It's not every day a whole new browser comes out. It's even less common for that new browser to make lead story status in mainstream media outlets*. But today both happened as Google released their browser, Chrome.
[Or, When Minutiae Attack!]
Update 2011.10.17: a quick test in Firefox 6 and 7 suggests they've finally fixed this issue. With the suggestion that FF 3.6 will reach EOL in the near future, we might finally be able to say goodbye to this curious little problem.
There's a perennial question that pops up on email lists and from other developers. The question goes something like this:
"This page works fine in everything except Firefox and I can't tell why... it's showing an HTML comment as raw code..."
It seems that Firefox 3 will include an option to treat meta refreshes much the same way as popups - blocking them and alerting the user what the page wants to do. It's another step forwards in letting the user take control.
Of course, Opera users already have this option; using opera:config#UserPrefs|ClientRefresh. Neat, although an alert would be good; as would site-specific settings. Hopefully the feature will be refined in future versions.
Really though, either way is good as it gives the user a little more control over their browser. Automatic refreshes and redirects break accessibility recommendations. They're one of those things which gets written up as "until browsers provide a way to control...".
As these features become more widespread, the importance of fallback options will become even more critical. Just like scripts need a <noscript>, meta refreshes need a link in the document. Many pages don't have them, though; so no accessibility or SEO juice for them!
It serves as a good reminder that we should provide alternatives any time we modify the behaviour of a page. I have had people say in the past that meta refresh was so simple nothing could go wrong. Well, that assumption will bite them on the arse...! :)
We should always assume that somehow, somewhere such features will be disabled. It's not hard to provide an alternative, so it should remain our habit to do so.
Back in 2005 I compared the patch rates of IE, Firefox and Opera. In the past few days the subject of browser security has come up a few times, so I thought I'd revisit the topic to see what (if anything) has changed.
I'm using Secunia advisories again, to keep the data source consistent. The product pages are:
Note that Secunia's data starts from February 2003, regardless of each product's release date. You can investigate Secunia's methodologies if you will; there are some quirks. However I'm not after a perfect scientific investigation, so much as a broad strokes impression.
Since each browser has a different release date and lifespan, comparing raw numbers of problems isn't really useful. However we can compare the percentage of patches/fixes from the vendor - it's not how many security issues were identified, rather it's about how many were fixed.
I would have added in "time to patch" and "days vulnerable" and so on, but Secunia doesn't currently graph that information (as far as I know).
I thought about sorting out standardised timeframes and so on, but the bottom line here is how secure can a user's browser be today? I say "can" since we can't assume that all browsers are up to date with the latest patch (or even close), but we can at least evaluate the potential for a conscientious user to keep up. After all, we can only apply the patches that are available.
Having discussed the user acceptance issue in the previous article (to patch or not to patch?) I won't rehash it here. However I will mention that according to Secunia Opera users really need to update their browsers.
First, let's remind ourselves of the data from 2005:
| Browser | IE 6 | Firefox 1 | Opera 7/8** |
|---|---|---|---|
| Number of advisories since Feb 2003* | 83 | 21 | 42 |
| Vendor patch | 55% | 81% | 100% |
| Vendor workaround | 1% | 0 | 0 |
| Partial fix | 13% | 5% | 0 |
| Unpatched | 30% | 14% | 0 |
* Firefox advisories start from August 2004.
** Opera 7 and 8 are combined to create a better comparison in terms of the number of advisories.
[Note - yes I know it didn't really make sense to combine Opera 7 and 8, but both had a 100% success rate so it didn't really change the outcome.]
First off, let's compare the patch rates of the same browsers (and we'll add Safari so people don't accuse me of forgetting Macs). Remember that these are all superceded versions now:
| Browser | IE6 | FF1 | Op8 | Saf1 |
|---|---|---|---|---|
| Number of advisories since Feb 2003 | 110 | 39 | 15 | 15 |
| Vendor patch | 67% | 87% | 100% | 93% |
| Vendor workaround | 2% | 0 | 0 | 0 |
| Partial fix | 11% | 3% | 0 | 0 |
| Unpatched | 20% | 10% | 0 | 7% |
So, no change for the three browsers compared last time. Safari slots in at second, after Opera and before Firefox.
Now let's have a look at the latest versions of the four browsers:
| Browser | IE7 | FF2 | Op9 | Saf2 |
|---|---|---|---|---|
| Number of advisories since Feb 2003 | 4 | 2 | 3 | 6 |
| Vendor patch | 25% | 50% | 100% | 33% |
| Vendor workaround | 0 | 0 | 0 | 0 |
| Partial fix | 0 | 0 | 0 | 0 |
| Unpatched | 75% | 50% | 0 | 67% |
This produces very clear results, but the low number of advisories exaggerates the margins. The previous versions all have a higher number of advisories, but actually the only change in ranking is that Safari drops from second to third. The sharp drop in patch rate between Safari 1.x and 2.x makes it hard to get any useful conclusions - has Apple really dropped the ball?
For the other three browsers, the rankings remain:
It's worth noting that the patch rate for both Firefox 1.x and IE 6.x improved between 2005 and 2007. However both dropped noticeably between their previous and current versions (same as Safari). The proportion is exaggerated by the low number of advisories for the newest products.
Well, one clear thing is that Opera is the only vendor with a 100% patch record according to Secunia. Opera is also the only vendor that maintained its patch rate between versions - in fact you have to go back to Opera 6 to find an unpatched advisory (and there's only one).
It's also clear that IE has the worst patch rate of all the browsers compared. You could say that's a result of having a much bigger user base and a correspondingly higher incident rate. But then Microsoft has more resources than the other three vendors combined so it's a pretty weak excuse for leaving security issues unpatched.
Meanwhile Firefox does pretty well for an open-source product, consistently beating IE - even if not by much. Apple meanwhile needs to get Safari 2 sorted out; but we'll see what happens as more data becomes available (for all four browsers).
So at this time Opera wins the patch stakes. The argument can be made that Opera attracts fewer attacks due to small marketshare. That could be true - there's no way to truly know, since malicious hackers aren't polled - but when I'm doing my banking I don't care if it's true. I just care that my browser is secure; and Opera currently has the best record for fixing security issues.
Computer lab only has IE? No problem! Bring your own browser. Most labs let you use USB drives, so use them to run software...
So what software do you use on your thumb drive?
This feeling has been brewing for a while: Firefox is turning into Internet Explorer. The attitudes surrounding Firefox draw ever closer to those attitudes it was supposed to destroy. Things people hate about IE and Microsoft are appearing in Firefox and the Mozilla Project.
What am I talking about? It mostly boils down to these points:
People weren't happy to let IE get away with that crap, so why are they taking it from Firefox?
I thought we were supposed to be getting away from the days where users with a small marketshare were told "too bad". Sadly an increasing number of sites are appearing which work in IE and Firefox, but not Opera or Safari. Users of these browsers can rant and rave but they are inreasingly being met with blank stares - why won't you just use Firefox and shut up?
Major players lament having to support "niche browsers" and even people who should know better are starting to go live with sites that only work in IE and Firefox:
AOL users, as well as those unaccounted for here (Netscape, Opera, etc.), would be wise to use Mozilla Firefox in order to have the best possible experience using Blogger.
What makes it even more interesting is the way many applications add support for other browsers later on. This indicates that it was entirely possible to support all browsers at go-live, but instead they ran early without bothering to finalise the product. Who cares about a few idiots who don't use IE or Firefox?
Most of this is based on the flawed assumption that browser stats are gospel. For all anybody really knows, Firefox's real marketshare could be a third of that reported in most log files (what with all the pre-fetching).
People ranted at length about being told to "go use IE", yet don't seem to blink when directed to Firefox. Just because Firefox has some open source cool points doesn't mean people should be forced to use it.
What happened to letting the user choose? Did we decide that Microsoft was right after all - everyone should use the exact same software?
Worse still, many people seem to think Firefox is the only alternative browser. Firefox has been pushed so hard, people are treating it like "the other browser" instead of "another browser". Gratingly, articles are appearing with titles like Why You Should Consider Budgeting a Site Redesign for Firefox 1.5 Now (Yes Firefox). Argh! No! Redesign with standards, not "for Firefox"!
There's been quite a bit of noise surrounding the <a ping> feature being proposed for Firefox. Actually, it's not a feature for Firefox so much as a bit of proprietary code for developers to add to their pages. Wow, there's a thought - let's code our pages for just one browser. Let's stick "Best viewed in Firefox!" buttons on our sites and get in the popcorn to watch Browser Wars 2.0 unfold.
Does one feature really mark an entire project? Not really, but have you ever had the feeling you've just seen the thin end of the wedge? The ping attribute isn't even being proposed for a good reason - the justification is that some sites do stupid redirection monkey business in order to track hits. Why a browser maker should get involved is beyond me. Let those sites break if their redirections fail, for all I care. Don't make up new code!
No matter what people say, Firefox is not perfect. It has bugs (table padding problems, anyone?), it has security flaws, it has a memory problem. It may have a larger marketshare than a few other browsers, but it's still a minnow compared with IE.
People should stop acting like it's perfect. Firefox is just another choice. Its marketshare could slip tomorrow - IE7 could take back everything Firefox gained; taking the "it's popular" argument back off Firefox. I'm sure Firefox users would still like to be supported even if that happened.
Unless we want to replace IE with Firefox, the industry needs to remember the original point of web standards: support standards, not browsers. Build once, publish anywhere, let the users choose the UA that's right for them.