Opera 11 - new and improved (Alpha) | Hacker News Reader

Opera 11 - new and improved (Alpha)

my.opera.com

56 pointsyread63 comments

I guess you have to credit Opera for sticking it out this long, years after they dropped off the radar of basically everybody. I just checked for a few of my sites, and I'm seeing about 1.3% of my visitors using Opera. I keep hearing that their mobile browser is big, but it seems that nobody has ever visited any of my sites using it.

The one thing I always liked about Opera as a developer is that they used IE as their spec. That is, if a site rendered correctly in IE, chances are it would do the same in Opera. The upside is that you'd only ever need cross-check for Opera compatibility a couple times a month (versus several times per day for IE/FF/Chrome), and it would generally still be working.

The downside of course, since they have such a small installed base, is that if ever your thing stopped working on Opera, you didn't have much incentive to spend the time and effort do do anything about it.

Opera has decent market share in central and eastern Europe.

In Ukraine and Russia it's huge:

http://clear.com.ua/projects/ee_browsers/?ukraine

Wow! Thanks for that. I really need to break the habit of assuming the world follows US patterns because Google, Silicon Valley, etc are in the US.

Also Germany, Scandinavia and parts of Japan I have found uses this browser. Europe more in percent than America. I am not sure why this is. If any one knows or guesses, I would be curious to find out.

For one thing, it's a Norwegian company, which explains its popularity in Scandinavia, and could mean more focus and expertise on internationalization. Second, some of its features are more suited to Eastern European markets. Computers tend to be older and cheaper, and Opera works well in low-CPU/low-memory environments[1]. Many connections are still on dial-up, and many ISPs charge by the megabyte, and Opera has a number of options for saving bandwidth, like disabling images or using a proxy compression server. These advantages are probably less important nowadays than 5 or 10 years ago, but by now it's an established player, which helps.

[1] On machines with lots of RAM it tends to be a memory hog, but it works perfectly fine in tighter environments. They claim their memory strategy changes depending on local conditions.

> I guess you have to credit Opera for sticking it out this long, years after they dropped off the radar of basically everybody.

I don't get it. They were never on the radar. They are now. They've got more than 140 million users.

> I keep hearing that their mobile browser is big, but it seems that nobody has ever visited any of my sites using it.

It has 1/4 of the mobile browser market, and mobile browsing is expected to be bigger than PCs in a couple of years (it's actually growing that fast).

Opera exists on the basis of the superb mobile browsers - Opera Mobile and _especially_ Opera Mini. (there was a previous post about this http://news.ycombinator.com/item?id=1882753)

In India, where smartphone penetration has just started, pretty much Opera Mini is defacto here. Especially with the cheap, underpowered Chinese smartphones which come pre-installed with Opera Mini.

In fact, a significant percentage of Nokia phones out here are preinstalled with Opera Mini (or Opera Mobile). I'm pretty sure there is a per-device licensing that is making money for them.

More than 1/3 of Opera's revenue is from the desktop browser. It's simply false to claim that it exists on the basis of mobile browsers (superb or not).

I am a Mac user and I have a new wish: browser vendors releasing alpha and beta versions named as such: so instead of Opera I would have Opera11a in .dmg. Then I would simply drag it into my Application. Now I drag it on desktop, rename, and drag into Applications…

They've previously done that with Opera alphas and betas. I don't know if they're doing it consistently, though.

Still waiting for news: Opera xx - now we support google services!

I dont know who downvoted you, but as an Opera user that uses Firefox to check Gmail, I'm waiting for the same thing myself!

Huh? I don't have any problems with Gmail under Opera. As for the "who downvoted" I guess it was someone who thinks that it's Google that should be blamed for most of "the problems" with Google's services under Opera. It's a browser that's very standard compliant and yet they don't even let you do such silly thing as changing background image on Google's front page (and of course it works fine if only you mask Opera as Firefox).

The last two builds broke gmail for me. However, the latest one works.

Could it be possible to have emails time tag to be correct?

Let me disagree here. Although Opera is positioned itself as standard compliant, it has many issues with javascript. Some functions are working incorrectly, some just not working, some events firing, some not. Right click is disabled too.. I'm not against Opera, but we should think about stuffs above. Regarding Google - google making their money from adverts, from people who using google. Do you really think that google could deliberately disable Opera support due to some internal polytics and renounce from money of many opera users? I think there were good reasons such as bad js support, because javascript is widely used by google in their services.

Your argumentation ignores that Google had to invest money to test for another browser, which may be more than the money they can make on adverts to Opera users. This has nothing to do with any supposed problems in Operas js handling (which I've never noted ... for me the js works just fine in Opera and disallowing js to handle the right mouse button is a great feature), but is a simple cost-benefit analysis.

Google preemptively tries to kick Opera out of the browser world. Many new services come up with dedicated code to block Opera claiming that this browser can't handle them. I can't tell you how many times I changed Opera's UA string to Firefox and those services start working. Magic! <sarcasm

Not testing for a particular browser is fine (and truth be told Opera just works in 99.99% of the time). Blocking it just because one can is pure evil!

Of course, Opera is the only browser that has challenged Chrome claims of the faster browser ever (in production builds, that's it).

> Blocking it just because one can is pure evil!

It means they value market share over standards. Very evil indeed.

> Do you really think that google could deliberately disable Opera support due to some internal polytics and renounce from money of many opera users?

Yes. Google Docs officially does not support Opera. Not Gmail either. This was a policy change that happened some time ago.

Sources: http://docs.google.com/support/bin/answer.py?hl=en&answe... http://mail.google.com/support/bin/answer.py?hl=en&answe...

In my experience, both of these always work in Opera though, if you actually try.

(disclaimer: Opera Employee, Core QA)

If you are having problems with javascript in Opera, I would really like to hear about them so that they can be fixed. There is a bug report form at [1], or reply here. Obviously it is nice if you have a reduced testcase, but just knowing that a site fails is immensely useful.

[1] https://bugs.opera.com/wizard/

http://groups.google.com/group/google-maps-js-api-v3/browse_...

http://www.questionhub.com/StackOverflow/3899985

Aggressive opera caching and different behavior from others browsers is not what I like.

Why the questionhub link? http://stackoverflow.com/questions/3899985/opera-error-uncau...

http://www.questionhub.com/StackOverflow/3899985

This is a known issue. No promises, but it looks like a fix might be ready sooner rather than later.

http://groups.google.com/group/google-maps-js-api-v3/browse_...

It's not instantly obvious to me what the problem is here, so I will need to investigate further.

Thanks for the bug reports!

> Right click is disabled too

It works in zimbra.

> Although Opera is positioned itself as standard compliant, it has many issues with javascript.

I would have to disagree. Some people keep claiming that, but they can never seem to provide any evidence that Opera has more problems than other browsers.

GMail failed in a number of the Opera 11 alpha builds and snapshots because our native implementation of Function.prototype.bind was broken. This was recently fixed.

Google can't even get Chrome to work on all their own services. How can you expect them to support Opera?

That address field feature is a cool idea. I'd expect other browsers "copying" it soon, which would make it another item on a list entitled "Invented by Opera, popularized by others".

The way Opera hides http:// is a nice touch (edit: because they avoided copy&paste problem that Chrome has), but such addressbar style has been pioneered by Firefox (extensions) and Chrome.

It's not. It breaks copy and paste.

Have you tried it? it doesn't.. when the focus is on the address bar it reverts to the whole URL.

Then they fixed it since last month.

This feature didn't exist last month -- it was added publicly on november 17th, in other words two days ago.

Dunno about Chrome but I don't have the copy and paste problem in Chromium 9.0.584.0 (66224) Ubuntu 10.10. Since Chrome is released from the same base, either it should already be there or would be there in the future releases.

The problem people complain about is that when you try to copy domain name alone (e.g. to ssh to it) http:// gets copied unexpectedly. It's a feature that for some is a bug.

Opera avoided the issue by showing full URL when you focus addressbar, so you see what you're copying and can edit protocol part.

I don't understand what you find "cool": not showing protocol, not showing query parameters, or both?

> Parameters at the end of web addresses are visually hidden when the address field is not focused

I think those are terrible, terrible ideas; to the best of my knowledge this behavior was pioneered by Chrome and is now growing like a cancer, under the false premise that "non-technical users don't understand urls".

How are they going to ever understand them if browsers don't show it to them??!? Hiding query parameters breaks the relation between a url and its content; now all of Google is under the same "www.google.com" url: how is that "better"?

By the same principle that "young people don't understand words" let's remove all printed words and simply print books with the same cover and the same title -- they're all "books" anyway.

This feature evokes unpleasant memories of Microsofts decision to hide file extensions and all the pain which came because of this. The only good news is that you can deactivate the feature in your preferences.

Too bad you can't do that in Chrome.

I keep a patched version of Chromium around for that very reason. All you have to do is change one line in the sources. I'll dig up the patch and send it to you once I get back on a real computer.

I don't like that they hide query string, but they've implemented hiding of protocol well (doesn't interfere with copy&paste).

I've realized that hiding of https:// actually makes sense. You could mistakenly believe that page delivered over https:// is secure, but there are several cases, when it's not true (e.g. page embedding JavaScript over plain HTTP).

A true security indicator replacing https:// in the addressbar makes more sense, and it shows secure only when connection is really secure (as far as SSL is concerned of course).

Actually I wasn't referring to the fact that they hide protocol, that is indeed questionable move. What it's cool it's how they present you an information about security level of the site - check it.

Edit: plus if this http://news.ycombinator.com/item?id=1921100 is true (don't have Opera 11 here to check it) then hiding protocol it's hardly a problem.

I'm sympathetic to the argument that hiding information is counter-educational. But nothing exists in a vacuum. There are serious problems to deal with, too. I suspect the concern is phishing. Consider:

  hxxp://mc564.mail.yahooo.cz/mc/welcome?.gx=1&.tm=1290173262&.rand=8a8hcjme2nb60#_pg=showMessage& sMid=13&&filterBy=&.rand=1020435964&midIndex=13&mid=1_41637_AFNv%2FNgAARb0TMWfVwycxVDDYoo&fromId=sendreceive@iso50.com&m=1_45237_AFRv%2FNgAALZZTMxQEQbt1QDuB60,1_44482_AFRv%2FNgAAX6dTMscRg7DzmBp0%2BM,1_43847_AFRv%2FNgAAWU6TMoFPgz8kG1tlus,1_43135_AFVv%2FNgAAPHWTMixWAJSDRUfECQ,1_42368_AFNv%2FNgAAEUnTMcCHAZk2EsgYvc,1_41637_AFNv%2FNgAARb0TMWfVwycxVDDYoo,1_40918_AFdv%2FNgAAOo%2FTMMVTQtEN0nYhxw,1_40277_AFdv%2FNgAAK2ETMGqmwMZ%2Fm865vA,1_39560_AFdv%2FNgAAJLQTMBzaQzIs28a2Uc,1_38916_AFJv%2FNgAAGfRTL97CgGm0yp6dBs,1_38170_AFZv%2FNgAAAK%2BTL4DxwdptRQfn90,&sort=date&order=down&startMid=0&hash=30c49f5b9f61273beed1953ae7674c2f&.jsrand=5917415

How likely are you to notice the problems with this URL? I'm pretty sure I wouldn't. I wouldn't be surprised to find out that non-technical users consider the address field a sort of write-only space; they type something comprehensible in there, and the browser destroys it. The simplified address field with the bold domain name is a lot easier to read, and that probably increases the chances that it will be read, rather than ignored. The full URL is just one click away.

Enter the URL with two spaces before it please, to place it in a code box.

It's changing the layout.

Sorry about that. But that's interesting, isn't it? The URL is so unwieldy that the site and/or the typical web browser cannot handle it properly. In fact, it's so long that most of it is already invisible on a typical screen anyway.

All modern browsers deal with phising without destroying information; just alert the user that the site they're visiting is not what it claims to be.

If users don't look at the address bar and don't understand urls, how are they going to tell something is amiss when they visit the following domains (all available at the time of writing):

- gnail.fr - gnail.me - gmail.se.com - gmail.uk.net - yahooo.uk.com - yahooo.eu.com

Only showing the domain does not help the non-technical user: you can have a fake domain look quite legitimate to the non-technical eye.

Hiding part of the URL is what every browser currently does, when the URL gets long enough. Does that count as "destroying information?" The Opera alpha shows the whole URL when you click on the address field.

I think I'd be more likely to notice typos like "yahooo" or "gnail" with the new display method, without the noise of a thousand-character URL surrounding them. That is by no means the be-all and end-all of anti-phishing technology. It's just giving me one more chance to see something that might be amiss.

Why should people have to understand URLs? Do most people understand how the engine of a car works? No. They just want to get from A to B. Most people don't care.

Hiding query parameters hides yet another thing that's mostly unreadable anyway.

Comparing this to "young people don't understand words" is quite silly. Printed words are important. The message is important. And in fact, language is constantly being simplified, and has been simplified throughout history. Just like the address bar in Opera is now simplified.

I was excited to see they decided to add extension support. Unfortunately it seems quite limited for now. Essentially you have greasemonkey scripts plus a button on the toolbar. Also the extension API is extremely buggy at the moment, which makes it painful to write extension but its still a development build so it's ok.

Looking at the changelogs I'm surprised to see regressions in basic features of the browsers. I wonder if the firefox and chrome trunks have the same kinds of bugs.

title should be clear this build is a nightly build, not retail

Scores only 179 on html5test.com. Chrome dev: 241, Webkit nightly: 243, Firefox 4 beta: 217

HTML5Test doesn't pay too much attention to stuff like SVG, which is Opera's strong point. One area where opera is lagging behind is webGL support and that reflects on the tests.

SVG is not HTML5…

That site isn't very useful since it only reflects what the author thinks is important. It doesn't show actual standards compliance. He chooses tests and assigns scores according to his own priorities. So it's misleading to call it "html5test."

Scores are based on the number of features supported. What's wrong about it?

What's wrong is that one feature gets one score, and another feature gets a much higher score. Scores are not based on the number of supported features, but on higher scores for certain features.

What are you talking about? One feature is one point.

I really like the way mouse gestures are shown and "taught" - feels almost like a teaching game. But I had to disable the URL parameter hiding..

Immediately used 160MB upon starting the program. Firefox is lean compared to that.

A random number like that is pretty useless. Opera has dynamic handling of RAM, so if you have a lot of it, it will use more. If you have less, it will use less.

I'm guessing you have at least 1-2 GB RAM.

By the way, claiming that Firefox is lean because it sets hard limits instead of adapting the memory usage dynamically is pretty silly...

You can manually set the memory cache in Preferences.

I figure that unused RAM is wasted RAM. If I've got 4GB of RAM and all that caching allows to browse faster (going back a page in Opera is lightening fast), then I applaud it.

I like getting downvoted for not knowing the minutia of a program's memory handling. Thanks, people.

To Osiris: Given I'm not aware of that 150+ MB being dynamic, I assume that's 150MB I'm not getting back if and when I need it. I play memory-eating video games and use VMs; 150MB for a web browser looks like a lot when you've got to trim the fat off your running processes.

Does Opera reduce its memory footprint when you are low on RAM?

Reply on news.ycombinator.com