Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Linux snickers at Microsoft's victory declaration

zdnet.com

Post reply on HN
Print view View original

41–50 of 53 posts

Re: Linux snickers at Microsoft's victory declaration

#41

Post by shriphani »

Ed Bott's article + this one are the digital equivalent of penis-length-measurement contests. Why is this cheap attempt at acting like Sherlock Holmes on the front-page of HN ?

I am an intern @ MSFT. I use win @ work. Windows+Ubuntu on my home box and load up FreeBSD in a VM (on my home machine) from time to time and used a Mac till very recently (it died not very close to a refresh of the MBP lineup to make it worth waiting and not too far from a refresh to buy the one on sale at the time).

Re: Linux snickers at Microsoft's victory declaration

#42

Post by rbanffy »

jff wrote: post #19

"Supercomputers, the fastest of the fast, run Linux almost exclusively." And it's so disappointing. Consider the Blue Gene series. Great new architecture, custom networks, custom hardware. Initially, they put Linux on the I/O nodes and a custom kernel on the compute nodes, because you want a minimum of OS noise (see the FTQ benchmark for more info on this). But everyone clamored for Linux on compute nodes--so now we…

Plan 9 was a very interesting OS, but it always gave me the impression to be too desktop/workstation influenced. While you could have your heavy lifting carried out on big processing nodes on the net, it depended on having the same one-size-fits-all OS on all boxes. There was a time when we envisioned powerful computers running powerful software on every desktop. That time has more or less passed - most people now are very happy with the browser, the modern equivalent of the 3270.

And I have to tell you that being able to sit down at any 3270 in any company building and to have access to your stuff readily available was, in several ways, a better (if uglier) experience than what desktop workstation provided.

Plan 9 also committed the cardinal sin of being different for the sake of being different. You can build a better Unix without making "vi" a MIPS emulator.

We invent things that are ahead of their time. It's a shame so much of the technology is lost when that happens.

As for the Fortran 77 thing... Well... They have lots of code that has already been written and tested. Do you really want to port it to something else? Besides, legend says F77 compiles to very efficient code.

Re: Linux snickers at Microsoft's victory declaration

#43

Post by tedunangst »

TheCondor wrote: post #36
dredmorbius wrote: post #24

Earlier quoted context omitted.

That's all well and good. But. But: I've been hearing this for well over a decade. "Microsoft's attitude toward Linux is changing." Again and again and again and again. And then some new instance of FUD emerges, or a legal attack on a Free Software / open source advocate or company, some trumped-up performance benchmark, chairs in low-earth orbit. Whatever. The behavior has always failed to match the rhetoric. I'm su…

It'll take a management change and some serious actions. Let's make no mistake, the company has been compeltely and 100% opposed to Linux for a very long time now. What exactly would it take to constitue a change of heart? Seriously? I would think they could extend patent amnesty to all Linux distributions. They could embrace the Mono project. They could acutally release some software on Linux. How about media player…

Just one example, but the asp.net mvc source/reference implementation is available under a liberal license and plays nice with mono.

Re: Linux snickers at Microsoft's victory declaration

#44

Post by tedunangst »

rbanffy wrote: post #42
jff wrote: post #19

"Supercomputers, the fastest of the fast, run Linux almost exclusively." And it's so disappointing. Consider the Blue Gene series. Great new architecture, custom networks, custom hardware. Initially, they put Linux on the I/O nodes and a custom kernel on the compute nodes, because you want a minimum of OS noise (see the FTQ benchmark for more info on this). But everyone clamored for Linux on compute nodes--so now we…

Plan 9 was a very interesting OS, but it always gave me the impression to be too desktop/workstation influenced. While you could have your heavy lifting carried out on big processing nodes on the net, it depended on having the same one-size-fits-all OS on all boxes. There was a time when we envisioned powerful computers running powerful software on every desktop. That time has more or less passed - most people now ar…

I don't get that impression for plan 9 at all. It included lots of features that suggest it was designed for heterogeneous clusters. Like simple convenient cross compiler support. You could run mips, your office mate could run sparc, and you could effectively and somewhat transparently use both to get your work done. I think plan 9 was actually trying to transform the network to be like your 3270 example.

Re: Linux snickers at Microsoft's victory declaration

#45

Post by Nelson69 »

tedunangst wrote: post #43
TheCondor wrote: post #36

Earlier quoted context omitted.

It'll take a management change and some serious actions. Let's make no mistake, the company has been compeltely and 100% opposed to Linux for a very long time now. What exactly would it take to constitue a change of heart? Seriously? I would think they could extend patent amnesty to all Linux distributions. They could embrace the Mono project. They could acutally release some software on Linux. How about media player…

Just one example, but the asp.net mvc source/reference implementation is available under a liberal license and plays nice with mono.

Mono isn't exactly killing it as the go to development platform.

Re: Linux snickers at Microsoft's victory declaration

#46

Post by rbanffy »

tedunangst wrote: post #44
rbanffy wrote: post #42

Earlier quoted context omitted.

Plan 9 was a very interesting OS, but it always gave me the impression to be too desktop/workstation influenced. While you could have your heavy lifting carried out on big processing nodes on the net, it depended on having the same one-size-fits-all OS on all boxes. There was a time when we envisioned powerful computers running powerful software on every desktop. That time has more or less passed - most people now ar…

I don't get that impression for plan 9 at all. It included lots of features that suggest it was designed for heterogeneous clusters. Like simple convenient cross compiler support. You could run mips, your office mate could run sparc, and you could effectively and somewhat transparently use both to get your work done. I think plan 9 was actually trying to transform the network to be like your 3270 example.

> It included lots of features that suggest it was designed for heterogeneous clusters.

... all running Plan 9. I see your point - when it was conceived we had RISC workstations and Pentium PCs and they all could run some flavor of Plan 9, but today we have x86 PCs and ARM gadgets running browsers and x86 boxes running app servers and databases buried in data-centers. It could be built on Plan 9, but that's not how it ended up being.

Re: Linux snickers at Microsoft's victory declaration

#47

Post by rat »

recoiledsnake wrote: post #16
justinsb wrote: post #10

I read Microsoft's original change of wording completely differently - it wasn't that they had defeated Linux, it was that Windows had already lost (largely because their competitors changed the game.) Competitors no longer compete only with Internet Explorer, they now offer an alternative to the entire Windows OS. Linux on the desktop (GNOME/KDE) didn't win per-se, but the browser did win, and Windows isn't a necess…

>it was that Windows had already lost (largely because their competitors changed the game.) Competitors no longer compete only with Internet Explorer, they now offer an alternative to the entire Windows OS. Linux on the desktop (GNOME/KDE) didn't win per-se, but the browser did win, and Windows isn't a necessary choice for a browser-first system. Is that reflected in sales figures? They do acknowledge competition fro…

Chromebook? hahahahaha

Re: Linux snickers at Microsoft's victory declaration

#48

Post by jff »

rbanffy wrote: post #42
jff wrote: post #19

"Supercomputers, the fastest of the fast, run Linux almost exclusively." And it's so disappointing. Consider the Blue Gene series. Great new architecture, custom networks, custom hardware. Initially, they put Linux on the I/O nodes and a custom kernel on the compute nodes, because you want a minimum of OS noise (see the FTQ benchmark for more info on this). But everyone clamored for Linux on compute nodes--so now we…

Plan 9 was a very interesting OS, but it always gave me the impression to be too desktop/workstation influenced. While you could have your heavy lifting carried out on big processing nodes on the net, it depended on having the same one-size-fits-all OS on all boxes. There was a time when we envisioned powerful computers running powerful software on every desktop. That time has more or less passed - most people now ar…

Possibly the biggest motivation for porting Plan 9 was that it provides a very light-weight yet featureful kernel. Linux is way too big, but the compute node kernels IBM and Cray have been coming up with may be too simple. Plan 9 provides a rich environment but doesn't demand much of the environment.

And yes, Fortran 77 can be quite efficient... but on the other hand, we're also seeing a lot of new software being written in Python, which (at least according to our process traces) tries to open thousands of non-existent files on startup. So efficiency isn't everything they care about... it's partly that they were used to Fortran, and now they're used to Python, and something scarily new like Plan 9 is, well, scary.

Oh, and good luck trying to convince people that MPI isn't necessarily the only solution for parallel programming. If the community at large has its way, we'll still be using Linux and MPI in 2040.

Re: Linux snickers at Microsoft's victory declaration

#49

Post by comex »

ajross wrote: post #30
sdkmvx wrote: post #28

Earlier quoted context omitted.

It's interesting that you say this. If POSIX was all that there is to it, then isn't Apple iOS/Mac OS the same as Android and (insert-your-favorite-distro-here) Linux? I think ultimately the differences between user-facing (and even developer-facing, API-level stuff) components are way more important than the fact that POSIX is so common. After all, we would't describe two computers as basically the same because they…

All I'm saying is that by any complexity metric you want to pick (lines of code, whatever) Android and "Linux" are far, far more alike than different. The layer you are picking as "way more important" only seems more important to you because it's the layer you live in. The rest of the software is really important too (where would you be without the networking stack or video codecs, for example?) and to those people (…

> All I'm saying is that by any complexity metric you want to pick (lines of code, whatever) Android and "Linux" are far, far more alike than different.

Citation needed. The Android-specific stuff makes up a lot of code.

Re: Linux snickers at Microsoft's victory declaration

#50

Post by rbanffy »

jff wrote: post #48
rbanffy wrote: post #42

Earlier quoted context omitted.

Plan 9 was a very interesting OS, but it always gave me the impression to be too desktop/workstation influenced. While you could have your heavy lifting carried out on big processing nodes on the net, it depended on having the same one-size-fits-all OS on all boxes. There was a time when we envisioned powerful computers running powerful software on every desktop. That time has more or less passed - most people now ar…

Possibly the biggest motivation for porting Plan 9 was that it provides a very light-weight yet featureful kernel. Linux is way too big, but the compute node kernels IBM and Cray have been coming up with may be too simple. Plan 9 provides a rich environment but doesn't demand much of the environment. And yes, Fortran 77 can be quite efficient... but on the other hand, we're also seeing a lot of new software being wri…

> and something scarily new like Plan 9 is, well, scary

Plan 9 would be more OK if it didn't try to be so not Unix.

That's too bad. It saddens me that, of the three more advanced widely used OS, two are variations of the 70's Unix theme while the third is the bastard child of VMS with lots of lipstick on it. :-( Plan 9 deserved better.

Post reply on HN