Skip to main content
Software Engineering

Timeline for Why was Tanenbaum wrong in the Tanenbaum-Torvalds debates?

Current License: CC BY-SA 3.0

24 events
when toggle format whatbylicensecomment
Aug 22, 2017 at 19:46commentadded Kaz Anything is a microkernel if we identify some handful of functions in it that are called a lot by everything else and then stretch the definition of microkernel to cover that handful of functions.
Oct 21, 2016 at 23:26history edited Cees TimmermanCC BY-SA 3.0
fix name typo
Dec 19, 2012 at 1:11history made wiki Post Made Community Wiki by david
SMay 19, 2012 at 22:29historysuggestedPeter MortensenCC BY-SA 3.0
Copy edited. Added some context.
May 19, 2012 at 21:18review Suggested edits
SMay 19, 2012 at 22:29
Mar 22, 2012 at 21:02commentadded regularfry I don't think point 3 is true in the sense that Tanenbaum meant. At the time, there was a lot of optimism around GNU HURD. Claiming Linux (in any variant) as a GNU OS win is just not historically valid. Userspace? Sure. But in a sense Linux the kernel won in spite of GNU.
Mar 22, 2012 at 16:58commentadded ccoakley @CraigStuntz: Thank you. Yes, and I think that as long as OS vendors have a decent value proposition for people, we won't all switch to free GNU OSs. I don't think Tan- could foresee the advent of non-free OSs using many GNU and other Open Source technologies. OS X would be nowhere without BSD and GNU tools. I doubt Android would exist without similar Open Source tech available.
Mar 22, 2012 at 16:49commentadded Craig Stuntz @ccoakley: Well, the claim was "everyone will be running a free GNU OS." If that were true, wouldn't Apple's market share be falling?
Mar 22, 2012 at 16:44commentadded ccoakley @CraigStuntz Bummer. You teased me with your "Seems relevant to the question at hand." I thought you were implying larger implications than what was discussed in the link, and I just didn't know which way you were going.
Mar 22, 2012 at 16:35commentadded Craig Stuntz @ccoakley: My point is limited to Macs: Apple seems unwilling to ship GPL v3 anything. See the link in my comment.
Mar 22, 2012 at 15:11history edited JeremyCC BY-SA 3.0
added 3 characters in body
Mar 22, 2012 at 15:06commentadded BlueRaja - Danny Pflughoeft Regarding point 2: I've been saying for years that once .Net becomes ubiquitous, MS will start moving away from x86. Windows 8 will run on both ARM and x86.
Mar 22, 2012 at 14:36commentadded ccoakley @CraigStuntz - elaborate on GPLv3 specifically? Are you saying that the adoption of v3 slowed the adoption of GNU tech, something that T- didn't foresee? Or do you mean "GPL weakens the value proposition due to concerns over IP leakage?" Because I think that was more relevant before the debate than after (empirically, businesses seem much more likely to use GPL software today than then, and I speculate that some of that has to do with reduced fears over IP leakage, founded or otherwise).
Mar 22, 2012 at 14:16commentadded Craig Stuntz "All Macs ship with an archaic version of GNU's bash," probably to avoid GPL v3 terms. Seems relevant to the question at hand.
Mar 22, 2012 at 11:13history edited Mark BoothCC BY-SA 3.0
Changed the 'Point X' references into titles, so the answer flows better.
Mar 22, 2012 at 9:24commentadded ccoakley @KonradRudolph recreating citation for total sales via google: 3 billion ARM chips shipped in 2009 (vanshardware.com/2010/08/mirror-the-coming-war-arm-versus-x86) while an estimated 400 million x86 chips sold in 2011 (4 times quarterly number reported here: computerworlduk.com/news/it-business/3345299/…). The embedded market is huge and largely non-Intel.
Mar 22, 2012 at 9:23commentadded ccoakley @KonradRudolph The claim that apple would outsell all x86 sources was by a semiconductor market analyst around the time of the new iPad announcement also based on rumors of many current intel channels jumping ship (related, but different analyst prediction via google: itproportal.com/2012/03/21/…). The total chips in use and sales was from an academic presentation.
Mar 22, 2012 at 8:14commentadded o0'. @KonradRudolph a citation from the future?
Mar 22, 2012 at 8:02commentadded Konrad Rudolph [citation needed] In particular the claim that Apple may outsell all x86 vendors combined.
Mar 22, 2012 at 4:02commentadded fluffy Also regarding part 2, the various x86 modernizations actually go a few steps beyond RISC and do some fascinating stuff with "micro-operations" internally which give even better scheduling flexibility with on-the-fly instruction reordering, which brings performance boosts that are far beyond what RISC adherents can even dream about. RISC CPUs could get that too but at this point you're not comparing RISC vs. CISC, you're comparing various in-hardware JIT strategies with fairly abstract ISA frontends.
Mar 22, 2012 at 1:34voteacceptRobz
Mar 22, 2012 at 1:07commentadded ccoakley @deadalnix - I meant to make an explicit mention of the scope on 3, restricting to desktop computers. If a google server farm runs linux (does it?) and each box counts toward the linux count, and my cable modem running a stripped down version of linux counts, and my android phone counts, then linux probably dominates. But while I've seen nice charts of ARM vs x86 processors sold/used over time, I've not seen one for the OSs, particularly when you mix in the embedded devices.
Mar 22, 2012 at 0:44commentadded deadalnix On point 3, linux is very widely spread, probably the most used OS, windows and VxWorks following. So point 3 can be considered as true. It make 2 point out of 3 correct, which is pretty good considering how unpredictable IT is.
Mar 21, 2012 at 23:33history answered ccoakleyCC BY-SA 3.0
toggle format
close