I daresay that marketing, startups, and quite a few large companies pushing for IPOs are not going to want to hear this. But when he's right, he's right.
Spot on Linus. I raise a toast to you.
Linus Torvalds believes the technology industry's celebration of innovation is smug, self-congratulatory, and self-serving. The term of art he used was more blunt: "The innovation the industry talks about so much is bullshit," he said. "Anybody can innovate. Don't do this big 'think different'... screw that. It's meaningless. …
All the faffying companies like BT spend trying to sweat their copper assets, all the bullshit analysis of obfuscated, bamboozled "upto" speeds, distance by cable, whether the cable is copper, aluminum, actively powered devices, non-actively powered devices. Whether than can squeeze extra distance by forcing existing customers off ADSL, to implement long range VDSL. It's all just delaying tactics / hyped vapourware crap of pointless "upto" copper carcass G.fast.
Bullshit paper shuffling Ofcom stats whether its better to keep Openreach part of BT / Separate. Non of this would matter if the copper cables in the ground were put through a programme of replacement to fibre. Sack all these anaylsis numpties, do the 'real work', employ the engineers/people to do the hard grunt - the real world jobs that actually matter. Fibre optic cables replacements on the ground.
Then, all these analysis/regulatory jobs are redundant. (though I'm sure they'd try).
It's times like these we need someone like LinusTorvalds to say 'everyone - shut the fuck up'. Let's start putting true Fibre optics in the ground FTTP, and move forward, for the sake of UK.
Copper is a dead carcass, going forward.
There is hard work that needs to be done, and it needs starting now. BT needs to be sidelined, if they are unwilling to go this route.
Enough talking Ofcom, we need some proper decisions, you need to be brave and sign your own death warrant, with the UK taking on a fibre replacement programme for new builds / end of life replacement. (as a bare minimum).
LOL, Trumpton, for all you lot abroad, over the pond etc, it's a kids TV show from way back when from the UK and is called 'Trumpton' believe it or not!
'The big, fat, orange idiot-in-chief can kiss my pasty white butt if it thinks otherwise.'
LOL, really, but do you think you'd get away with saying "The big black idiot-in-chief", a while back? Just saying, and your not alone mine is pasty and white too.
er.... I think you'll find its actually a reference to a very old card game or 10. One suit trumps another, the Ace trumps everything and so on, nothing to do with any individual. And since I were a wee lad in the 60's, trumping is a reference to farting. And Trumpton was one of my favorite shows..
One thing about Windows that really annoys me is when a fundamental operation keels over, but the animation continues as if no error occurred.
Note to Microsoft: Ever thought of assigning the team that does the animation to work on such things as programming operating system commands?
They learned that lesson very very well. The one about the on-screen cursors and their animation, that is. In the book Revolution in The Valley: The Insanely Great Story of How the Mac Was Made there is an awesome story about Bill Gates visiting the Apple campus for some demo or other, Microsoft was an early and huge Mac software developer in the 80s, and Bill was absolutely baffled at how the Mac cursor could move without any flicker. Bill kept asking if there was another computer doing the demo in another room, or that they were using special dedicated hardware for it. Jobs just smiled and said, nope it's all in software. Microsoft figured it out in the end, and must have got really good at it. :P
Open Source is people. Hence the difficulties, and the victories.
Open Source is not about being fettered by corporate, rather it is about being self-harnessed and self-directed towards common goals together with others of like mind.
Think of a couple of open source projects you use - most likely the contributors have shifted around greatly over time, with others coming in to infuse new oomph and ideas every so often. This is a hidden strength of open source projects. I'm thinking of two where the original developers haven't been heard from for more than a decade, but the code is maintained and expanded by what is perhaps the 4th generation now. (And, yes, it's a small %age of new projects with lots of counter examples - no panaceas here)
Ideally there's a balance of some innovation delivered with solid program running.
My instinct is a lot of a Linux kernel is about protocols and protocol implementation.
I wonder how many of the developers do this as an FSM with the code written by the tool?
Or how many (eventually) implement it as a bodged up FSM?
"My instinct is a lot of a Linux kernel is about protocols and protocol implementation."
Download it and have a nose around. Even for a simple sysadmin like myself there are some interesting things in there.
I think you'll discover it is rather more than just a finite state machine's output. Although there is one actually in it.
Female seeking male? Absolutely not. The LKML ain't a meat market. <sfsf>
More seriously, the only person I can speak for is myself ... The only FSM I use is the one between my ears; no tool generated code here. Reading through the source code, I suspect I'm not alone. The source is open, you can read along for yourself, if you like.
Disclaimer: I'm not a coder, per se, but I have been contributing for a long time.
 Yes, I know, I'm such a tool ... Feel better now?
Female seeking male? Absolutely not.
I stared at the letters FSM for an embarrassingly long time before the words "Finite State Machine" trickled reluctantly to the forefront of my consciousness. It wasn't until I read further down the comments that "Flying Spaghetti Monster" even occurred to me.
Does that make me a bad person?
Err, it's an IT site. IBM manuals tend to describe protocols using state diagrams.
Obviously such practices are less common than I thought.
Perhaps that's why such work is so hard. *
Note. True FSM's are not Turing complete. The question is how much of the functionality in the kernel needs needs something more powerful than an FSM to deal with situations due to massive growth in the number of possible states.
Torvalds said he subscribes to the view that successful projects are 99 per cent perspiration, and one per cent innovation.
"Genius is one percent inspiration and ninety-nine percent perspiration." - Thomas A. Edison
Gee, same concept, a century or so later. I'm not surprised. Everything goes to crap when shortcuts are taken.
You could argue (I won't) that Linux cloned a lot of ideas from Unix/BSD/Minix. Lots of perspiration, not all that much innovation. In fact, a troll is surely waiting in the wings to chip that in. Or maybe whine that it's not a micro-kernel OS.
But git seems to be another beasty entirely. Yes, it's a VCS, and yes, those have been done before. But, it's a different VCS in how it stores data. And it is a very different VCS in how its API interacts with standard file system commands.
I confess I am a fairly clueless person wrt git. The railroad diagrams used to explain commit points just make me scratch my head. Lots of the advanced command and option combinations are incredibly complex and don't make that much sense, so I tend to cut and paste them from SO.
But unlike some of the other VCS I have had the misfortune to interact with, I get the feeling that it's actually pretty elegant, if I was clever enough to grok it fully. Commands like git bisect, once you get with it does, just are "wow, this is so useful, why did no one do this before?". I am not sure how long he took to put git together, but as I recall, he had to do it when some (other) Linux devs kinda pulled a fast one on the vendor of a commercial VCS that they had been using (for free, IIRC).
ClearCase was one of those VCS "misfortunes" I worked with. "Abortion" is a way better term for CC tho ;-)
To be fair, with git it's not that advanced commands don't make sense - they work and work well. It's just that if it's not your primary job to wrangle with it, then it's not worth the effort to really know obscure stuff and edge cases. But since git is command line - copy, paste, adjust, and run is doable. Much more so than with a super complex nested GUI like CC.
...many great improvements in computing came from people who were just doing something properly. Just think of UNIX. You have a bunch of people who were trying to put in a semi-minimal amount of work. Only features that had most "bang for the buck" were implemented, and the whole thing has a "can't somebody else do it" attitude. (having small programs for everything)
Today many people see innovation as doing trivial things more and more complex. Android, systemd or much of the Freedesktop projects are prime examples for this. I think this is because we have an excess of bad programmers who all want to do something... without understanding how to do it in a minimal way. That way they create lots of code that doesn't do anything productive.
The two parts of real success in creating and sustaining anything are, to me, (1) focusing on solving a practical, actual need and (2) work with a tried and trusted group of people.
This has always got results, in my experience. Sometimes the team came together as a bunch of unknowns and, through the project, we found out who were the good 'uns and who were not, but in the end you have a core of people who do their best, who are dedicated, and who are bright. Now point them to a real, present problem that needs to be fixed. Job done.
"Is it just me or does Linus tend to have a lot of process problems"
It's not just you as others seem to come to the same conclusion. It's a consequence of your looking at the exceptions, not the rule. Consider the situation:
- He has a huge number of contributors
- He's probably never met most of them
- He never recruited any of them
- He doesn't employ any of them nor work for a company that employs them
- He doesn't provide any annual assessments of them
- He doesn't recommend pay levels
- He can't fire them
If you were in that position and responsible for a project of such magnitude what management tools would you have to hand and what process problems might you experience?
From my company, hence AC.
"We have innovatively wrapped some PHP around some Windows Perfmon output and now we can innovatively see how all our servers are performing on our innovative dashboard!
If you're interested in this innovative system, please contact our innovators!"
Well, I'm exaggerating but really not very much at all. The press release would only be marginally more embarrassing with a few Eleventy ONES!!! in it.
"... Linux is the dominant operating system for servers. Almost all high-performance computing runs on Linux. And the majority of mobile devices and embedded devices rely on Linux under the hood...."
Well, this is not really true. On large business servers segment, Linux does not even exist. That segment belongs to IBM Mainframes, SPARC, POWER. There are no large business servers based on x86, because they did not even existed until just recently. When you need large business workloads, such as SAP, databases, etc - you have no other choice than Mainframes, SPARC, POWER. Just check the SAP world records, the top spots all belong to RISC. The fastest x86 is a very new 16-socket server (which is a redesigned HP Superdome Unix server) and that SAP number is quite bad. Also, all various TPC benchmarks all belong to RISC. There are no x86 servers on the top.
If you need good scale-up perfomance, you need Unix running on RISC. Linux does not scale good on scale-up servers. Linux scales fine on clusters with many small nodes, but does not scale beyond 4-8 sockets on a single server.
Show me a good SAP benchmark running on Linux. You will not find any. Why? Because such scale-up business servers need a well scaling OS. And Linux is not one of them.
If *sap* is your benchmark for "large* server workloads, your view is too narrow.
RISC/spARC/PaRISC are finding narrower and narrower scopes these days. X86_64 running linux is taking more and more territory over. Even in the *largest* enterprises. Mainframes will never go completely away, but they TOO are running more linux now than ever before.
RISC and spARC both have linux implementations. And *wince* I've actually dropped linux on an AIX frame, replacing only the OS and the application code.
Now back in the day the SP2 boxes where the choice for oil companies crunching through their geophysical data for the umpteenth time. I'm not sure if this is still the case.
If anyone has the right to say they originated "data mining" it's the oil and gas industries because the data sets were (and are) vast and the cost of repeating a survey eye watering.
Are you not aware Linux runs happily on IBM Mainframes, it's even heavily promoted by IBM for those who work with Linux but want the perceived advantage of a mainframe compared to other options (e.g. Intel server clusters)
Disclosure I now a few people who were involved in getting NIXes to work on IBM kit way back in the day that led to current state of NIXes running fine on System Z etc.
> all various TPC benchmarks all belong to RISC. There are no x86 servers on the top.
Well in the case of the TPC ones there's a catch. If you want to run a TCP benchmark using Oracle's DB then Oracle have to sign it off. If it competes with their systems they won't. I've seen numbers which embarrassed SPARC boxes but can't be published as they won't ever be signed off.
AC for obvious reasons.
Agile & DevOps are only good for ephemeral shit, like that new-fangled WWWthingie. They are only used by OnPaperItWorks companies looking for angel investors, and FlyByNight industries in search of venturecap funding. No buzzwords, no free money. In other words, outfits that "talk of tech innovation", not the ones that shut up and get it done.
"Torvalds said he subscribes to the view that successful projects are 99 per cent perspiration, and one per cent innovation."
And with that 1% innovation, he has achived a hell of a lot more than most people have. Including you.
did anyone ELSE notice how 'Agile' (aka FRAgile) wasn't mentioned in Linus' development process? It sounds kinda "top down" to ME! And it's SUCCESSFUL!
That kind of success requires leadership and proper management. Maybe Micro-shaft can learn a thing or two from Linus? (but they probably won't)
Uname -a tells me that I have 2.6.32-504.el6.i686 running at present on my old Thinkpad X61s (Stella Linux, a Centos 6 'remix). Hardware vintage matches well with kernel. Newer hardware not so much. Plus Centos/Redhat backport certain fixes to their stable kernels.
If it works for you, then go with it! The freedom of OS.
Biting the hand that feeds IT © 1998–2021