"Browser sniffing — unless you’re writing a web stats application — is always a bad idea. It’s a misguided attempt to send different content to different user agents. This is never scalable — you can’t change every website you’ve ever made every time a new browser version comes out. It is also not future-proof, as highlighted by this article."
I think they should just stick to /10.00 and let bad code fail. This kind of backward compatibility will only create more crud on the internet. Opera's market-share is low enough so as not to cause too much fuss (changing IE would cause too much backlash) but high enough to prompt authors of browser sniffers to act.
I don't think it is high enough to prompt authors of browser siniffers. More likely result is that, sniffers will stay the same and opera users will suffer.
The absolute last thing you want to do with your 1.5%-market-share-having browser is allow it to show people broken web pages. They're correct here that nobody is going to change their site to accommodate a dying browser. The best thing they can do is what they've done.
If you're Firefox with 25% share and a strong developer following, then MAYBE you could get away with that. Opera, treading water as fast as it possibly can, would simply die if it didn't render every existing page on the web correctly.
I'm not sure where you got the idea that Opera is a dying browser. They have a significant market share in the mobile market. As Wikipedia says (http://en.wikipedia.org/wiki/Opera_browser):
Approximately 40 million mobile phones have shipped with Opera pre-installed. Opera is the only commercial web browser available for the Nintendo DS and Wii gaming systems.
You could certainly read that to mean that Opera has doubled its market share since early 2003.
You could also read that to mean that Opera has remained at about 2% market share for six years now, during which time Firefox has been born and risen to about 40% penetration, and during the last six months of which Google's browser has been both released and passed them by a factor of 3.
So yes, Opera is not going anywhere. You can read that to mean that they're going to stick around if you'd like. You can also read that to mean that they are not now, nor have they ever been a real competitor in the Browser market.
I don't know if those statistics are really that accurate. They are often biased towards the desktop/laptop, & often the USA/English as well. Also from that Wikipedia article:
As of September 2008, usage data on English-language sites show Opera's share of the browser market as being below 1%. The browser has seen more success in Europe, including about 18–20% market share in Russia and Ukraine,and 5–6% in Poland, Latvia, Lithuania, and the Czech Republic.
Unlike any of the other browsers listed in those statistics, Opera is developed by a company based in a non-English language country. They have their own niche languages/platforms and I'm sure they're doing quite well in them.
That said, I live in the US and find Opera to be entirely usable on my laptop. In fact, I asked on another thread (http://news.ycombinator.com/item?id=622038) why Opera doesn't have a larger market share, and most of the reasons given were non-technical.
I always liked Opera's philosophy of "do what IE does" rather than Firefox's policy of interpreting the more ambiguous parts of the standards in a way that conflicts with IE's pre-existing way of rendering. It meant that pages just plain worked first try on Opera no matter which browser you targeted initially. That definitely wins me over.
That being said, Opera is the only modern web browser that we no-longer test for on Twiddla. For a while we were putting in little Opera-specific hacks to work around its painfully slow rendering engine (which ironically is the reason it's such a pleasant browser to develop against), but eventually it became too much effort to be worth it considering the 1.35% of our traffic that it comprises.
So yeah, I'm with you in hoping it doesn't go away. Still, I'll stand by my statement that breaking the user experience to make a point would be a fatal mistake for them.
>> "I always liked Opera's philosophy of "do what IE does" rather than Firefox's policy of interpreting the more ambiguous parts of the standards in a way that conflicts with IE's pre-existing way of rendering."
So Opera duplicates IE's idiotic bug ridden excuse for rendering, whilst Firefox does things properly. I'd rather firefox/webkit TBH.
IE is and always has been the standard. It's only in the last few year that Firefox has captured enough market share that the W3C "standards" have any real bearing on the situation.
Of course, had the W3C simply used the box model in place in IE and Netscape at the time it wrote its CSS specification, rather than inventing its own, nobody would be calling the IE way of doing things idiotic. It would simply be the logical way of rendering content in boxes.
For a while we were putting in little Opera-specific hacks to work around its painfully slow rendering engine (which ironically is the reason it's such a pleasant browser to develop against)
I've only used Opera since it became free in version 9, but it's always seemed a lot faster than Firefox and IE. It's even gotten noticeably faster upon all the significant updates I've seen.
It's surprising to see you describe Opera as IE-like (broken) and slow, but since you probably have more experience in these things, I'll have to take your word for it.
I actually thought of the same. The trouble is that most enterprise application vendors will not be so willing to change their application. And this in turn will reduce the rate of adoption of the new browser in an environment already very slow in adoption rates
How do you propose that would work? The problem is that scripts detect Opera as version 1, not 10, and then failing - "if digit 1 of version < 4 then fail". To the browser, the looks like (and IS) the expected behaviour, so there no not-horribly-broken automatic fallback.
In a way this solution is compatibility mode. "Broken" detection scripts see version nine (which is good enough), scripts that know about Opera 10 sees version 10.
Think about the people who have to deal with the UA string for legitimate reasons (analytics...) they will have to deal with a special case (if (UA == "Opera/9.80...")).
IE compatibility mode is broken. It has two "modes", one of which changes the user agent version and one that doesn't (but still changes the functionality of the browser). Guess where the sniffer scripts get the version from.
The only way to make it work reliably is to look at the window.documentMode property which requires changing your code, thus defeating the point of the compatibility mode in the first place.
"Browser sniffing — unless you’re writing a web stats application — is always a bad idea. It’s a misguided attempt to send different content to different user agents. This is never scalable — you can’t change every website you’ve ever made every time a new browser version comes out. It is also not future-proof, as highlighted by this article."