Safari matters - support it or lose credibility

Mike Rundle has had it with web applications and websites that don’t work in Safari. From Mike’s rant Either Support Safari, Or Lose Customers:

A web team not supporting a crappy, old browser that doesn’t work well is one thing, but not supporting a modern browser like Safari is just ridiculous.

Agreed. In most cases, developers blaming bugs or Safari’s lack of support for something should blame themselves for not doing things properly, since Safari has excellent support for web standards. The comments on Mike’s post contain the usual “nobody uses Macs anyway” and “I don’t have a Mac to test on” excuses.

Well, for those of you who don’t have a Mac to test on, you really need to have a serious discussion with your boss. No web professional can get by without having a Mac to at least test their work on. Come on, you can get a Mac Mini for $499, so that is not a valid excuse. If you want to be taken seriously, you test your stuff on a Mac.

Posted on December 21, 2005 in Quicklinks, Web Standards


  1. Safari, as far as I know, respects standards very well. So I think if your pages do the same (and work well in Mozilla & Opera) Safari will have no problem to display them correctly.

  2. There’s another thing to point out. Safari is now the 3rd most user browser (behind IE and Firefox) with around a 3.5% market share.

  3. It would be really nice if all browsers were completely standards compliant but we know that this is just not reality. When there are different development groups working on browsers at the same time each one will develop its own quirks due to the very nature of the development process.

    I hope that my markup would work across all browsers but in the real world we need to know that its true. The problem I have is that I currently only 2 WinXP boxes and 3 Linux boxes running Debian. I already test my work in IE6, Opera 8, Firefox 1.x, and also Lynx. Once I find the time between classes (in school full time working on my Masters in Mathematics) and the back office responsibilities of running my business I will get the build bugs sorted out so I can test in Konqueror as a hedge for Safari. I would love to have a way to test in Safari but I currently cannot justify the cost of buying an OSX machine just for that purpose and all of my efforts to track down an emulator have failed…

  4. Maybe Safari should start supporting form labels, because most of the old crappy browsers already do.

  5. Safari supports form labels, it’s just not possible to bring focus to the nested form element by clicking on the label :)

  6. Roger, it’s really good browser but you have to admit that Safari has many weird bugs (mostly 1.2.x branch) and even the latest 2.0.2 have some annoyances when a combination of keywords and numerial values in CSS properties occur or there are still some PNG-gamma issues. Its JS support is suprising from time to time… So yes, one should support Safari, but she’d need to test extensively to avoid any of these bugs/differencies.

    Aaron, last year I used PearPC and it worked pretty well. There are also some other possibilities like OSx86, but rather illegal…

  7. December 21, 2005 by Roger Johansson (Author comment)

    Jan: I don’t agree that Safari has “many weird bugs”. A few, yes. But I very, very rarely run into any problems at all with Safari.

  8. I’ve been wanting a Mac for a long time specifically for testing in Safari, but there’s no way my boss will approve it. I will keep trying, but getting management to spend money around here is a near-impossible feat.

  9. So I am the first to remind that Safari was first publicly released browser to pass Acid2 test?

  10. Thanks for the link Roger, as I explained in that post, I wrote it as a bit of a rant after being fed up with people who say Safari support is worthless. I’ve been a big Mac user since the Performa so it’s a subject near and dear to my heart ;)

  11. BTW: The reason that entry has like 75+ comments is because somebody submitted it to Digg and it made it on the front page. Go figure lol.

  12. I use Safari for testing purposes only. I can say that bother me a lot: not meaningful JavaScript error messages. Developers sometimes just can’t catch simple bugs, especially when using object oriented JS and events.

  13. All of you that don’t have approve or budget for a Mac - use PearPC emulator on PC, as I do at home. Only thing you need except emulator is MacOS installation CDs. It’s not work very fast, but for browsers testing it’s more than enough.

  14. Acid test is theory, what about practical uses and abuses? ;-)

  15. Most places that I’ve run into will work with Safari, but have an ancient browser sniffers that look for IE and Netscape.

    For instance, the Montana State University Online sniffs out my Safari 2 and tells me I should download a newer browser such as Safari 1.3 or Internet Explorer 5.2. Grrrrrrrr.

  16. You could ever get a used iMacs Flat panel over at Small Dogs for $499. Doesn’t seem like much to pay to support a browser.

  17. There’s no need for an emulator or a $500 mini — you can get a used iMac (the first model) for under $100 on eBay. Add a copy of Tiger and some extra RAM if it needs it and you’re still looking at around half the price of a Mac mini. It won’t be good for much else, but it’ll test pages in Safari just fine. I keep a copy of 10.3 Panther on an old iMac to test older versions of Safari, Explorer, and Netscape.

  18. I totally agree that any web developer should support safari. But the argument of not having a mac to test on, at least here in Brazil, is unfortunately valid. A mac here (even the mini) costs a lot. For small agencies it may be difficult to spend that much on a Mac. A question: testing in konqueror is somewhat valid for safari, since they share the same engine? Are there many differences?

  19. I did read Mike’s article, and I think he has a valid point. However, my biggest problem with testing on Mac is this (and please, if there is a way around it, someone let me know):

    I have an older iMac - it has 10.2 on it. Which has Safari 1.2 or something. However, now that Safari is up to version 2, I can’t download or test with it unless I spend even MORE money to upgrade my OS to 10.4.

    So my issue is that I DID spend the money to buy a used Mac so I could test on Mac - but the fact that Apple wants to KEEP making me spend more and more and more to keep up? No, thanks. I’ll download a FREE browser like Firefox and test on that. As far as I’m concerned, Safari is no longer free!

    Again - if there is any way around it, I’m open to suggestions!

  20. Like Bruno I regularly test in Konqueror, and the most frustrating aspect of Safari is that in a significant number of cases it actually has rendering problems where Konqueror is fine. It’s all well and good that Apple has gotten the engine to pass Acid2, but it seems a number of other display bugs have crept in.

  21. Exactly, Bruno. I live in Argentina and a Mac Mini costs around $900. That, and keep in mind that salaries are about a fifth compared to countries like the US.

  22. December 22, 2005 by Anonymous

    Forget Credibility… Support it or LOSE SALES! I have forgone purchasing anything at Old Navy because their site specifically forbids the use of Safari.

  23. I don’t want to be a spoiler….

    I will say one thing about Safari in the fact that it supports CSS very well, but the DOM - not ‘as’ well. Unfortunately passing the Acid2 test doesn’t prove it’s a DOM-compliancy…and yes, i’ll continue to use e.preventDefault() as the simplest of all examples.

    it’s rather silly that anytime I was to prevent ordinary browser actions, i have to set its eventHandler to false with an anonymous function. Crappy dappy.

  24. Sure, the latest 2.0.x revisions are pretty fine, but those bugs left are highly unpredictable for someone who’s not aware of them and not used to avoid those constructions.

    Don’t get me wrong — I use Safari as my primary browser, but I often see colleaugues hitting these issues when developing web sites.

  25. (And sorry for the missing interpunction in my posts, I haven’t noticed you have the site in Latin1, not UTF, thus the missing dashes or ellipsis.)

  26. *coughs * ….(This is the reason why I’m hopefully getting a mac-mini as soon as I can after the New Year…and I freelance - so it’ll come out of my own pocket). Articles like this one only confirm something I knew much earlier in the year but haven’t had the cash yet to implement.


    I bought a mac to test on but then I found it was also useful for storing my photography :)

    Anyone who thinks they can get away with not supporting Safari needs to rethink their ideas.

  28. “The problem I have is that I currently only 2 WinXP boxes and 3 Linux boxes running Debian.”

    Forgive my ignorance, but wouldn’t using Konqueror provide a pretty good approximation of Safari?

  29. Hmmm, perhaps the Gap should read this article. To even think a major company doesn’t support it, is completely insane.

  30. December 22, 2005 by Paul D

    I’m developing a fairly complicated web app right now, and I find it easiest to use Safari as my test browser. No problems yet, and everything that works in Safari also seems to work in Firefox just fine.

    Quirks like focus support on form labels are easily remedied with a line or two of Javascript.

    I am worried about testing IE6 later on; I don’t have a PC either, so I could go have a pity-party like those anti-Mac commenters on Mike Rundle’s blog. However, for me, getting modern browsers working for the 1.0 release is more important.

  31. Alex, Konqueror and Safari share most of its core engine, however not all changes from KHTML are patched to WebCore and vice versa.

  32. December 22, 2005 by Maarten Leewis

    I’m looking for a new phone, so tried to find one on the Dutch nokia. Just try and get past the first page… And it’s not just on Safari, but it doesn’t work on any mac browser. I had to go in to Virtual PC to find the page behind it.

    It’s a shame they make such good phones, otherwise i wouldn’t have bought a nokia, just because i can’t view their website properly.

  33. I think Safari is quite good in supporting standards. The main problems while developing our toolkit qooxdoo is that there are many javascript related problems. The most are just not correctly trackable and so this makes it very hard to fix them. Probably it’s the best to test Safari against some Gecko Tests (I think they will do this already). But if we should really support Safari sometimes it should be optimized in the future. There are currently many more problems with Safari than in other (older) browsers like IE 5.0 under Windows.


    Exactly Ryan. Can’t hurt to say it twice. Browsercam to me is an absolute necessity.

  35. December 22, 2005 by gerben

    Wasn’t the idea of standard that we didn’t have to worry about browsers? The fact that we have to design for safari means it’s just not standards-compliant enough!

    I too don’t have the money to test on Mac so I test on IE5.5, IE6, OP7, OP8, FF107 and just hope for the best when it comes to Safari.

  36. December 22, 2005 by Roger Johansson (Author comment)

    The fact that we have to design for safari means it’s just not standards-compliant enough!

    We don’t. JavaScript may apparently be an issue (not that I’ve had any problems), but its HTML and CSS support is excellent.

  37. Aaron, last year I used PearPC and it worked pretty well. There are also some other possibilities like OSx86, but rather illegal

    Installing OS X on PearPC is also illegal, as OS X can only be installed on Apple branded hardware.

  38. The fact that we have to design for safari means it’s just not standards-compliant enough!

    You’ll need to clear up all the grey areas in the specifications, have the specifications say what the default margin, padding, etc. should be for each element…

  39. You’ll need to clear up all the grey areas in the specifications, have the specifications say what the default margin, padding, etc. should be for each element…

    You don’t. Specification is not putting you against the wall with commando waiting to shoot behind your back. It’s more of recommendation.

    I am against all those blindly followers that say web standards are saviours. Again: they’re just recommendations to be followed (not blindly) for both the browser companies and web designers/coders.

  40. Jan - Thanks for the heads up but I was already aware of PearPC. In addition to what Geoffrey mentioned I have heard about some true horror stories with PearPC “quirks”. Legality issues aside, I don’t believe Pear is stable enough for production testing yet. My opinion, I reserve the right to be completely wrong :)

    dusoft - Specifications are there to be followed and should be written as such. I will be the first to admit that the W3C are too vague. If you want a specification to be followed don’t make application developers guess about what you mean, tell them straight out. It also just might be a good idea to wait until the development community actually has time to get at least one specification working before releasing 3 additiional versions of it. It might actually let you get some actual real world testing and feedback for the next version. It’s a thought anyway…

    However, that doesn’t mean we shouldn’t follow what is there as closely as we can. Imagine if people used the logic you presented when writing banking applications. We don’t really need to worry about the exact routing number specification do we? I mean, the money will get there eventually right?

  41. December 24, 2005 by Roger Johansson (Author comment)


    What would be the problem with adhering to web standards? No, really, please explain. I don’t get your point.

    What would be the benefit of web developers and browser developers guessing instead of knowing? Would web developers benefit from having to test every single line of code they write in every possible browser? Would web browser developers benefit from having to spend even more time on algorithms for error correction? Maybe all web browsers should be required to have some kind of artificial intelligence so they can get even better at figuring out what to do with the tag soup fed to them? I think not.

    Sorry for ranting.

    Merry Christmas to those of you who celebrate it!

  42. Roger

    We can read dusoft’s comments in more than one way. I have no problems with it - maybe because I see standards as tools, not rules of law.

    Browser-defaults don’t have to line up with any standard, even though W3C have suggested some defaults. It is when we override those browser-defaults in accordance with standards that we should expect predictable results in all standard-compliant browsers.

    Browsers should not be limited to, or by, standards. All browser-developers should be able to add whatever they want as long as it doesn’t interfere with the browser’s standard-compliance.

    Additions in one or more browsers might become parts of tomorrow’s standards, and it doesn’t really matter what comes first - standards or browser-response, as long as we are going forward. That’s in part how today’s standards, and revisions to standards, are worked out anyway.

    BTW: I have no problems when it comes to supporting Safari. Sometimes Safari may have slight problems supporting me and my standard-compliant code though :-)

  43. December 25, 2005 by Roger Johansson (Author comment)

    Georg: I agree. Browser defaults and extensions/new features can to some extent be left to browser developers (but it would be excellent if all browsers could settle on a common default stylesheet).

    However, as web developers we have to be really careful about when, how and if we use proprietary features. I don’t want to go back to the browser sniffing and code forking ways of the nineties.

  44. December 25, 2005 by Roger Johansson (Author comment)

    I have forgone purchasing anything at Old Navy because their site specifically forbids the use of Safari.

    When I bought a new car a couple of years ago I was very close to buying a Ford, but since did not let Mac users in I spent my money on a Peugeot instead. So the lost sales are real.

  45. Let’s start by getting the WordPress support forums ( ) to work with Safari!

    The Username/Logins of the posts are linked to the commenters Blog - but in Safari (1.2.4 I dont know about the more recent version safari) the links do not work - I have yet to see an answer for this.

    all in all - most pages work nicely with Safari - and as old as my mac is, Safari is still faster than Firefox, unfortunately.

  46. I don’t test in Safari. Never. Primarily because it usually follows Firefox’ rendering pretty decently, but also because Macintosh OS X is so relatively small on the OS market - let alone Safari. I know, I should design for all browsers, but I don’t. And probably never will.

  47. Should you develop to support Safari, of course, do you have to? It depends.

    Let’s say, hypothetically, you have a huge site, like CNN and it was originally built w/o support for Safari - it may look decent in it, but it is not ready to say “It is ready for Safari.” If you are in this scenario, and then you analyze the statistics of the browsers that people use when viewing the site, then you can make an extremely reasonable judgement when it comes to figuring out if you should do the overhaul.

    If you fall into a similar scenario, let’s say you have a web site that is geared for Internet Explorer. It contains tips, tricks, etc. Then, by looking at your stats, you see that 95% of users are using IE, while 4% are using FF, and 1% are using Safari (with average traffic being around 20,000 a day). Is it worth the time to make it Safari compatable? The answer can vary depending on many other factors, but the answer is, probably not.

  48. Aaron, the testing was pretty effective if you don’t count the different gamma between PC and Mac. But sure, it was only a temporary thing and it’s a shame I wasn’t aware of this.

    Anyways, you can easily justify buying a Mac, belive me :)

  49. December 30, 2005 by Marc Luzietti

    You knew what Apple was before you picked it up.

  50. For all I know, Safari has some javascript issues. One of my scripts which use no browser detection and all DOM methods are reportedly not working in Safari, with no proper console errors. I am not even sure my top link navigation works in Safari actually. So Safari users would better disable javascript on my site (I am not doing any specific blocking)

    For Web Developer companies, having no mac is a sin, yes, but for freelancers it may be hard to own one. I personally say I have done enough when my stuff works in Opera, IE6, Firefox. And it is a bit of a paradox as to why I have to worry if I follow standards, I have to worry about any specific browsers, but that’s sadly the case here, speaking of javascript support, that is.

  51. Maarten: The Nokia site seems to work fine for me in Safari… maybe it was fixed recently.

  52. January 9, 2006 by Hally Stortjen

    I hate to say it but Safari’s PNG gamma problem ruins it all for me. I hope they fix it.

  53. January 9, 2006 by Roger Johansson (Author comment)

    Hally: While that may be slightly visually annoying to the designer, it doesn’t prevent a site from working. But yes, it would be nice if all browsers on a given OS would display images the same way.

  54. Some insight might be useful for future latecomers to this thread, like me.

    After some pressure from the blogosphere and a very successfull visit of Nokias mobile browser developers to this years aKademy (KDE annual conference), there is much more cooperation going on among all the KHTML developers. It was sad to see good work on KHTML not going into Safari.

    So hopefully, as the cooperation improves and the codebases converge, the differences between Konqueror and Safari will only be marginal and temporary.

    Good times ahead.

Comments are disabled for this post (read why), but if you have spotted an error or have additional info that you think should be in this post, feel free to contact me.