Software

Close Encounters of the Redmond Kind

Oh, Linux is Linux, and Windows is Windows, and never the twain shall meet — isn’t that how the saying goes?

Well, not quite, of course, but it might have been if Kipling had only lived long enough to get to know our favorite operating system. The fact of the matter is that Linux and Windows must and do meet up in a variety of ways — several of which have figured prominently on the Linux blogs in the last week or two.

The result? No end of fodder for conversation.

Dual-Boot Dell Laptops

To wit: When the EE Times reported earlier this month that Dell had released a hybrid laptop running both Linux and Windows, bloggers on Slashdot, DaniWeb and elsewhere pondered what it might mean for Linux in the long run.

“For a certain class of business user, this could be useful,” wrote DaniWeb’s Ron Miller. “I just don’t see it having wide market appeal.”

Even more definitively: “No!” wrote erroneus on Slashdot. “Rebooting is a chore. Once people start up, they don’t want to shut down to start up another application. It’s not what they are used to.”

Then again, “if this were done as a VM where the Linux machine were to boot and they installed Windows XP in a VirtualBox or some other VM, then that might be acceptable,” erroneus added. “Then they would have their safer, virus-free environment for email and web browsing and then a VM to host the applications they need to run. This stuff works really well.”

‘An Endless Source of Problems’

Along similar lines: “Dual-booting is an endless source of problems,” Slashdot blogger drinkypoo told LinuxInsider. “Windows is bad at working with Linux filesystems (there is some very limited and crashy Ext2 support, which has always made Windows XP extremely unreliable for me) and Linux is slow when working with NTFS, so just keeping track of where your files are and dealing with them becomes a hassle.”

That said, “there are already some very successful mini-media-Linux systems in the BIOS of some Asus motherboards, which seems like a genuinely useful form of Linux multi-boot — although there is little to no reason to implement it in the BIOS,” drinkypoo added.

On the other hand: “I think this accomplishes a lot,” Slashdot blogger yagu asserted. “It lets people get their warm fuzzy blankets (Windows and all that they run on it) *and*, for free, it lets them kick the Linux tires. This is good.

“I’m willing to bet a lot of people forget they’ve booted Linux and spend many productive hours/days and have their own ‘ahas’,” yagu told LinuxInsider. “Not saying it’s a sea change, but every little bit helps. Good for Dell. Here’s hoping it isn’t buried so deeply in their web site that it might as well not exist.”

Firefox Faster on Wine?!

Also drumming up some significant controversy in the blogosphere was a report in which TuxRadar conducted a cross-platform comparison by running some simple JavaScript benchmarks of Firefox 3.0 using both Windows and Linux.

The result? “Pretty bleak for Linux,” the report says. Further, Firefox was actually found to run faster on Wine than it did natively.

Oooh, them’s fighting words!

‘I’m Not Surprised’

“Mozilla created Firefox for Windows, and then they made a half-assed version for Linux,” wrote iYk6 on Slashdot, where almost 500 comments joined the near 200 or so on Digg. “I’m not really surprised that the Windows version runs faster. Wine usually runs programs at about the same speed as the Windows version. Sometimes a little more, sometimes a little less.

“I don’t see how this ‘looks bleak for Linux’,” iYk6 added. “Damn trolls.”

Another view: “Check the doco — Firefox 3.0 built for Windows was PGOed (Profile Guided Optimisation),” noted tqft. “PGO was not yet enabled for linux builds. Try a newer build. FAIL.”

‘The Benchmark Was Flawed’

Indeed, “the benchmark was flawed due to the fact that the Windows version is compiled with better optimizations than the Linux version,” Montreal consultant and Slashdot blogger Gerhard Mack agreed.

“PGO vs non-PGO can show exactly the performance difference they saw in the benchmarks,” Mack told LinuxInsider. “Had they benchmarked using Swiftweasel, I’m sure the results would have been more even.”

Explained a different way: “Mozilla Firefox is compiled on Windows by the non-free Intel compiler which, of course — being designed by Intel, one of the largest manufacturers of central processing units in the world — is naturally going to have optimizations for Intel SIMD technologies such as SSE-SSE3,” Slashdot blogger hairyfeet told LinuxInsider.

Dueling Compilers

“Any builds running native on Linux will be compiled using the GNU compiler, which, while following the FLOSS philosophy by allowing viewing of the source code, doesn’t natively support CPU-specific flags such as SSE or 3DNow! — which means that it will always lose to the SIMD-aware Intel compiler,” hairyfeet added. “While I understand that there is an Intel compiler for Linux, there is no distro that I know of that uses it for their distributions due to the lack of source code.

“So to make this a truly fair comparison you would have to compile both the Linux and Windows version with either the GNU compiler or the Intel one,” he added. “Until that happens, I’m afraid the Windows-based Intel-compiled version will always win.”

Widely interpreted languages like Javascript tend to be “much slower than compiled languages,” but “these days, we have so much performance we can afford to use them,” blogger Robert Pogson told LinuxInsider.

‘Perhaps We Need a Fork’

“I know that when one optimizes code, one optimizes for one particular thing at a time,” Pogson explained. “Only the Firefox team knows what they optimize for, and it may well be that other OS.”

Clearly, “not all benchmarks are testing the same thing because some show better for GNU/Linux and some for that other OS,” Pogson added. “Perhaps we need a fork of Firefox if the team is optimizing for that other OS, or perhaps it does not matter if performance is good enough on either system.”

Summing up, the report “does say good things about the overall quality of Wine; it says little to nothing about Firefox,” drinkypoo added.

MS and Red Hat, Sittin’ in a Tree?

Perhaps most shocking of all on the Redmond front in recent weeks, however, was the news that Red Hat and Microsoft have signed an agreement for patent-free virtualization interoperability.

Yikes — an actual agreement!

Not surprisingly, the announcement was immediately pounced upon at Slashdot, Digg, the Linux Loop and LXer, where bloggers far and wide debated the question, “Is this a good thing?”

“It’s interesting to me that Microsoft cares more about interoperability in this case than its stupid patent threats,” Mack said.

But does it really?

‘Nope, Not Buying It’

“Microsoft has never been and will never be trustworthy,” drinkypoo warned. “Those who forget the lessons of history, et cetera …”

Put more explicitly: “I can’t get right with this,” yagu said. “Microsoft just needs to stop, go away. They are clear they are a company that wants domination only. These ‘efforts’ are canards.

“I’m biased, but my bias is from 20+ years watching Microsoft over and over offer olive branches and repeatedly turn on their partners,” yagu explained. “Anyone who thinks Red Hat and Microsoft are a great thing, a great match, should spend time reading the history of Microsoft and IBM circa 1991 and their partnership on OS/2.”

In short, “it’s really amazing what people forget,” yagu concluded. “Nope, not buying it.”

Leave a Comment

Please sign in to post or reply to a comment. New users create a free account.

Related Stories

LinuxInsider Channels