A Blog From Behind the Trenches

Time to say goodbye

Today, we announced that Opera has reached 300 million active users. At the same time, we made the official announcement that Opera will move from Presto to WebKit as the engine at the core of the browser.

With this, Opera will be the first major browser to switch to a completely new rendering engine.

Read more...

About a month ago, I posted a request for people to list their top 3 problems with Opera. I received more than 250 responses, and I'd like to thank all of you for your comments.

Read more...

Another "cloud browser" seems to have closed its doors.

WAP Review reported last week that the "Bolt" proxy browser service has been shut down due to "economic circumstances". The browser has been pulled from Android Market, and the browser's official site has been replaced by a brief text explaining the situation. Some of you may recall that I called out BitStream, the creators of Bolt, on their outlandish speed claims a couple of years ago.

There used to be a wider selection of "cloud browsers", but the list is now narrowing down considerably:

• Skyfire shut down its worldwide proxy browsing service to focus on Western markets, and today it mostly seems to work as a video/media transcoding service on top of the system default browser
• TeaShark created some initial headlines a few years ago, but apparently haven't been heard from since

The reason for this mostly seems to be that running these services is not exactly cheap, which we can definitely relate to. While Opera Mini is currently bringing in a lot of revenue, the expenses are also something we have to take into consideration. We've had to focus a lot on building efficient infrastructure, seeing as we're processing more than 86 billion pages every month!

And let's not forget that this is a highly competitive market.

Now Bolt is closing down, or at least the free service is.

But who knows, maybe they have some good engineers that would like to continue working on browsers. Remember: We're hiring!

A couple of years ago, I reported on how Apple was using patents to block a W3C specification.

The end-result was that the patent didn't seem to be relevant to the specification at all, and one or both of them were even rejected by the patent office. That Apple would use invalid or irrelevant patents or patent applications to block or delay an open standard seems odd, but if you look at their general behavior during the whole thing, it is easy to conclude that the intent was indeed malicious.

And while I didn't report on this (I suppose I should have), Apple actually used two patent applications to prevent a W3C standard from proceeding in 2010. This patent claim, too, seemed to have been filed at the very last minute, much like the patent claim in 2009.

As I mentioned, the submissions in question were still only patent applications, and Apple were basically saying that the claim covers not only the patent applications as they were then, but also "any claims that issue in any continuations, divisionals, continuations-in-part, reissues or other counterparts" in future versions of the applications.

Not exactly helpful.

And guess what, Apple is at it again. Another year, another attempt by Apple to block open standards using patents.

This time they have four claims - three patents and one patent application - that threaten to block the W3C Touch Events Specification. They filed their patent claims a little over a month before the time limit expired (the claim was filed on November 11, and the time limit is December 26, 2011).

The odd thing is that Apple chose not to join the working group that handles touch events. If they had joined, they would have been forced to file the patent claims far sooner. So now we know why they didn't join. What we don't know is why Apple insists on waiting almost until the last minute before filing its patent claims.

I'm starting to see a pattern here:

• 2009: Apple discloses one at the last minute. Patent Advisory Group created. Patent was found not to be relevant.
• 2010: Apple discloses two patent applications at the last minute. PAG created. One patent found to not apply. The other deemed not relevant.
• 2011: Apple discloses three patents and one application at the last minute. PAG likely to be created. Unclear what the result will be.

What makes this matter even worse is that this doesn't just affect these specific standards. The Patent Advisory Groups could in fact slow down the development of other standards by pulling people from other projects in order to investigate these claims. The investigation can take several months, and will take time, resources and money to complete.

That's time, resources and money that could have been spent on improving various other work-in-progress standards.

If you take a look at Dev.Opera, Opera's resource for developers, you will notice that the site now has a brand new design (including inline comments rather than links to a separate forum thread). The developer relations team hasn't posted any articles on the redesign yet, but you can reach them on Twitter for now if you have any feedback on the new site.

Today we released a new Opera 12 "Wahoo" snapshot with full ECMASCript (JavaScript) 5.1 support.

Now that we've added proper ES 5.1 support, including things like strict mode, the test262 test suite looks much better in Opera. However, you may notice that one of the tests fails. As it happens, this particular test seems to be invalid, so it is supposed to fail in browsers with a proper implementation.

And this highlights a point I made a couple of months ago when yet another "benchmark" from Microsoft's marketing department was making the rounds on the web:

The test262 test suite still isn't finished, and there may be faulty or invalid test cases in addition to all the test cases that have yet to be written.

It might be tempting to portray the test262 test suite results as the final word on ECMAScript 5.1 compatibility, but that is far from the case. In fact, there may be other faulty tests in there.

I hope all Opera users out there will be careful not to misrepresent the test results!

In my opinion, the main focus when discussing ECMAScript 5.1 and test262 is how great this is for interoperability. With all the major browsers implementing the same standard and aiming to conform to the official test suite, interoperability on the web will be be greatly enhanced.

So don't get me wrong. If you had the impression that test262 is of no use, this is far from the truth. It will be a great tool to aid interoperability as it develops further, and more tests are added and invalid tests are fixed or removed.

It's just that it goes beyond mere marketing. It's going to be an important tool to keep the web open to all.

People seem to believe that a video uploaded five days ago with the title "Wybuch w Oslo - Explosion in Oslo - 22.07.2011" predicted the explosion. However, Google reveals that the original title was actually "SZOKUJACY FILM nawałnica burza nad Polską 20.07.2011".

Here is a screenshot showing the original title, as cached by Google:



Click the image to see more of the search result page.

The evidence clearly seems to show that it is fake.

It's that time of the year again.

NSS Labs publishes yet another report on socially-engineered malware, and Microsoft's browser incidentally wins it with a huge margin again. Strange, isn't it?

There's no need for me to reiterate all the problems with the report in this post. My post on the subject from back in 2009 is still mostly valid, so feel free to take a look.

The report is still equally unverifiable, poorly put together, and seemingly "tampered" with to boost IE's score. Does anyone really believe that IE is not only better, but far superior to any of the other browsers?

If they do, they really should consider the major issues with the report.

Update 2011-07-21: An anonymous/undercover NSS Labs employee denied that the report was sponsored by Microsoft. Although they have a history of posting misleading claims, I edited the post to avoid it taking away the focus from the many problems of the report itself, such as the poor methodology.

I notice that yet another "benchmark" blog post from Microsoft is making the rounds on the internet.

This time they are bragging about having the highest score for JavaScript test suite "test262".

As one might expect, the impression people are getting when reading the blog is not one that matches reality. The results are misleading unless you read the "fine print", which seems to escape most people who comment on it. You see, while the IE10 preview appears to come out on top, the results are not what most people think they are.

Read more...