedia? To me this is just circular reasoning. "We do so just because everyone else does so" —Preceding unsigned comment added by 85.225.219.105 (talk) 22:27, 4 December 2008 (UTC)

See also the discussion in Talk:Usage_share_of_desktop_operating_systems. Especially this link [2] which proves that the Net Applications data is extremely biased towards USA and UK. If there are no better alternatives, there should at least be a warning on top of each companies table, that describes what to expect from the data. —Preceding unsigned comment added by 85.225.219.105 (talk) 07:29, 5 December 2008 (UTC)

To offer a sample interpretation of the link in the paragraph above, if you calculate the percentage of the top american ISPs it adds up to around 40%, while the first german ISP, Deutsche Telekom, is around 0,8%.

Germany has more than a fourth of USA:s population (82 million compared to 300 millon), has a similar technical level but is in the stats accounted for 1/50th of the internet traffic.

Of course nothing is known of how Net Applications weigh their data, but I allow myself to speculate, this bias could for example give higher points to usage of Apple products who often focus their launches to USA. —Preceding unsigned comment added by 85.225.219.105 (talk) 11:45, 5 December 2008 (UTC)

Janco Associates

I've really begun to wonder about the data from Janco. They show usage share of browsers jumping way up, then way down again, without apparent pattern. Examples are when Firefox was shown at over 12% in 2005 then mysteriously jumped back down to below 9% three months later, and even more strangely, Netscape dropping down to 0.15% and bouncing back to over 2% three months later. Their January 2006 press release includes numbers for previously reported time periods that don't match the previously reported numbers, so we'll end up with multiple conflicting reports as we already have for April 2005. I know browser stats are fairly unreliable, and I can understand a small amount of "jitter" in reported numbers, but these mysterious and huge leaps are just too much for me to take. Is it just me, or does Janco's data just not seem credible any more to others as well? Should we just remove all their data from the page? -- Schapel 03:37, 23 January 2006 (UTC)

Because no one has responded, and because having Janco's data in the article seems to lessen the overall quality, I'll just go ahead and rip it out. -- Schapel 14:48, 24 February 2006 (UTC)

Web browser usage share

I recently moved this page to Web browser usage share, but it was moved back again. Why? Andy Mabbett 00:12, 23 November 2005 (UTC)

Because we have stuffs like list of web browsers and comparison of web browsers. So just to make the name consistent with each other. --minghong 01:50, 23 November 2005 (UTC)

Wouldn't it have been more logical to have changed those other articles to Web browser list and Web browser comparison? Then the specific topic ("Web browser") would be in the most signifcant position in the names, as opposed to having general categories ("list" and "comparison") in the most significant position. That might help with searches. Chris Loosley 23:20, 2 January 2007 (UTC)

I just checked on the w3c site for browser statistics and it displayed numbers completely different from the pie-charts given in the article (w3c is given as a source on the article. (http://www.w3schools.com/browsers/browsers_stats.asp). The pie-chart is extremely misleading. 24.30.15.20 (talk) 17:10, 12 July 2008 (UTC)

No, our chart uses data from W3 Counter, not W3 Schools. We do not use W3 Schools because its numbers are based only on visitors to its own site (and one with a distinctly unrepresentative audience), not a range of sites. Actually, I have doubts about W3 Counter's sample, and whether we should be using it for a chart. (I wonder whether it was chosen because it shows the lowest IE numbers.) But that's another question. --Groggy Dice T | C 19:54, 12 July 2008 (UTC)

The relevance of search engines and cellphones

I miss information about how much of the traffic is from search engines and how large a proportion of visits come indirectly through search engines. The impact of search engines is steadily increasing. I guess a very larg emajority of new visits tio sites comes through search engines like Google, Yahoo, Altavista, MSN, Clusty, Inktomi etc. Search engines probably do not search inside Macromedia presentations, Active-X and such. Many web designers might not be aware of loosing this opportunity to be found by prospective visitors. I think not only the browsers and rendering count but the access of search engines to the information counts as well.

Other interersting points would be to look specifically at "market share" of browsers on devices other than personal computers, such as mobile phones and PDA's. Opera is quite strong in that segment. - Mokgand 11:06, 7 December 2005 (UTC)

Merger of Usage share of web browsers and Usage share

"Usage share" could apply to many things other than web browsers, yet the the page talks only of browsers. I suggest merging into this article - Chris Wood 12:43, 27 January 2006 (UTC)

Or we could generalize the Usage share article so that it applies to search engines and other things usage share could apply to, and then the articles should probably be kept separate. -- Schapel

I agree. We should merge the content on Usage share in and replace that page with a stub for the time being? - Chris Wood 14:10, 31 January 2006 (UTC)

Thank you, Tom2006, for doing the effective merge and redirecting Usage share. - Chris Wood 20:22, 20 February 2006 (UTC)

Table widths

I reverted the table width changes, so that nearly no table rows need to word wrap on window widths down to 900 pixels in IE and Firefox. Without the widths, many table rows needlessly word wrap, and tables are harder to read because the rows are of differing heights. If someone can change the tables so they look significantly better with narrow windows (around 800 pixels) and not significantly worse with more typical width windows (around 900-1000 pixels), go ahead and do so. Note that about 80% users have 1024x768 resolution monitors or better. [3] -- Schapel 14:04, 24 February 2006 (UTC)

Note that screen resolution is misleading — many people with large screen resolutions don't maximise windows, particularly as long lines of text can reduce readability. I have a 1680 by 1050 screen resolution, but only run my browser at 1024 by 768, and I'm certainly not unique in this respect. --Safalra 14:19, 9 June 2007 (UTC)

Right. Still, the majority of users have their browser windows at 900-1000 pixels. The tables should not be made to look better at widths of 800 pixels at the expense of being significantly worse at wider widths. Can you see a way to make them look much better at a width of 800 pixels and not much worse at wider widths? -- Schapel 15:37, 9 June 2007 (UTC)

Opera as IE

"Browsers such as Opera are set by default to report in as Internet Explorer, presumably for compatibility reasons." I thought they had changed this as of version 8. And does anyone know for sure if it was for compatability reasons? -DevastatorIIC 04:31, 21 March 2006 (UTC)

No, Opera 8 defaults to cloaking as IE. Opera 9 currently defaults to identifying itself as Opera, but the final version might go back to cloaking as IE by default. According to the Opera Press FAQ, Opera cloaks as IE by default for compatibility reasons. I'll take the presumably out.

Layout engine / web browser usage share

"A rough estimation of usage share of layout engines/web browsers." Page nowhere states how this image has been done, and on what data it is based. Page states just that "this article aims to be an unbiased historical record for the usage share of web browsers, based on statistics and articles published by well-known websites. One of the uses of such statistics is to create a graph that roughly represents the browser wars", but gives no indication how this image has been made. —The preceding unsigned comment was added by 62.78.155.119 (talk) 06:33, 13 February 2007 (UTC).

It's based on data from this page. You can find more information if you click on the image itself. -- Schapel 15:39, 13 February 2007 (UTC)

In General

User:Digita has added a table to this section. It appears to me to be completely pointless. I've reverted it twice, but he's put it back each time. Am I missing something or is this table just useless clutter? -- Harumphy 17:23, 7 March 2007 (UTC)

I think it's worse than useless; it's downright confusing. I'll remove it. Anyone who thinks it's helpful at all, please discuss first before re-adding it. -- Schapel 18:10, 7 March 2007 (UTC)

I further updated the table, maybe this will be an improvement. However, even the initial version was hardly 'pointless' - it directly relates to the paragraph next to it. Digita 18:24, 7 March 2007 (UTC)

I should make it clear that I did not revert it 'twice', the second time I added new data to the table to make it more relevant as well as adding wikicode reference footer. Now, I have added all-new 2nd table that you may find more relevant (now added to the page). Finally, please do not delete the reference within the paragraph, and reference footer. Digita 18:43, 7 March 2007 (UTC)

I agree that these tables should be removed. They provide data from single source, whereas most of the article shows data from variety of sources. Making one of those sources more important then others seems contradictory to the spirit of the article and can also be considered POV. Wikiolap 19:29, 7 March 2007 (UTC)

The point was not to provide an overall comparison, but to provide an example of a comparison as talked about in the paragraph. Overall I disagree, but Wikio has a valid point that it seems to place one source of higher importance then the others. On that basis I will remove (both) the tables, and I thank Wikio for actually giving a logical reason for his opinion. Digita

I think I finally understand what you were trying to do with the table. Until you described it, and I thought about it, it made no sense at all. The reference still doesn't really make sense, because the example describes the current 83% usage share for Internet Explorer, but the reference gives an two-year-old example of 83% usage share for Internet Explorer version 6. Because of that difference, it's still confusing. Could we change the reference to a current one that shows all versions of Internet Explorer having 83% usage share, so the example and reference would match? -- Schapel 20:14, 7 March 2007 (UTC)

I added the 83% reference for 2005, simply because thats when it was 83% (rather then updating pre-existing text). However, this was an obtuse way of adding the data and I agree that section needs improved. Feel free to update that section as needed, having newer data and better references would certainly make more sense. Digita 20:25, 7 March 2007 (UTC)

I found two references from 2006 that say IE (all versions) has 83% share. We really don't need them, as it's only an example. We could just as well give the example of Sneeflob browser having 42.356% share (just as an example), or Firefox having 24% share. However, I agree that having an example that verifiably reflects the current situation (or at least a fairly recent situation) is best. It just wasn't clear at all from your table that that was your intention. -- Schapel 23:19, 7 March 2007 (UTC)

Yes, I see what you mean. Im glad we have worked this out and I am fine with the page. Digita 23:29, 7 March 2007 (UTC)

Is the omission of Konqueror intentional?

The usage of Konqueror is quite low compared to the other browsers mentioned. Its usage is so low that most stats sources don't even mention it. I find it only natural that this article would not mention it either. -- Schapel 01:45, 14 April 2007 (UTC)

Summary

I think there should be a brief summary of the top three in the opening, something to the effect (using invented stats) "Currently, approximately 50% of users are using ie6, 30% use Firefox and 10% use ie7, the remaining 10%, and caveats, are detailed below." Andy Mabbett 11:35, 5 April 2007 (UTC)

I would rather not have that. This is a rathole which will open door for POVs from all sides. Wikiolap 14:42, 5 April 2007 (UTC)

There are methods for dealing with PoV pushing - I came to the page seeking exactly that kind of information and I'm sure others do, too. Andy Mabbett 15:02, 5 April 2007 (UTC)

If you want to do that, I would suggest attributing a source (which is one way of handling PoV pushing), and using Check Upsdell's browser stats (listed in the External Link section). He does a very good job of summarizing the current overall usage of the main browsers. -- Schapel 15:11, 5 April 2007 (UTC)

Sorry, but choosing one source out of the dozens to use in summary is POV. The way the page it is now - it simply shows what different sources report - this is neutral and unbiased. Wikiolap 15:33, 5 April 2007 (UTC)

But it isn't one of the current sources. It is a completely separate website that uses different sources and provides Chuck Upsdell's summary of those sources. Attributing that summary to the author would not violate POV in any way that I can think of. I agree using one of the current sources in the article and putting that up as a summary would be POV, as it would be giving undue weight to that one source over the others. That is exactly why I am proposing doing something different from that. -- Schapel 16:43, 5 April 2007 (UTC)

(outdent) How about something like:

Estimates vary greatly, but, as of March 2007, about 20% of users have ie7, 50% ie6 and 20% Gecko-based browsers such as Mozilla Firefox. The remainder, and the reasons for treating these figures with caution, are discussed below.

citing Upsdell, whose figures I've averaged and rounded? Andy Mabbett 18:07, 5 April 2007 (UTC)

Upsdell is just another source (and another point of view therefore). Are his averages weighted ? There are big differences in the volumes in the sources he averages over. Some of them overlap with the sources in the article. Why Upsdell's numbers are so special we would want to use them in the summary ? Do they have more fidelity ? I don't think so. It's OK to add his stats as another section to the article, but not to the summary. Wikiolap 18:18, 5 April 2007 (UTC)

OK, how would you summarise the current percentages, in a sentence or two, at the start of the article? Andy Mabbett 18:25, 5 April 2007 (UTC)

IMO, any attempt to summarise the current percentages will constitute OR (since it will be non-trivial synthesis). This is why I vote against having summary section - it will either be POV or OR. Wikiolap 18:41, 5 April 2007 (UTC)

I know you don't think we should do it; you've already said so. I'm asking you if we do do it, then how would you like it done? Andy Mabbett 13:49, 6 April 2007 (UTC)

There are three sources of recent global figures, so you could say something like: "Globally, IE has about an 83% share, Mozilla and derivatives 13%, Safari 3% and Opera 0.6% (average of most recent data from TheCounter, Onestat, and NetApplications)." In view of the accuracy of the data, it would be misleading to use additional significant figures. Harumphy 12:44, 6 April 2007 (UTC)

I'd be happy with that. Thank you. Andy Mabbett 13:49, 6 April 2007 (UTC)

I'm happy with that too, as long as everyone else agrees. -- Schapel 14:42, 6 April 2007 (UTC)

I am not happy, but if everybody else thinks this should be done, than this approach is probably the best. On related note, there is a graph at the top of the article, and I wasn't able to locate what sources did the author of the graph used for his numbers. If we will go with the Summary section, would it be a good idea to update the graph accordingly ? Wikiolap 15:19, 6 April 2007 (UTC)

I have already updated several of the usage share graphs using averages from several different global usage share sources as described. I'll also update the one on this page and the one on the browser wars article in the same way. -- Schapel 23:20, 6 April 2007 (UTC)

(Outdent)

So, where are we with this? Andy Mabbett 20:55, 13 April 2007 (UTC)

I think we have consensus on the User:Harumphy's proposal:

There are three sources of recent global figures, so you could say something like: "Globally, IE has about an 83% share, Mozilla and derivatives 13%, Safari 3% and Opera 0.6% (average of most recent data from TheCounter, Onestat, and NetApplications)." In view of the accuracy of the data, it would be misleading to use additional significant figures.

Wikiolap 00:23, 14 April 2007 (UTC)

I think we have a consensus on how to do it, but not necessarily on whether to do it. --Harumphy 09:21, 14 April 2007 (UTC)

Proposed merge from Market adoption of Mozilla Firefox

The Market adoption of Mozilla Firefox page is a strange beast. It is unique, not only in the field of web browsers (neither IE, nor Opera, nor Safari have such pages) but in the whole of Wikipedia -- there are no other pages of the type "Market adoption of X". Of course, uniqueness is no bad thing if there is a good reason for it; but as far as I can see, there isn't here: there is nothing on that page that would not be better encapsulated on a more general page about the comparative usage share of web browsers -- i.e. this page. On the contrary; keeping a seperate page merely contributes to the unnecessary splintering of information (see WP:MM#Overlap). Any comments for or against a merger? -- Simxp 21:24, 28 May 2007 (UTC)

You should discuss the merger in the Firefox article, as the Market adoption of Mozilla Firefox is a daughter article of that one. It's not really a "strange beast" at all; it's a daughter article like thousands of others in Wikipedia. -- Schapel 00:42, 29 May 2007 (UTC)

I've removed the tag; no support to merge has emerged, and Schapel has a point about the daughter article; though I still retain my scepticism about the justification for the existence of such a page in the absence of similar pages for any other web browser, software, or indeed anything at all. -- simxp (talk) 06:17, 15 June 2007 (UTC)

Scientific support

Isn't the whole article misleading, since there is no scientific study for market share? There are statistics, sure, but there's no scientific study to support the accuracy of those statistics. Everyone assumes these statistics have validity, but there's absolutely no basis for that, other than it's a generally-agreed-upon assumption that these are good ballparks. I just think the article language should say that a statistic of 68% usage share for IE does not necessarily mean that 68% of people are using IE. It's just the percentage that the statistics support, based on little more than a user agent string. I'm not for changing any of the numbers, I'm just for educating people that a statistic of 68% could theoretically be 58%. Or heck, 48%. No published study in a scientific journal has been conducted on the people sitting using the browser, so until that happens, the numbers should be treated with suspicion.66.119.170.242 (talk) 22:51, 6 January 2009 (UTC)

Isn't there already a large section at the top of the article that goes into excruciating detail about how and why particular browsers might be overcounted or undercounted? -- Schapel (talk) 23:12, 6 January 2009 (UTC)

Yes, but my point is that, in particular, these sentences are not correct: "Usage share, in web browser statistics, is the percentage of visitors to a group of web sites that use a particular browser. For example, when it is said that Internet Explorer has 68% usage share, it means Internet Explorer is used by 68% of visitors that visit a given set of sites." That's not true. The data does not, in fact, mean that Internet Explorer is used by 68% of visitors that visit a given set of sites. An estimated 68% of visitors are using Internet Explorer, the key word being estimated. That's one part that bugs me. It's a contradiction to the section following. 66.119.170.242 (talk) 02:53, 13 January 2009 (UTC)

No, there's no contradiction. The first section defines what the usage share for a set of sites is. The second section explains why what is measured as the usage share might differ from the actual usage share. If 68% of visitors to a site use IE, then the usage share for IE is 68% for that site, by definition. On the other hand, for various reasons, the usage share might be measured as 67% or 69%. There is no contradiction between the different numbers; in fact the section explains in detail why the measured usage share may differ from the actual usage share, which explains the discrepancy. -- Schapel (talk) 11:51, 13 January 2009 (UTC)

W3 Counter

I've added another source of stats today. I'm a bit concerned that this appears to sample once every 10 days. This is perhaps too frequent for us and it might make sense, once we've got a few months' data, to take monthly averages. --Harumphy 18:23, 4 June 2007 (UTC)

Reverse chronological order

Ever since there was a change to put the more recent data on top, I wonder whether it makes sense to reverse the tables to show more recent data on top. This is also consistant with how W3Schools web site shows their browser statistics. What do people think about it ? Wikiolap 15:57, 28 July 2007 (UTC)

My two cents I suppose I would prefer your suggestion, but I don't think the current format is deficient. Does anyone know of similar articles that do use one or the other format? -Justin (koavf)·T·C·M 23:28, 28 July 2007 (UTC)

This has been done. Why not? It looks better, and current information is what we are after. 203.109.189.179 05:57, 29 September 2007 (UTC)

Thanks a lot for doing it ! Wikiolap 15:44, 29 September 2007 (UTC)

W3Schools

Why is W3Schools' data now on this page, listed as "Global Usage Share Data"? It's not, it's usage share data from their site only, which they admit is highly skewed. Let's just remove it before even more people think it's actually representative of global usage share. -- Schapel 13:56, 1 October 2007 (UTC)

I listed it under "Global Usage Share Data", because I understood "Global" as "Worldwide". I will comment it out for now, but let's see how (or whether) other people reply here. Perhaps there will be enough arguments to restore it. Wikiolap 17:27, 1 October 2007 (UTC)

The "global" used in the other data sources are global in the sense that they are a composite of data drawn from thousands of servers worldwide. The W3Schools data is from only one web server, one that is not representative of web usage because mostly web developers and other web enthusiasts visit that site. That's why the numbers for Internet Explorer are so low and those of other browsers are so high, as most people visiting the site are aware of alternatives to Internet Explorer. It's just that reason that makes the data inappropriate for use in an article purporting to list overall usage share. I explain that site at the bottom of the article simply because so many people are mistaken that the W3Schools data represent global usage share. -- Schapel 16:44, 2 October 2007 (UTC)

I'm thinking keep, but explain that the audience of W3schools is likely to be more tech-savy than other markets. Since the data is used for market analysis, presumably some website admins would like to be able to look this up and justify extending support for firefox and other browsers which have a higher market share according to the w3schools data --91.84.69.180 (talk) 17:16, 16 March 2008 (UTC)

Sure If someone wants to put an explanatory note, go ahead. It's pointless to have it in a comment indefinitely, though. -Justin (koavf)❤T☮C☺M☯ 06:13, 9 June 2008 (UTC)

I still think the W3Schools data is misleading. The "local" designation still does not explain that the data is for only one site, and that the site is heavily biased towards "alternative" browsers. Besides, why include that specific site's stats and not other sites' stats? Doesn't that just open the door to including any and all sites' stats in this article, making it hard to see the real overall browser usage stats? I'm for including only stats sources that give a representative sample of the usage of various browsers, and not including any individual sites' stats as they are likely to be biased samples. -- Schapel (talk) 17:12, 11 June 2008 (UTC)

Local data It is worth keeping only to the extent that they have a lot of traffic. If they don't, then I'm fine with deleting it. The thing that I don't want is to keep the information, outdated, in a comment in the article. That is useless. —Justin (koavf)❤T☮C☺M☯ 17:17, 11 June 2008 (UTC)

W3Schools doesn't seem like a major site. It's a site about web browsers, and it also shares its web browser stats, so many people mistakenly think those stats are for the entire web. It's just a misconception. I'll just delete it. -- Schapel (talk) 17:26, 11 June 2008 (UTC)

Misidentification of uncommon browsers

There is something important about web browser statistics not mentioned. Some less common web browsers might be incorrectly identified. For example, I tried using Mozilla Shiira on a blog. When I checked the sitemeter, it identified as Safari 1.2. Another time, I tried with a nightly build of Mozilla Firefox 3 beta 2, and it registered as Mozilla 2. This info maybe should be included in the article, that data collectors might not be able to correctly identify a large number of browsers, instead giving credit instead to IE, Firefox, Safari, whatever. Althepal (talk) 23:49, 16 December 2007 (UTC)

Netscape vs FireFox in TheCounter.com section

In the theCounter.com section, the table of statistics contains the heading Mozilla/Netscape, which I have changed to Mozilla/FireFox. This was reverted with the reason "keep this table consistent with others". This reason makes no sense, as all of the other tables list the usage stats for FireFox under either "Mozilla" (OneStat.com), "Mozilla FireFox" (ADTECH, Net Applications, W3 Counter, WebSideStory) or "FireFox" (OneStat.com country specific stats). Some of these (but not all), include a separate "Netscape" column to list statistics of the Mozilla based Netscape browsers. The statistics of theCounter.com themselves list most of this percentage as FireFox. HermanHiddema (talk) 14:22, 18 April 2008 (UTC)

I would also like to know the relationship between the users Schapel and Mozillar, as the revert history for this article is suspect and smells of sockpuppetry or similar HermanHiddema (talk) 14:22, 18 April 2008 (UTC)

I don't understand the reasoning. The TheCounter table lists Firefox under Mozilla, as do many other tables. Netscape is included because some tables break it out into a separate column. Therefore, Mozilla/Netscape indicates that Firefox, Mozilla Application Suite, and Netscape are included. Leaving out Netscape makes it looks like Netscape is not included, and some people might think Netscape is included in the Netscape Navigator column.

I have no idea who Mozillar is. Please assume good faith. Thank you. -- Schapel (talk) 14:35, 18 April 2008 (UTC)

Er, no, there is only one other table that lists FireFox under Mozilla, the first OneStat.com table. All the other tables have a separate entry for either "FireFox" or "Mozilla FireFox", including the four other tables from OneStat.com.

The current format will leave people wondering why theCounter.com doesn't count FireFox. Also, the contribution to the overall stats in the Mozilla/Netscape column by either Netscape or Mozilla is extremely small (the Netscape release based on Mozilla has been in fact been discontinued), some 99% of that total is from FireFox. (Which is probably the reason that the ADTECH stats have not included Netscape since 2006).

I will assume good faith on this, but please in the future be more careful with your edit summaries. In my edit [reverting Mozillar] I mentioned that his revert gave no reason. If you then give in [your edit] an edit summary saying "the reason was given before", that makes it seem like you are referring to the edit of Mozillar, and sound like you are defending the action of yourself or of a friend. HermanHiddema (talk) 19:09, 18 April 2008 (UTC)

The problem is that the term Mozilla is overloaded. In the tables where both Mozilla and Firefox are present, the term Mozilla refers to the Mozilla Application Suite. Why don't we convert those columns to say Mozilla Application Suite? In the future, please use the talk page instead of engaging in an edit war. Then you will not be confused by edit summaries that are being used as if they are abbreviated entries on the talk page. -- Schapel (talk) 20:43, 18 April 2008 (UTC)

I did not engage in an edit war. I made a normal, useful edit, which you reverted twice. So please in the future use the talk page instead of using multiple reverts. HermanHiddema (talk) 22:33, 18 April 2008 (UTC)

I stand by the opinion that in the context of that table, at the top of that row, Netscape is a virtually meaningless term, as is Mozilla. The current title is completely confusing to any reader not intimately familiar with the Mozilla/Netscape/Firefox family tree. Also, please check the stats at theCounter.com and see when the last time was they actually included Mozilla in their stats. I cannot find it mentioned after 2004. HermanHiddema (talk) 22:33, 18 April 2008 (UTC)

Netscape means Netscape-branded browsers and not any other browser, very consistently through the tables. Firefox means Mozilla Firefox and not any other browser, very consistently through the tables. Mozilla is the only inconsistent heading I can find, as sometimes it means Mozilla-branded browsers, and sometimes it means just the Mozilla Application Suite. At least it is consistent in that the article linked to depends on the meaning. Do you have a concrete suggestion that will make the table headings more understandable, while not making them less consistent than they already are? The only way I can think that's possible is to change the text used for one or both of the headings currently listed as Mozilla, as that is the only inconsistent heading.

Your edit was not useful. It made a situation that I admit is somewhat confusing into a situation that is even more confusing. When you find multiple editors reverting your changes, please use the talk page instead of putting your edit back in, and claiming that a reason was not given when I gave the reason in my first revert. That's called an edit war. Let's be productive, alright? Do you have a concrete suggestion for improving the article, one that actually does improve the article? -- Schapel (talk) 12:31, 19 April 2008 (UTC)

Regarding consistency:

Source	Netscape	Mozilla	FireFox
theCounter.com	under Netscape/Mozilla	under Netscape/Mozilla	under Netscape/Mozilla
OneStat.com	under Netscape	under Mozilla	under Mozilla
OneStat.com (countries)	under Netscape	-	under Firefox
ADTECH:	under Netscape	under Mozilla	under Mozilla Firefox
Net Applications	under Netscape	under Mozilla	under Mozilla Firefox
W3 Counter:	-	under Mozilla 1.8	under Mozilla Firefox
WebSideStory	under Netscape/Mozilla	under Netscape/Mozilla	under Mozilla Firefox

Biggest inconsistency: FireFox is not listed by name in two of them, despite being present. The first source groups all three browsers under the heading of the two of them that make the smallest contribution. Now either you list all of them, or you list the biggest two (which I tried), or you list only the biggest. Listing the smallest two is nonsense. HermanHiddema (talk) 19:49, 19 April 2008 (UTC)

No table is listing the "smallest two" browsers. As I have mentioned before, Mozilla is an overloaded term, which can be used to refer to all Mozilla-branded browsers, including both the Mozilla Application Suite and Mozilla Firefox. What you tried to do was replace a column that included Netscape- and Mozilla-branded browsers (which is correct and consistent with the other tables) with a column that indicated only Netscape and Firefox, leaving out the Mozilla Application Suite (which is both incorrect and inconsistent with the other tables). Also note that in 2004, a year that the table includes, the Mozilla Application Suite was the most used browser of those three. Leaving that browser out would be not just a little wrong, but way, way off. The only inconsistency I can see is that sometimes Mozilla refers to the Mozilla Application Suite and sometimes Mozilla refers to all Mozilla-branded browsers, although which meaning is meant is indicated by the article the text links to. It's not perfect, but it's the best anyone has ever suggested. If you have a suggestion that makes the tables better without making them incorrect or less consistent than they already are, I'd be interested in discussing it. -- Schapel (talk) 22:48, 19 April 2008 (UTC)

If you have a suggestion that makes the tables better without making them incorrect or less consistent than after my change, I'd be interested in discussing it. HermanHiddema (talk) 18:50, 20 April 2008 (UTC)

That is what we have been discussing. You removed the Netscape from the column, even though the column includes Netscape usage. That made the table incorrect and inconsistent. By reverting, I fixed the problem. Do you still not understand? If not, all I can do is simply ask you not to change what you do not understand. -- Schapel (talk) 20:14, 20 April 2008 (UTC)

Yep, that's what we've been discussing. You removed the Firfox from the column, even though the column includes Firefox usage. That made the table incorrect and inconsistent. By reverting, I fixed the problem. Do you still not understand? If not, all I can do is simply ask you not to change what you do not understand. HermanHiddema (talk) 20:23, 20 April 2008 (UTC)

No, because Mozilla Firefox is a Mozilla browser. -- Schapel (talk) 20:48, 20 April 2008 (UTC)

Sigh. I give up. I have tried to make my point, but apparently you are unwilling to read or consider my arguments. I'm not going to spend endless time debating you over this, I have better things to do. HermanHiddema (talk) 11:25, 21 April 2008 (UTC)

If you have a concrete suggestion for improving the article, I'd love to hear it. What you have been doing is simply saying the way it's done now is wrong because Firefox is not included in TheCounter.com. That is obviously incorrect, as Mozilla Firefox is a Mozilla browser, so is included in the column labeled Mozilla / Netscape. -- Schapel (talk) 11:57, 21 April 2008 (UTC)

This column includes three browsers. Netscape, Mozilla (App Suite) and Firefox. Two of these browser have been discontinued. The third contributes 99% of current hits. Why are you so intent on not mentioning Firefox? HermanHiddema (talk) 12:03, 21 April 2008 (UTC)

I'm not intent on not mentioning Firefox in the least, as I have said many times. I'm intent on keeping the headings in the article consistent. Why are you intent on not adding Firefox to the OneStat table so the change is at least consistent? If nothing else, let's make changes consistent. That has been the problem with your edits. If you still do not understand, please read this post again until you do. Thanks. -- Schapel (talk) 12:08, 21 April 2008 (UTC)

Keeping the headings consistent? Wow. You must not have actually read anything I wrote. Try looking up the word "consistency" in a dictionary, then refer to the table I made earlier which shows how inconsistent the headings are. HermanHiddema (talk) 12:14, 21 April 2008 (UTC)

And now you have made them even more inconsistent, by applying a change to one table and not another. Please explain how your edit to add Firefox to TheCounter and not to OneStat fixes a problem. -- Schapel (talk) 12:18, 21 April 2008 (UTC)

Sure, OneStat changed as well. Was that so hard? HermanHiddema (talk) 12:22, 21 April 2008 (UTC)

No, not at all. What was your issue with simply proposing that change and getting consensus before making it? That's all I was asking for. I propose that we change the Mozilla heading to point to the Mozilla Application Suite in all tables, because now the meaning of the Mozilla heading has changed. -- Schapel (talk) 12:26, 21 April 2008 (UTC)

What do you think I've been trying to do for the last few days? I tried to improve the page by making a constructive edit. You reverted that and then never came up with any constructive suggestion yourself, arguing instead for the article to remain as is. HermanHiddema (talk) 12:31, 21 April 2008 (UTC)

Well, if you're the one wanting to change it, why should I propose a change? If you want to make a change, you should propose it. There's another inconsistency in the article now. In some headings Firefox is listed as Mozilla Firefox and in others it's listed as just Firefox. If, as you say, many people are not aware that Firefox is a Mozilla product, will that not confuse some people to think that Firefox might be different from Mozilla Firefox? Gee, I wish we could have discussed these changes before starting to make them. -- Schapel (talk) 12:34, 21 April 2008 (UTC)

The way wikipedia usually works is that people work together, making edits to improve the article, without discussing every change on the talk page first.

So if you note that the "Firefox"/"Mozilla Firefox" thing is inconsistent: Go ahead and change the article! This is a wiki, you don't need to get approval before editing. HermanHiddema (talk) 12:38, 21 April 2008 (UTC)

Oh, well, I would rather change it back to the way it was, before these new problems were added. Is that okay? Alternatively, we could discuss further changes. I will leave it up to you. -- Schapel (talk) 12:44, 21 April 2008 (UTC)

What new problems? HermanHiddema (talk) 12:53, 21 April 2008 (UTC)

Among the new problems are the following:

• You have linked to Firefox, which seems to keep bouncing back and forth between being a disambiguation page and redirecting to Mozilla Firefox. We should avoid redirects and disambiguation pages.
• Sometimes Firefox is referred to as Mozilla Firefox and sometimes it is referred to as just Firefox. It should be the same everywhere.
• It seems as though the term Mozilla now always means Mozilla Application Suite although the article linked to is Mozilla instead.
• Sometimes Mozilla and Firefox is listed as Mozilla/Firefox and sometimes it is listed as Firefox/Mozilla. Is there a difference? Is not, it should be the same everywhere.

There may be other problems as well, but a quick look reveals at least those. -- Schapel (talk) 17:07, 21 April 2008 (UTC)

• When I created the Firefox link, it was redirecting correctly, I was not aware there was an edit war going on there. We should avoid linking to disambiguation pages, but redirects are ok (see WP:REDIRECT#NOTBROKEN).
• I have already given my opinion on the naming issue below. You're the one insisting on discussing it before changing, so please respond regarding the naming issue.
• Easily fixed, no need to discuss, IMO
• It is never listed as Firefox/Mozilla, but is listed once as Firefox/Mozilla/Netscape. This is my preference as it is in order of current percentage share.

IMO, instead of complaining about such simple issues, you should fix them. I'll do it this time HermanHiddema (talk) 18:49, 21 April 2008 (UTC)

Uh, when I fixed the issues you created, you threw a fit because I didn't fix it the way you wanted. That's why I asked you to discuss the issues first. If you have no problem with me fixing issues without discussing, please don't revert my fixes. You can't have it both ways. -- Schapel (talk) 20:20, 21 April 2008 (UTC)

You didn't fix the issues, you simply reverted. When I made that edit, I didn't do it to be an asshole or make a point, I did it because I found the article lacking and tried to fix it. The fault I found with the article was the fact that Firefox wasn't mentioned in a place where it obviously (IMO) should be mentioned. Now it seems to me that you could, from there, take two basic approaches.

1. Decide that Firefox did not deserve mention in that place, in which case you could decide to revert with an edit summary saying something like "Firefox is not important in this context" (Which was something I would have contested).
2. Decide that mentioning Firefox in that context is a valid idea, but that it introduced issues of inconsistency, which you could then have tried to fix (or asked me to fix).

The second would have been my approach, it is IMO the productive way to move forward. Instead, you chose to revert, which is IMO an unproductive moving backward. HermanHiddema (talk) 09:29, 22 April 2008 (UTC)

I don't know if they can be fixed without reverting. You have given no proposal for how the new inconsistencies can be removed. You've added inconsistencies, won't fix them, and refuse to discuss how to fix them. You just insist that I should do it. I know one sure-fire way of doing it, which you don't like. Either let me fix them my way, you fix them your way, or we can work together to discuss how to fix them. Do bark at me to do it, then bark at me again for fixing it wrong. -- Schapel (talk) 16:31, 22 April 2008 (UTC)

You seriously think that there is no possible way that we can include Firefox in the column heading without introducing unfixable inconsistencies?

Further, I would like to note that at least one of the "new problems" you listed was in fact an old problem already there before my edit. Both "Firefox" and "Mozilla Firefox" were already being used. Furthermore, there was already an inconsistent linking to Netscape, with links exiting to both "Netscape" and "Netscape (web browser)", which I have now fixed.

I have never insisted that you should fix issues, only suggesting that you could, but also that you could ask me to do it. And saying that I "won't fix them" is equally untrue. I have fixed many issues, both those pointed out by you and others that I have since found myself. HermanHiddema (talk) 18:57, 22 April 2008 (UTC)

Regarding "Firefox" or "Mozilla Firefox", I would prefer just "Firefox". The software maker isn't mentioned in the Internet Explorer column either, and I don't think changing that to "Microsoft Internet Explorer" would improve the readability of the article. HermanHiddema (talk) 12:53, 21 April 2008 (UTC)

I don't care, as long as it's consistent. Will you revert my changes if I pick one or the other? -- Schapel (talk) 20:20, 21 April 2008 (UTC)

No, I will not revert, I have never reverted any change of yours that was productive and moved forward. I have only reverted your own reverts, because in my opinion they kept the article in a place where it was lacking.

I have now fixed all references to be consistently named "Firefox" and link to "Mozilla Firefox" (to avoid that constantly switching Firefox page). HermanHiddema (talk) 18:57, 22 April 2008 (UTC)

It looks okay now. Thanks for fixing everything. -- Schapel (talk) 19:09, 24 April 2008 (UTC)

1999 to present vs. Present to 1999

As the dates are listed backwards chronologically, I feel it would make more sense to label the second section 'Present to 1999' rather than its current '1999 to Present'.

Introducing dates backwards is not unprecedented: many local history books (e.g. some by Reece Winstone [4]) write dates this way round when the subject matter is presented in reverse chronological order. Thoughts? LHMike (talk) 16:53, 19 April 2008 (UTC)

Seems reasonable to me. Wikiolap (talk) 19:23, 19 April 2008 (UTC)

Ok I've made the change. If there are any further thoughts please mention them here. LHMike (talk) 12:45, 20 April 2008 (UTC)

This is illiterate nonsense. Time flowed from the past to the present. The heading should do the same, i.e. "1999 to present" or "present from 1999", but not "present to 1999". How can you go to something that disappeared nine years ago? --Harumphy (talk) 14:26, 20 April 2008 (UTC)

No one is saying time flowed backwards. It's a sort order. Normal alphabetical order is A to Z. Reverse alphabetical order is Z to A. Compare lowest to highest, highest to lowest, 1999 to present, present to 1999. The order used in the tables indicates it's sorted from the present to 1999. Personally, I don't really care one way or another, but claiming it's illiterate nonsense is a strawman argument. -- Schapel (talk) 14:37, 20 April 2008 (UTC)

Yes, I changed the heading to reflect the fact that the tables show the information from the Present to 1999. If the tables were sorted the other way round, the original heading would have been the one to use. LHMike (talk) 14:51, 20 April 2008 (UTC)

Proposal: New split by time

Currently we have two sections - "Before 1998" and "1999 to present". From looking into the tables, it looks like there was a significant change around 2004-2005. Netscape either disappeared completely from some sources, or carries very small percentage. Safari, on the other hand, appeared around same time and now has non-trivial numbers in the tables. So I propose to make a split and have sections 1999-2004 (or 2005) and "2004 (or 2005) to present". This should also help with Netscape/Firefox/Mozilla confusion, since Firefox these days is clearly stands on its own. Opinions ? Wikiolap (talk) 21:14, 19 April 2008 (UTC)

Why do we have two sections in the first place? Why not simply order the tables in some sort of consistent chronological order? If we are going to have two sections, I would rather it either be like it is now, or divided into actively updated stats and older inactive stats. If we divide chronologically at 2004 or 2005, that breaks several of the tables up into two sections and makes the information less readable. -- Schapel (talk) 13:14, 20 April 2008 (UTC)

I am mostly looking for a way to go into the future without having columns which are empty now, i.e. Netscape. We don't have frequent changes in the browser market when one browser disappears and another is born, so when something like that happens, it seemed to me that it would justify split. Wikiolap (talk) 22:45, 21 April 2008 (UTC)

XiTi?

I wonder why this article doesn't include any numbers from XiTiMonitor. Their numbers are already being used in usage share of desktop operating systems. Although they only give out figures on non-Firefox browsers for Europe, the ADTECH numbers are also Europe-only. They seem to be respected, and I would put more stock in their numbers than in TheCounter.com, for example. TheCounter's browser numbers seem plausible, but going on to click their OS numbers, those results seem a little weird.

Another point where I feel this article could be enhanced, is that the "#Overestimation and underestimation" section only discusses technical problems related to identifying unique user agents. It doesn't discuss at all the issue of how representative a set of sites is. For instance, I suspect that Net Applications has a US-centric group of sites, and that this is one reason their Safari (and Mac) numbers are higher than most other sources. It's also interesting that OneStat, which reports some of the highest Internet Explorer numbers, is headquartered in the Netherlands, which is a particularly Firefox-resistant bastion of IE usage (according to XiTi and OneStat itself). I've seen blog talkbacks commenting on the outsized (though still only 2%) number of Latvian visitors in W3Counter's stats. Many of these firms are probably reporting "global" numbers that are actually skewed towards a particular country. And that's even before taking into account other demographics, whether a site is more likely to be accessed from a work or home computer, and other factors. Plus, these numbers do reflect usage, not necessarily users, as any measurement based on web hits is going to skew in favor of the browsers used by heavy internet users and those with fast connections. --Groggy Dice T | C 14:06, 27 April 2008 (UTC)

Ads and Ad Blocking

Is Adtech an ad merchant, as it appears? Perhaps there are other sources for Europe. I am sure that Adtech is a totally reputable, trustworthy, unbiased and accurate source... But consider the above (#1?) point about the versioning. The argument was made that anybody who makes the effort to download an alternative browser is also mroe likely to ugrade it. I expect that someone desiring this kind of control over their system might also be more inclined to want to remove annoyances from web pages (such as ads).. As far as I know, IE doesn't come with any built in ad blocker (maybe IE7 or IE8 does?)... Mozilla makes Adblock Plus available through their own site as an extension. Opera comes with a built-in ability to filter ads, I believe. I think that you do have to turn on both of these things - but I also think that anyone who is aware that there exists more than one way to view the web would also be likely to consider the possibility of viewing it without ads.
Programs like Norton Internet Security block ads....
Blah blah blah blah blah... Basically what I am saying is, I _suspect_ that Internet Explorer users are probably less likely to filter out advertisements - and that therefore there may be an unfortunate Skew on the Adtech data? Just a suggestion, not a criticism. Not so sure of it that I'm going to go find more sources of statistics, but I thought I'd put it out there. I didn't notice it anywhere on this talk page. Cheers :-) 125.236.211.165 (talk) 09:28, 7 May 2008 (UTC)

Absence of the America Online Browser

Why isn't there an AOL browser in the article or in the graph? I thought that a lot of AOL users used such a browser in the past and I figured that at least a few are still using it. However, the article and graph do not mention the browser at all, even though browsers with tiny usage shares are included, such as Opera, which the graph gives a 0.66% usage share. I looked for an article on the browser and it appears that there have actually been several of them. I cannot find an article for the original browser, but the others are AOL Explorer and AOL OpenRide. -- Kjkolb (talk) 14:10, 24 May 2008 (UTC)

Because stats sources generally do not give stats for AOL Browser. -- Schapel (talk) 14:22, 24 May 2008 (UTC)

Because it's basically Internet Explorer ?

Description: AOL 9.0

User Agent: Mozilla/4.0 (compatible; MSIE 6.0; AOL 9.0; Windows NT 5.1)

--87.194.174.252 (talk) 12:46, 3 December 2008 (UTC)

Visualization - graphs

There is a ton of table data on the page - what about some graph visualization?--Kozuch (talk) 12:10, 12 June 2008 (UTC)

How about this? HermanHiddema (talk) 14:08, 12 June 2008 (UTC)

Great That's pretty fantastic - I would kindly request that you consider using the same colors that I used in my pie chart for consistency's sake. I chose those colors because they are emblematic of the browser itself. —Justin (koavf)❤T☮C☺M☯ 18:17, 12 June 2008 (UTC)

Also You may notice that two quarters are missing. —Justin (koavf)❤T☮C☺M☯ 18:19, 12 June 2008 (UTC)

Ok, sure, I just chose some primary colors. By the way, I understand the blue for IE, orange for FF and red for Opera, but why is green emblematic of Netscape? They always had a dark blue & black logo. Anyway, if you check the URL for the chart, you can change the colors right there, right after "&chco=" there are five hexadecimal colors listed.

The missing quarters are missing in the table too, aparently theCounter.com has no data for the second half of 2003. I could interpolate perhaps, is that an idea? HermanHiddema (talk) 08:14, 13 June 2008 (UTC)

Clearly I did not pay attention to the missing quarters - obviously. If there is no data, there should be no data presented, of course. I chose green because I recall green in Netscape 8 and 9 - the only versions I ever used (E.g. the buttons in this menu.) I don't know if this is just a problem with my eyes confusing a hunter green with dark blue; maybe I just have some problems seeing that end of the color spectrum. Anyway, from my recollection and this picture, Netscape was green. It's also nice considering how it's another simple primary/secondary color. I'm not terribly familiar with Google's chart API, but looking at the URL it can clearly be modified. Thanks. —Justin (koavf)❤T☮C☺M☯ 08:59, 13 June 2008 (UTC)

Ah, I see. For theCounter.com, the Netscape referred to in the table is Netscape Navigator (which has this logo), the versions of Netscape up to 4.7, popular in the 90s. The newer Netscape versions (version 5 and up) are listed in the "Netscape/Mozilla/Firefox" column, as these browser share the same original code base. How about the following color scheme:

• Netscape: dark blue
• Internet Explorer: light blue (as in your chart, perhaps slightly lighter to distinguish it from Netscape)
• FireFox: orange (as in your chart)
• Opera: red (as in your chart)
• Safari: green (another primary color, I dislike using grey because it is so often used for "other" in all sorts of charts).

-- HermanHiddema (talk) 10:36, 13 June 2008 (UTC)

Green and grey That logo still looks dark green to me rather than dark blue. Does this look green or blue to you? I guess we're just perceiving it differently, but I'm seeing a more-green-than-blue teal. I still prefer the colors that I had before for the rationale that I had before - Safari has that brushed chrome skin and silver in its compass logo, so silver (rather than grey, actually) is what I used. I arbitrarily chose light blue for "other" since it stood out compared to other colors; I would have chosen yellow if it had enough contrast with white, but it doesn't. —Justin (koavf)❤T☮C☺M☯ 19:58, 13 June 2008 (UTC)

They are not far apart in color, though one uses a gradient and the other doesn't. They are both pretty close to teal (images and color boxes added for reference). In my perception, this is closer to blue than to green, but I guess it comes down to just that, a matter of perception. Teal is described as "greenish blue", and its hexadecimal value is exactly at the halfway point between green and blue. Like cyan, but darker. I still dislike using grey for safari though, no matter how grey or silver their browser looks. It just doesn't look right, grey is too much of a neutral color. HermanHiddema (talk) 13:21, 19 June 2008 (UTC)

Here's a new effort. I've changed the colors, changed the type of chart and shown only percentage over 50%. As this is a line chart, I can simple let the line from 2003 Q2 to 2004 Q1 run straight over the area with missing data, which makes the chart look nicer. HermanHiddema (talk) 11:11, 13 June 2008 (UTC)

Table formatting

I have, as an example, reformatted the first table (theCounter.com), giving every browser column the same width (in the em measure, so it scales well with font size changes). Until now, the width of the column was determined automatically, based on the lenght of the name of the browser, which looked somewhat messy. Any objections to formatting more tables like this? HermanHiddema (talk) 12:34, 12 June 2008 (UTC)

Width Personally, I am almost never in favor of stating the width of the columns and letting the browser and user do that themselves. —Justin (koavf)❤T☮C☺M☯ 18:18, 12 June 2008 (UTC)

The problem I have with that is that it makes all the columns different widths, so the data for browsers with short names (Opera, Safari) gets crunched together, while that of browsers with longer names (Internet Explorer, Netscape Navigator) gets a lot of whitespace around it. I don't know of any way to tell the browser to make a certain set of columns all the same width without also telling it, to a certain extent, how wide. Using widths in em at least has the advantage that the width still depends on font size, so people with different font sizes (eg those with bad eyesight) still get a clean looking table. HermanHiddema (talk) 08:20, 13 June 2008 (UTC)

Style and preference And I personally have no problem with the columns having different widths. If this is something you really like, I'll defer - it's a matter of preference, really. Now, one way to tell it how wide to make the columns without defining the width per se is to use a percentage, which is the only kind of width I ever use (except for the occasional IE CSS hack with pixels.) Again, if you like ems, go for it; there is something to be said for it. —Justin (koavf)❤T☮C☺M☯ 09:01, 13 June 2008 (UTC)

Percentages are good too. Like ems, they do not define the width in pixels but let it depend on another measure (browser window size for percentages, font size for ems). Both are good methods in my opinion, and the choice between them often depends on the visual effect you want to achieve. I will adjust the other tables in a few days or so if no other editors raise objection. HermanHiddema (talk) 09:24, 13 June 2008 (UTC)

Usage stats from all MediaWiki projects?

I'd like to know if stats are kept for individual projects or aggregated across them all, and whether they are big enough a group of sites as far as visiting numbers go to be representative sample for inclusion in this article. —Preceding unsigned comment added by 207.10.232.172 (talk) 05:00, 14 July 2008 (UTC)

This isn't really possible at the moment as hit counting and other types of logging has been disabled on Wikipedia servers. Xenon54 12:44, 14 July 2008 (UTC)

This article needs cleanup

I don't understand the need for data from SO MANY sources. As the article explains, there are a lot of factors as to why the market share numbers ARE inaccurate. The purpose of the market share numbers is to give a general idea as to the market share of the different browsers. Not to add as many market share data from different providers so people have a hard time interpreting the data. Adding more and more data does not increase the accuracy of the data.

I wholeheartedly agree. More data of this kind does not mean it is more accurate, since usage share calculation via user agent strings is known to be inaccurate, and the amount of inaccuracy is not known, and has never been seriously studied. 66.119.170.242 (talk) 02:53, 13 January 2009 (UTC)

No, it does not make the data more accurate. That is not the point. The point is to have a neutral point of view, so that we don't favor one source of stats over another unless there's reason to believe the source is seriously biased. -- Schapel (talk) 11:59, 13 January 2009 (UTC)

My suggestion would be to choose two data sources to include in the article TheCounter.com and OneStat.com seem good as these have a lot of data points and seem to represent the average.

Also, please only one pie chart at the top of the page. The pie chart is to give a quick interpretation of the market share numbers; can you imagine how confusing two pie charts is. —Preceding unsigned comment added by 145.97.203.86 (talk) 10:23, 7 August 2008 (UTC)

Pie Chart

Does the pie chart have to be so HUGE? 87.82.130.228 (talk) 11:11, 31 August 2008 (UTC)

It doesn't, but I think the current size is OK, since it fills the right side of the content menu, but doesn't overflow beyond that. Wikiolap (talk) 18:16, 31 August 2008 (UTC)

The pie chart also doesn't reflect the numbers under it. I'm not normally a fan of IE or Microsoft but I can see the IE slice is less than 75% but that the number underneath for it is 78% or so. So the pie chart is out of sync with the numbers. Martin Packer (talk) 09:03, 8 September 2008 (UTC)

Forget it... The numbers changed under me. :-) —Preceding unsigned comment added by MartinPackerIBM (talk • contribs) 09:04, 8 September 2008 (UTC)

Chrome

Hasnt chrome already exceeded operas share? —Preceding unsigned comment added by 92.16.148.143 (talk) 12:23, 14 September 2008 (UTC)

Yes and no The data are mercurial, especially upon the release of a high-profile piece of software. Whether or not this trend is sustained and seen noted amongst a variety of counters has yet to be seen. —Justin (koavf)❤T☮C☺M☯ 12:51, 14 September 2008 (UTC)

Who must recognise what has been speculated to be their share for it to be considered a likely reflection of the actual trend? —Preceding unsigned comment added by 92.16.148.143 (talk) 01:23, 15 September 2008 (UTC)

No one and everyone There is no central authority for web usage statistics; otherwise, we could simply quote them and be done with the plethora of tables here. As numbers are published by sources (e.g. third-quarter numbers from W3Schools), then Chrome can be taken into consideration. Right now, the numbers are only daily usage statistics, which are ephemeral. —Justin (koavf)❤T☮C☺M☯ 02:04, 15 September 2008 (UTC)

TheCounter.com stats

TheCounter.com stats for Mar 2008, Apr 2008, May 2008, Jun 2008 are all for more than one month. They have a starting date of the 1st of Feb 2008, so those stats end up covering 60, 90, 121 & 151 days. So just wondering if the stats for 2008 Q1 & 2008 Q2 should still be shown (or shown with a more of a warning than usual about the values). -- 194.74.176.98 (talk) 12:35, 16 September 2008 (UTC)

I have to chime in here. That latest stats from TheCounter.com are not sourced. It seems to me that the person who edited the table calculated the numbers from the material he found on TheCounter.com. I think that is original research and beyond the scope of Wikipedia. Those statistics should either be correctly sourced or removed. What do you think? -- Repetition (talk) 22:00, 28 September 2008 (UTC)

Stats that can be easily verified by a reader doing simple calculations from numbers in a cited source is not original research. In this case, the cited source directly supports the stats in the article according to the no original research policy. It is not an original synthesis of materials. -- Schapel (talk) 00:24, 29 September 2008 (UTC)

I accept the calculations, because TheCounter gives the number of hits on which their results are based. If the calculations were based on averaging the percentages for the three months, I would agree that this had original research problems, because traffic could be higher or lower in particular months. But since the specific number of hits can be added up and then tabulated into percentages, there are no assumptions involved.

Note that I do have problems with TheCounter as a source. I know Vista is having a slow adoption, but their Vista numbers are blatantly preposterous. But that's separate from any OR issues. --Groggy Dice T | C 17:39, 1 October 2008 (UTC)

I agree that the counter is a less than reliable source, and I would agree to removing it from the article. But the at least it should be demoted from the first to the last source referenced in the article. - Misty Willows (talk) 15:51, 2 October 2008 (UTC)

What is the ecidence that TheCounter is unreliable source ? Just your perception about Vista adoption ? Are you claiming to be more reliable source than TheCounter ? It should not be removed from the article. Wikiolap (talk) 17:15, 2 October 2008 (UTC)

I agree. There is something wrong with the numbers from TheCounter. It says 14 % Windows NT and 0 % Windows Vista.???? TheCounter should be removed. Mortenwendelboe (talk) 14:32, 3 May 2009 (UTC)

The information on usage share of operating systems is not used in this article, so whether it's accurate or not has no bearing on whether the information on usage share of browsers should be removed. -- Schapel (talk) 15:33, 3 May 2009 (UTC)

Where Are the Wikipedia-User Browser Statistics?

I've made comments on this subject several times off in the labyrinth of administrative sections of Wikipedia/media -- with of course, who knows who ever having seen them.

Wikipedia needs to make Wikipedia-visitor browser statistics available!

First, just simply by issuing a press release each month with browser statistics, by Wikipedia language edition, and by IP-detected region.

But then with a technical implementation to have a "Toolbox" link in the lefthand nav of each article page to see what browsers (and regions, by IP) have viewed that article.

What browsers are Pokemon fans in Japan using? Inquiring Web minds want to know.

Other than Google, WIkipedia is the best representative slice of the general Web population I can think of -- and its public information service mandate means it's really inexcusable not to make this information publicly available.

So who knows someone on the Wikimedia board? Who knows how to get a concept like this through that labyrinth of administrative pages to where core Wikipedia decision are made?

—Preceding unsigned comment added by 75.36.147.156 (talk) 08:52, 4 October 2008 (UTC)

You shouldn't have to know somebody on the board. Anyone with access to the logs can gather this info. 72.228.150.44 (talk) 07:26, 8 December 2008 (UTC)

New stats from Net Applications (Q4, 2008)

Firefox is now at 20.06% ([5]). --82.171.70.54 (talk) 21:02, 8 November 2008 (UTC)

Thanks for the tip, but remember that Wikipedia is the encyclopedia that anybody can edit. Rwxrwxrwx (talk) 11:01, 9 November 2008 (UTC)

Errors in recent OneStat data

If you click on the source links to the TheCounter.com data for recent months, the time spans are shown as follows:

• For September: Fri Feb 1 00:01:02 2008 - Tue Sep 30 23:58:00 2008 243.0 Days
• For August: Fri Feb 1 00:01:02 2008 - Sun Aug 31 23:58:00 2008 213.0 Days
• For July: Fri Feb 1 00:01:02 2008 - Thu Jul 31 23:58:00 2008 182.0 Days

...and so on. If these are correct, then the derived quarterly figures in the article are incorrect, being weighted towards February. —Safalra (talk) 10:42, 24 November 2008 (UTC)

I simply subtract off the earlier stats to arrive at the correct figure. For example, to get the stats for September, subtract the stats from August from the stats for September, and you can get the stats for September only. -- Schapel (talk) 12:26, 24 November 2008 (UTC)