The Browsers of 2009

In a follow-up to this question I’ve begun pondering what the most-relevant browsers of 2009 will be. I tend to determine relevance by the question “Is this browser cost beneficial to us supporting it” being answered by a significant number of developers and corporations.

For example, I would probably rate the current, 2008, list as follows (in order of cost-benefit):

• IE 6
• IE 7
• Firefox 2
• Safari 2 & Safari 3 (tie)
• — common cut-off point —
• IE 5.5
• Opera 9.2

Starting in 2009 I predict that the list will be similar to the following:

• IE 7
• IE 6
• Firefox 3
• Safari 3
• — common cut-off point —
• Opera 9.5

Not a whole lot changed, but a couple points to consider:

• I wouldn’t be surprised if the switch between the dominance of IE 6 and IE 7 finally occurred.
• Firefox 3 will probably be released late spring/early summer – thankfully uptake will be fast (as it has, generally, been with past versions).
• Safari 3 is already shooting way up, Safari 2 will be a thing of the past come 2009.
• Opera will definitely have 9.5 out by then – maybe 10? Although, even if 10 is out, it definitely won’t have a market share by next January.
• IE 5.5 will definitely squeeze out it’s last remaining % into cost-benefit obscurity.

What’s interesting about analyzing the cost-benefit of a browser is that it’s done completely differently from straight-up analysis of browser market share. It’s really a combination of market share and time that’ll be spent customizing your application to work in that browser. Here’s a quick chart to represent my choices:

The “Cost” is represented by the % of time that will be spent, beyond normal application development, spent exclusively with that browser. The “Benefit” is the % of market share that the browser has. Note that any browser that has a higher cost, than benefit, needs to be seriously considered for development.

What’s interesting about this is that it use to be much-more clear-cut when choosing to develop for IE 6 – it had 80-90% market share so it’s benefit was always considerably higher (or, at least, equal to) the time spent making it work in that browser. However, in 2009, that percentage will be considerably less (I’m estimating 35%, to IE 7’s 45%) making it far less attractive as a platform. Note that Firefox and Safari, due to their less-buggy nature (and standards compliance) always have a higher benefit than cost, making them an easy target to work towards. Opera is problematic, however. It’s, continued, low market share makes it a challenging platform to justify. It’s for this reason that major frameworks, like Prototype, have ignored working with Opera up until this point – and understandably so.

Now it’s not always a one-to-one trade-off in-between cost and benefit. I think it would even be safe to say that benefit is, at least, twice as important as cost. Ultimately, however, this depends upon the choices of those involved in the decision making and the skill of the developers working on the compliance. Is an extra 4-5% market share from Safari 3 worth 4-5 developer days? What about the added overhead to Quality Assurance and testing? It’s never an easy problem – but it’s one that we can, certainly, all get better at over time – but really only through hard work and experience.

Posted: January 31st, 2008

48 Comments (Show Comments)

1. Ludovic (January 31, 2008 at 3:28 am)

   How about IE 8 ?

2. John Resig (January 31, 2008 at 3:32 am)

   @Ludovic: IE 8, final, definitely won’t be out this year (or even if it is – it’ll be late this year). While I definitely expect an IE 8 beta to be ready by the start of 2009 (and, hopefully, be released during early 2009) it won’t have any traction amongst the general user population by that time. Thus, it’s not really relevant to these calculations.

3. timoni (January 31, 2008 at 3:32 am)

   When you twittered “The Browsers of 2009” I was really just hoping for an animated .gif of Tom Cruise playin’ Wii without a remote in “Minority Report.” Ah well. ;)

   Do you really think many people are going to ignore the mandatory IE6 to IE7 upgrade next month to warrant IE6’s top placement on this year’s list? Also, I know you weren’t dealing with mail apps, but in terms of ‘time spent exclusively’ dealing with a Microsoft product that displays HTML, I’d say Outlook 2007 is going to be the biggest pain in the @$$ for years to come (at least for those sites that have formatted emails).

   As for your final question re: the worth of Safari 3 development, it seems that 3-4% market share would never be enough on its own to merit significant development time unless you had some serious user complaints. But depending on your industry, you might have a much larger percentage of viewers using Safari 3, right?

4. Michael(tm) Smith (January 31, 2008 at 3:33 am)

   Hi John,

   Any thoughts on what a similar breakdown for 2009 browsers on mobile devices might look like? Or for Web-capable devices other than PCs (e.g. Nintendo Wii)? Or thoughts on if or how much more relevant you think mobile browsers and browsers on non-PCs night become in 2009 as an significant consideration for Web developers as a target for delivering content to.

   It seems like for 2008 at least, the worldwide non-PC browser landscape looks similar to what you have here — except just turned on its head. That is, with Opera at the top, then Webkit, then the common cut-off point, then (pocket) IE, and Firefox not on the board at all.

   I think the Nokia acquistion of Trolltech and Qt (which now includes Webkit as a core component) is going to make things at the top two non-PC positions at little more interesting in 2009. And news about Skyfire is encouraging. And it would be great to see Christian and team have a new browser getting preinstalled on devices by 2009.

5. blufive (January 31, 2008 at 3:41 am)

   I assume that graph is for the 2009 figures? If so, I think we’re already part-way there. The browser stats I’m getting off the sites we maintain (UK insurance) are already showing IE7 ahead of IE6 by a statistically significant margin. We’re also seeing more Safari 3 than Safari 2, but that’s less clear-cut statistically. Those may be geographically-restricted assumptions, though.

   Ludovic: I think that depends when it’s released (given past experience and MS’s current position, I’d be surprised if IE8 went final much before the end of the year; the IE7 public beta period was over 9 months) and how fast it’s picked up (IE7 took over a year to overtake IE6, though it was significant much quicker than that).

6. John Resig (January 31, 2008 at 3:47 am)

   @timoni: I suspect that a large number of corporate users are still stuck in IE 6 land for a bunch of fun reasons (intranet apps, security concerns, who knows). I’m estimating that it’s market share will be approximately 1/2 of what it is now – which seems reasonable (possibly even a little aggressive, considering its current trend). I would love to be surprised here. One thing is certain, though: Whatever the final number is, it’ll still be high enough to warrant concern for 2009.

   It really will depend on your industry for a lot of these. For example, right now Safari 2 is of concern for a number of users who have old computer bases that haven’t upgraded yet (like universities with computer labs). This will undoubtedly change this year as everyone gets slowly up-to-date. But, yes, it really does depend on how much that time is worth to you. I’d argue that since the cost of development for Safari is so low, at this point, that it makes for an easy 3-5% market share win.

   As far as email goes, I feel real bad for email newsletter developers. Hearing their woes makes me feel happy that we have it as good as we do now.

   @Michael: I wish I knew! Major mobile sites are notoriously crass about releasing their mobile browser usage data – it’s worth good money to them. I think it’ll be safe to say, though, that by 2009 the number of standards-capable browsers will be much larger than it is now. If nothing else, I think that this shift in the capability of browsers and the bandwidth of the mobile providers will lead to web sites gracefully fitting in on the mobile web. If nothing else, the iPhone is starting to teach web developers that their current, standards compliant, sites can work just fine – just trim and gzip your contents and you’ll be that much faster.

   @blufive: Good point – I may have even been a little cautious with my rankings. Whatever happens, though, I think we can all be pleased with a large downward shift in IE 6 and Safari 2.

7. Anonymous (January 31, 2008 at 4:01 am)

   These statistics are flawed, you cannot compare percentages this way.

   Lets say a supporting browser A costs 40%, and it has 40% market share. In your graph the costs would be the same as the benefits.

   However, most businesses have the intention of making money. One of the ways they do this is by only taking actions that make more money than they cost. So the costs of supporting all the browsers they support, is supposed to be less than the benefits of doing so.

   Imagine supporting all the browsers they support costs $100.000 a month. Their website is working, it rakes in $200.000 a month. So you could say that supporting browser A costs $40.000 and has $80.000 benefit, so overall that’s positive.

   Now imagine supporting browser B costs 50% and has a 30% benefit.
   It would look like a really bad idea on your graph, the costs are higher than the benefits. In reality, its reversed. The costs are $50.000 and the benefits are $60.000.

   You can’t fix this without knowing what the ratio of costs vs. benefits is. All you can do is recognize that the units of the two series are not really comparable. So a more correct way to show would be to have two graphs, or perhaps one with costs going down from the middle and benefits up or something like that.

8. Pete B (January 31, 2008 at 4:54 am)

   I think ie6 support is going to be essential for a few years yet. But to be honest, I’m so used to working round its bugs it’s a lot easier than I used to find it.

9. pd (January 31, 2008 at 6:57 am)

   John you seem to be limiting your discussion to the start of 2009.

   Assuming IE8 is released at least 1st quarter 2009, given the way they are building it to seamlessly replace IE6 and 7 with the meta switch, what is to suggest that MS will not release IE8 to update channels with significant (at least 20%) penetration to at least non-pirate users by mid 2009?

   FWIW I have access to one site with a substantive audience (if 2,300 visits a day is substantial) based in Melbourne, Australia. We’ve just seen the share of IE7 overtake IE6, most likely due to MS allowing pirates to get IE7.

   I guess if MS stagger the release of IE8 first to customers, then pirates and then force it on everyone like they’ve done with IE7, we could well be waiting until 2010 for IE8 to get traction as you suggest.

10. Robert Kaiser (January 31, 2008 at 7:13 am)

    It sucks that supporting different browsers is even a question of cost, and it sucks even more that everything “below” the cut-off point is usually explicitely excluded from visiting the web page, which discriminates against minorites and is a racist attitude even though it would be cost-neutral to just send any unknown browser the most standards-complaint version you have (usually the Firefox version) and get additional visitors for free, as the vast majority of the otherwise cut-off browsers are very similar to the supported ones, just with a different skin color (user agent string).
    Send the Firefox/Gecko version to every unknown browser instead of closing those out, and you get more users for free!

11. Andrew (January 31, 2008 at 8:14 am)

    Cost-benefit?

    I wonder why Opera and Safari are so expensive. Personally I start developing my pages against those two – the most advanced – browsers, and than move on to Foxy, and IE6. I found that it’s much easier to move on in this particular order: Opera 9.2, Safari 3, FF 2, IE 6. If you start with Firefox you are doomed to tweak your page both ways – to more standard-oriented (O and S) and to less standard-oriented (well, completely disoriented – IE).

    I believe it is a common problem of today’s web design: people tend to start from FF because it’s very powerful and at the same time very-very forgiving. I can’t say the same about Safari, so it’s painful to move towards it, and as a result no one does it. Webmasters tell users: “Choose a better browser, man”. That makes me feel sad.

    I don’t mean that FF is less standardized than Opera or WebKit, though, so don’t use my post to flame about which is the best. I’m the most browser-neutral person – I use four different browsers daily (Well, ok, IE sucks, and I use it for Yahoo Music and Yahoo Security only).

    What really surprises me is why people who use powerful JavaScript libraries like YUI or jQuery cannot support all four families (IE, FF, O and S).

    I’m puzzled.

12. Andrew (January 31, 2008 at 8:24 am)

    @Robert Kaiser

    1++++

    I love the way Yahoo handles their new mail. They bought it from Oddpost guys, who wrote it originally for IE users.

    Today it’s a pain for Y1 developers. For about a month I was unable to send emails with Opera, and Saf is still not supported. So what they do? They apologies and suggest me to switch to Mail Classic “just for this time”.

    For now it’s fine for me – Y!MC is a nice Web 1.0 web client, and I still can get things done..

    So if you unable to serve my browser why don’t you give me a simplified (possibly mobile) version without all you bells and whistles, Mr. Webmaster?

13. Apreche (January 31, 2008 at 9:27 am)

    I actually think the update of Firefox 3 will be faster than ever before because it provides such a dramatic improvement in performance and stability.

14. David Bloom (January 31, 2008 at 9:29 am)

    Why the higher “cost” bar for Opera 9.5?

    With Opera’s upcoming developer tools and the many bugfixes in Opera 9.5, I think it would be reasonable to say that achieving Opera 9.5 compatibility would be no harder than achieving Safari 3 or Firefox 3 compatibility.

15. Eric Perduel (January 31, 2008 at 9:31 am)

    I wonder what is this thing in Opera which makes it problematic… after all it’s perhaps the most standards-oriented browser there is.

16. Steve (January 31, 2008 at 10:16 am)

    I think the benefit for Safari will be much higher. What isn’t taken into account is the implementation of Safari for mobile devices (iPhone, Android, I think Nokia has a WebKit-powered phone browser, etc.).

    Fortunately, at least Safari 3 is the only thing we’ll have to worry about. Provided Apple brings it out of beta on the Windows side anytime soon. If there was a section on the graph for Safari 2, it’d look an awful lot like the section for IE6.

17. David Martin (January 31, 2008 at 10:27 am)

    John you’re later point regarding stuck in IE land is absolutely correct. We have clients who still require OS/2 Warp support due to legacy systems in place. IE6 had such a stranglehold on the browser industry for so long… it’s going to be a long road to eliminating so many closed systems using it.

18. wortwart (January 31, 2008 at 10:58 am)

    I disagree with you in several points.

    – Most important: What is “normal application development”? If I’m a MS addict, designing for IE will be “normal”. If I stick to web standards, Firefox, Safari and Opera will cause almost no trouble at all.

    – Where do those funny yellow bars come from? Is this anything else than a rough guess?

    – IE8 and Opera 10 might not be final by end of this year, but they can’t be ignored.

    – I don’t think it is “understandable” that common JavaScript libraries ignore Opera (which is more popular in Europe than in the US), causing problems for millions of users on many websites.

    – I think you underestimate the importance of browsers for mobiles.

19. John Resig (January 31, 2008 at 11:12 am)

    @Anonymous: I think you completely missed the last paragraph of my blog post, I’ll quote the important part here: “Now it’s not always a one-to-one trade-off in-between cost and benefit. I think it would even be safe to say that benefit is, at least, twice as important as cost. Ultimately, however, this depends upon the choices of those involved in the decision making and the skill of the developers working on the compliance.”

    @pd: I see no reason why they wouldn’t stagger the release – like they did before. I think we’ll be waiting until early 2010 before IE 8 is a dominant browser.

    @Robert: Who said anything about excluding users from a page? If you develop your web pages in a gracefully degrading way unsupported browsers will certainly get an acceptable – albeit crippled – result. There’s simply no way around this. Personally, I like supporting Opera – it’s a great browser and has a strong user base. But that’s definitely close to the exception, rather than the rule. Especially when you start considering older versions of Opera (Opera 7? 5?).

    @Andrew, David Bloom, and Eric: For clarification: both jQuery and YUI support Opera. At least in my case, I also support Opera in the applications that I develop. I based the cost percentages on the time that I’ve, personally, spent working around browser-specific issues. I’ve encountered a handful in both Firefox and Safari – and a few more in Opera. Going through the jQuery source, for example I found 3 cases of Firefox-specific changes and 7 instances of Opera-specific changes.

    @Steve: Considering that Firefox currently has an (estimated) 150 million users and, roughly, 15% market share – it’s safe to say that there’s a billion browser users. Even if every iPhone sold was actively used (4 million), that’d be, at most, .4% for WebKit (just to bring it in to perspective). While it’ll definitely help, I’ll be surprised if it pushes above 5-6%.

    @wortwart: I’m assuming that normal application development is actually following web standards to develop your web applications. I’d position the bars differently if someone ignored standards and developed IE-specific sites, but that’s more of the exception rather than the rule.
    The bars are based on my personal experience – of course they will vary for you. I just showed them as an example of how I made my predictions for next year’s browsers. If you know every IE quirk known to man then your IE bar will be much lower – if you’re stuck on an older version of Safari, then your Safari bar will be much higher. Your mileage may vary.

    Ignoring IE 8 and Opera 10 is definitely up to you. At this point it would probably cost you nothing extra to support them on top of your current development cycle (IE 8 would come free with the meta tag, Opera 10 would be pretty simple with their standards compliance). However, developing your applications for any random browser that will eventually come out can be a fools game.

    Do you have statistics on the number of active mobile browser users? This data is incredibly hard to come by – and while I do believe it’s very important (as I mentioned above) that would have to be a decision that you would have to make in your development cycle.

20. wrock (January 31, 2008 at 11:26 am)

    @John Resig: Could it be that you find more issues in Opera because you develop for other browsers first and only then deal with Opera? If you started with Opera, I bet you would find more issues with the other browsers. As for mobile users, Opera has some numbers for cumulative Opera Mini users.

    @Steve: The obvious problem here is that Nokia and Apple’s browsers are fragmented. Not the same browser. They used to be, but are diverging now AFAIK (Google will add further WebKit fragmentation). And if you take WebKit for mobile phones into account, then Opera for mobile phones should be taken into account as well (being the most widely used mobile browser and all).

21. John Resig (January 31, 2008 at 11:29 am)

    @wrock: No. Counting nothing but tangible, unresolved, bugs – Opera has more of them. This causes me to have a slightly higher development time in Opera than in Safari and Firefox – not significantly so, but slight. A large number of its bugs are related to its attempts to implement IE-specific features (like making getElementById return elements by name and id), which causes frustration.

22. Wade Harrell (January 31, 2008 at 11:32 am)

    A couple observations from consulting experience (my personal experience here, no hard and fast rules):

    1) The percentage of managers and execs using safari 2 is always MUCH greater than logs indicate the user base is, but since they sign the checks the extra time is spent to make it work; or, more likely, features are dropped. The “computer for non-computer people” problem, they never download and consider a new computer the only way to upgrade.

    2) PC based client-side developers work in firefox the most (thanks to firebug) test in IE frequently and keep safari 3 around to tell managers they tested in safari. If after getting everything 100% in those three they are forced to test in opera by an insistent client they always have a significant amount of work to still be done to get all four playing nicely. It is typically more cost effective to look for clients that do not care about opera, and not work with those that do. (I’ve had clients dropped over time lost in attempting opera support)

    I would also like to point out that people keep talking about browser support in a bubble when in the real world a developer has to factor in how much extra work is needed to get a site working in all of them at the same time. this in itself is probably the biggest thing hurting opera, and as john has pointed out their attempts to ‘unstandardize’ themselves to level the playing field is also the source of many of their bugs.

    The browserX and IE differences are mostly well documented and getting up and running with IE and one other browser is not too bad. Toss in browserY and it gets more complicated, then when you get to browserZ it really makes the head hurt. The factors in deciding which browsers to be your X, Y, and Z are many, but at this time opera usually gets the last slot.

23. Wade Harrell (January 31, 2008 at 11:45 am)

    I would also like to point out that people keep talking about browser support in a bubble when in the real world a developer has to factor in how much extra work is needed to get a site working in all of them at the same time. this in itself is probably the biggest thing hurting opera, and as john has pointed out their attempts to ‘unstandardize’ themselves to level the playing field is also the source of many of their bugs.

    The browserX and IE differences are mostly well documented and getting up and running with IE and one other browser is not too bad. Toss in browserY and it gets more complicated, then when you get to browserZ it really makes the head hurt. The factors in deciding which browsers to be your X, Y, and Z are many, but at this time opera usually gets the last slot.

24. David Bloom (January 31, 2008 at 11:57 am)

    getElementById was fixed in Opera 9.5 so it no longer returns elements by name, and a lot of the IE API support (such as document.all) is being either removed or masked (so it won’t screw up object detection).

25. John Resig (January 31, 2008 at 11:58 am)

    @David Bloom: That’s excellent news! I’m looking forward to its release! If that’s the case then the ‘cost’ of developing in Opera 9.5 would surely be decreased.

26. Andrew Dupont (January 31, 2008 at 12:00 pm)

    Hey, Prototype supports Opera now! As of… last Friday!

27. David Bloom (January 31, 2008 at 12:03 pm)

    If there’s anything we can do to get that yellow bar lower, let us know at https://bugs.opera.com/wizard/ :-)

28. Nathan Youngman (January 31, 2008 at 12:23 pm)

    The browser we use for development definitely has a big impact on the cost. Tools like Firebug mean most developers will be testing in Firefox every day, the cost is minimal. The annoyance that is “supporting IE” is made even worse by this fact… especially the longer it is left unchecked.

29. David Naylor (January 31, 2008 at 12:24 pm)

    I don’t know if anyone already said this, but don’t you think Firefox 3 will gain a pretty substantial share of the current Firefox 2 installs very soon after release? It would surprise be if the Firefox install base wasn’t at least 90 % by the end of 2008.

30. John Resig (January 31, 2008 at 1:00 pm)

    @Andrew Dupont: Yep! Hence the “up until this point” clarification ;-) But yeah, I fully support your position – glad that the leap was made!

    @David Bloom: Sure thing – I’ll pull together a couple and file some bug reports. Are you guys shooting for Opera 9.5 or heading straight for Opera 10? Regardless, since those aforementioned IE-specific bugs will be solved, I’ve re-worked the chart to have its ‘cost’ be equal to Firefox and Safari.

    @David Naylor: Yep, I mentioned that in my chart – I full expect Firefox 3 to completely overtake Firefox 2 usage by the end of this year.

31. Wade Harrell (January 31, 2008 at 1:19 pm)

    @David Bloom: what do your upgrade stats look like? do opera users jump on new versions like the firefox crowd? is so then 9.5 could be a turning point.

32. Steve (January 31, 2008 at 1:27 pm)

    @John Resig: One of the other highly variable things to watch for in 2008 is the adoption of lots of small devices which are Internet-connected and have sophisticated web browsers. The obvious example is the iPhone, but don’t forget about all the competitive products (LG Voyager, for instance) which are now starting to feature similarly powerful browser environments. These are significant for a few reasons:
    1) WAP browsers can’t get to the majority of web sites without going through WAP converters like Google
    2) Every WAP browser I’ve ever used has a simply dismal UI
    3) The UI of a web site itself is different. E.g. if I go to Digg on just about any browser, Mac/Win/Lin, it’ll look almost exactly the same. Then try through a WAP browser. It looks totally different. Normal users aren’t going to have the same “trust” for the website because it isn’t what they expect.

    The benefits provided by a “real” mobile web browser should cause normal people to use them more than earlier mobile WAP browsers. In that respect, I feel your chart of “browsers for 2009” to be off. But it’s also highly variable; it all depends on what products are released and which ones sell a lot of units quickly. Regardless of whether it’s rendered by WebKit, Gecko, Opera, or something totally different, I’m going to hedge my bet on the fact that web developers will come to hate 2008 for this reason. :)

33. mermax (January 31, 2008 at 1:48 pm)

    John, how about giving us that list of bugs you’re comparing?

34. John Resig (January 31, 2008 at 2:08 pm)

    @mermax: From the jQuery source:
    Mozilla
    – Mozilla and Safari > 2 does not include the border on offset parents however Mozilla adds the border for table or table cells.
    – Mozilla does not add the border for a parent that has overflow != visible

    Opera
    – Handle the case where IE and Opera return items by name instead of ID
    – Opera sometimes will give the wrong display answer, this fixes it, see #2037
    – We have to loop this way because IE & Opera overwrite the length expando of getElementsByTagName
    – Do a quick check for the existence of the actual ID attribute to avoid selecting by the name attribute in IE also check to insure id is a string to avoid selecting an element with the name of ‘id’ inside a form
    – An additional CSS load detection is implemented in order to verify that all CSS files are loaded on document ready (this is not necessary in Mozilla-based browsers).
    – Remove parent scroll UNLESS that parent is inline or a table to work around Opera inline/table scrollLeft/Top bug

    So final count is 2 for Mozilla-based browsers and 6 for Opera (some of either of those may be fixed in Firefox 3 or Opera 9.5 – I look forward to their release).

    Looking through the Prototype source I see 7 Opera-specific browser checks and 2 Mozilla-specific checks.

35. timothy (January 31, 2008 at 2:35 pm)

    So far for me, what works in FF usually works in Safari. IE is always a problem.

    I start in FF because of Firebug. Is there anything as good as Firebug for any other browser?

36. Steve (January 31, 2008 at 2:57 pm)

    @timothy: There’s the Safari Web Inspector in Safari 3, although I don’t think that is available in the Windows version. It also doesn’t have a debugger with breakpoints for JavaScript, which is (to me) one of the biggest reasons I use Firebug so much.

37. David Bloom (January 31, 2008 at 3:14 pm)

    “- Handle the case where IE and Opera return items by name instead of ID”
    getElementsById is fixed in Opera 9.5, known issue: getElementsByName still matches ID (Opera bug 281988) but jQuery doesn’t use getElementsByName

    “- Opera sometimes will give the wrong display answer, this fixes it, see #2037”
    Apparently fixed in Opera 9.5 – 9.5 seems to pass and 9.25 fails with provided testcase (using modified jquery.js to remove workaround)

    “- We have to loop this way because IE & Opera overwrite the length expando of getElementsByTagName”
    I’ll look into this. If you have a testcase or steps to reproduce, that would be helpful.

    “- Do a quick check for the existence of the actual ID attribute to avoid selecting by the name attribute in IE…”
    Fixed in Opera 9.5; therefore, this:
    “…also check to insure id is a string to avoid selecting an element with the name of ‘id’ inside a form”
    …is no longer necessary in Opera 9.5 :-)

    “- An additional CSS load detection is implemented in order to verify that all CSS files are loaded on document ready (this is not necessary in Mozilla-based browsers).”
    Opera bug 272870. HTML5 only defines the DOMContentLoaded event as firing anytime after parsing and before onload. We’re considering modifying our behavior to be more like Mozilla’s.

    “- Remove parent scroll UNLESS that parent is inline or a table to work around Opera inline/table scrollLeft/Top bug”
    Fixed in Opera 9.5. (also, Opera 9.5 supports getBoundingClientRect and getClientRects, and jQuery uses that instead when it is available :-D)

    So, only 2 of those issues remain in Opera 9.5 (DOMContentLoaded timing with CSS and possibly getElementsByTagName length property overwriting). That number could go down even more before the final release, too :-)

38. John Resig (January 31, 2008 at 3:54 pm)

    @David Bloom: This is fantastic news! I’m incredibly excited. I’d like to strongly recommend the two changes that you suggested – it would make life a lot easier for us developers. If there’s any way for us to put extra weight behind them to push them forward, please let me know!

39. Richard Kimber (February 1, 2008 at 6:51 am)

    I hope you’re wrong about IE6.

40. George (February 1, 2008 at 8:15 am)

    I think IE7 and FireFox will surpass IE6 already some time this year. With Microsoft allowing all users to download IE7 (started last month, being rolled out automatically this month), IE6 will start to be fully replaced by IE7 and FireFox 2/3 will surpass it.

    One one of the most trafficked sites I have access to stats for, localized in Norway, the share within IE has been the following:

    January 2008

    IE7: 54%
    IE6: 46%

    December 2007:

    IE6: 51%
    IE7: 49%

    October 2007:

    IE6: 59%
    IE7: 41%

    March 2007:

    IE6: 74%
    IE7: 26%

    In other words, a very clear trend towards adoption of IE7.

    According to W3Counter, global stats:

    October 2007:

    IE6: 69%
    IE7: 31%

    January 2008:

    IE6: 65%
    IE7: 35%

    So in Norway, IE6 has dropped 13% since October, while globally it has dropped around 4%, with the rate of change increasing the last month. If IE6 continues to drop as steadily as it has in the past 3 months, IE6 will be all but obsolete on the Norwegian market by the end of the year.

41. Upanisad (February 1, 2008 at 9:13 am)

    Opera is more popular than Safari, here in Europe as some recent surveys suggest:
    http://www.xitimonitor.com/en-US/index-1-2-3-117.html?xtor=AL-16

    (and it’s even much more popular in Australia as far as I know). USA developers are just too American-centric (that’s alway been a problem in the web world, since all major web innovations still come from the US…which is the most browser-conservative area in the world: a real paradox!),
    All my own JS libraries support Opera and I’ve never had many problems supporting it. Well, I’ve to say that Opera is the only browser I use for web sufing and Firefox, IE and Safari are just launched when I need to do debugging and testing. If my sites didn’t support Opera, basically I couldn’t even use them! :)

42. Upanisad (February 1, 2008 at 9:21 am)

    I think Opera users are just as prone to update as Firefox ones, for one simple reason: both programs warn you that a new version is available, at start up! You’re obsviously encouraged to upgrade. Why Microsoft hasn’t done the same with IE stille buggles me! I hope they do that when IE8 comes out. Windows Update is just not enough! IE itself has to tell the user to upgrade.

43. Wladimir Palant (February 1, 2008 at 11:05 am)

    I largely have to agree with your analysis. IE updates have never happened quickly before and I don’t see enough evidence supporting that it would be different this time. There are several factors stopping IE 6 from dying out:

    * There is no IE 7 for Windows 2000 – but there is lots of Windows 2000 in large companies. When it comes to private users, Windows 98 is also still in use (numbers are thankfully low by now but that depends on the region).
    * The already mentioned intranets will further slow down adoption in large companies
    * Due to the radically changed user interface some users will prefer sticking with IE 6
    * By now many users will have automatic updates activated – but a large percentage still doesn’t. These will only update when they buy a new PC.

    Altogether I see IE 6 stabilizing somewhere around 30% for a few years despite Microsoft’s efforts.

    As to Opera, I would put the cost of supporting it certainly higher than for Firefox. With Opera I regularly hit the problem that it seems to support a standard on the first glance but you start hitting subtle bugs when you go into detail. The most annoying thing is however that this is extremely inconsistent between Opera versions, including minor releases! Major Opera releases aren’t stable at all, minor releases often fix non-critical bugs and introduce others (sometimes the ones that have been fixed before). This drives the cost of continuous Opera support way up because you have to look for new issues with every new minor release. Given its low market share, my approach could only be making sure that the site doesn’t totally break in Opera and hoping that somebody tells me about the other issues.

44. David Bloom (February 1, 2008 at 12:15 pm)

    Wladimir:
    If a minor version release of Opera causes any regressions on a site, please report it at https://bugs.opera.com/wizard/ (even a bug report that just says “this site used to work and now it doesn’t” is helpful if you don’t have time to do any analysis). Any unintended regression in behavior caused by a minor version release is taken very seriously.

45. Ben Tudball (February 1, 2008 at 5:59 pm)

    @John: wortwart and Upanisad have a great point. In Australia, Opera has a greater market share than what your stats show.

    The reality is that most Websites around the world attract either a local market or a country/regional market. But your cost-benefit figures are based on either general world wide stats or USA stats.

    Opera is problematic, however. It’s, continued, low market share makes it a challenging platform to justify. It’s for this reason that major frameworks, like Prototype, have ignored working with Opera up until this point – and understandably so.

    It is this kind of thinking that hampers library adoption in other regions of the world. If authors of application libraries and open-source projects use the exact same cost-benefit figures and “common cut-off points” as you suggest, then it cuts off Opera for other regions around the world where Opera is actually above the cut-off point.

    How about spreading some more positive views about all major browsers, and quit singling out and dissing Opera?

46. John Resig (February 1, 2008 at 6:09 pm)

    @Ben: You’re missing the point of this whole discussion – it’s all about how to plan for future browsers and performing cost-benefit analysis for your own web site. I looked at my data (world-centric applications) and came to my conclusion: A large number of developers don’t support Opera in their development due to how hard it is to explicitly plan for (see above where I pointed out a number of Opera-centric bugs).

    Now, Opera has been doing some fantastic work and is rapidly decreasing their bugs – Opera 9.5 should be a stand-up release. This doesn’t change the fact that their market share is, comparatively, quite small – but it does help to justify its development cost (which will be minimal).

    I said the same for Safari 2 – it has a significant number of bugs and memory issues that make it really hard to develop for – and many choose not to (such as Google). Again, it’s all about taking these numbers into consideration and making a decision based upon the case-by-case basis of your application.

47. Ben Tudball (February 2, 2008 at 12:50 am)

    @John: You commented:
    “it’s all about how to plan for future browsers and performing cost-benefit analysis for your own web site.”
    “Again, it’s all about taking these numbers into consideration and making a decision based upon the case-by-case basis of your application.”

    Exactly. How about clearly stating these lines in your blog post and highlighting the part: for your own Website.

    You are a highly respected author of a very popular JavaScript library. As a result, you have a stronger influence on your audience than most. Some of your world-wide blog readers will use your stats and your “common cut-off point” to justify not supporting Opera. However, no doubt a number of your readers develop Websites for regions of the world where Opera has a greater market share.

    You also commented: “Now, Opera has been doing some fantastic work and is rapidly decreasing their bugs – Opera 9.5 should be a stand-up release. This doesn’t change the fact that their market share is, comparatively, quite small – but it does help to justify its development cost (which will be minimal).”

    Did you read previous my comment? Opera has a higher market share in other regions around the world, where development costs ARE justified.

    I am not requesting you to change your cost-benefit figures. You have explained your reasoning in these comments and this is your blog. But I do request you clearly indicate in your post (not just in these comments) that the data is based only on your world-centric data, and that readers should use their own data. And I also request that the “common cut-off point” phase is removed or clearly explain that it is your cut-off point, because it is actually not common at all in other regions around the world.

    Now I request this because if developers take on board your cut-off point and don’t test their sites in Opera, then their will be less and less developers finding bugs and reporting them to Opera. That is what I mean by spreading some more positive views and quit singling out and dissing Opera. You know damn well they deserve more credit.

    On a side note: in the paragraph below the graph you state:
    “Note that any browser that has a higher cost, than benefit, needs to be seriously considered for development.”

    I think you meant:
    “Note that any browser that has a higher cost, than benefit, needs to be seriously reconsidered for development.”

48. Yereth (February 7, 2008 at 5:26 am)

    I think George makes a good point.

    I found that with most of the popular websites we develop (mostly dutch websites) that IE7 already surpasses IE6 usage, which is very positive I think. Still, it will take a while for IE6 to be considered an unimportant browser in development and it depends on your target group, as some companies are refusing to upgrade even from NT.. (!!)

    But since IE6 will still have quite a marketshare, I guess I’ll have to publish my improved png fix in the meantime. :)

Comments are closed.
Comments are automatically turned off two weeks after the original post. If you have a question concerning the content of this post, please feel free to contact me.