Slashdot Mirror
Linus Torvalds Receives IEEE Computer Pioneer Award
mikejuk (1801200) writes "Linus Torvalds, the 'man who invented Linux' is the 2014 recipient of the IEEE Computer Society's Computer Pioneer Award, '[f]or pioneering development of the Linux kernel using the open-source approach.' According to Wikipedia, Torvalds had wanted to call the kernel he developed Freax (a combination of 'free,' 'freak,' and the letter X to indicate that it is a Unix-like system), but his friend Ari Lemmke, who administered the FTP server it was first hosted for download, named Torvalds' directory linux. In some ways Git can be seen as his more important contribution — but as it dates from 2005 it is outside the remit of the IEEE Computer Pioneer award."
He is very deserving of the award. Well done.
"Git can be seen as his more important contribution"
Umm no. The early 1990's were dark days. Linux was/is a big deal. Where would we be without Linux? It changed the world! The same can't quite be said about Git (although great in its domain).
Git is a nice tool, but it's not even close to his work on Linux. Orders of magnitude less important. You'd be paying for commercial licenses of Solaris, GNU would have seen far less of an audience and not progressed nearly as fast as it has, and you'd be paying VMware license fees every time you started up an EC2 instance. If Git disappeared tomorrow, I'd switch to svn and probably grumble a couple more times than normal. If Linux disappeared tomorrow, I'd be bankrupt and broke.
there _was_ no free operating system for industry standard hardware, much less a Unix-like one, and the commercial offerings were all platform-specific.
If you wanted a real computer that could do real stuff (as opposed to a DOS box, which wasn't even network aware in any substantive way, and even in non-substantive ways required $$$ for bare-bones, single-function software tools that were cobbled together out of batch files and nonsense), you had to:
- Get your hands on dedicated Unix workstation hardware, which was often poorly documented/supported outside of a corporate sales account
- This meant either $tens of thousands for current workstation hardware or $thousands for last-cycle hardware if it was even available at all (university and government surplus lots were the primary suspects)
- Phone up the one or two providers that offered OSes for the system
- Shell out $many thousands for a license (and often $thousands more for media)
- In many cases, because non-current hardware was tied to non-current OSes no longer for sale, port the current tree yourself to the non-current hardware after spending the $thousands you spent for a license
In short, it was substantively impossible for—say—a small company, a startup, or a CS/CE student to get their hands on anything beyond a DOS box with Windows 3 on it. With money and time, they MIGHT get web BROWSING working on Windows 3—in unstable ways. Developing software was a nightmare on these DOS/Win3 boxes as well—compilers were expensive, proprietary, and often required runtimes that had to be licensed on a per-user basis (i.e. you spent $200 on the compiler that spoke a non-standard dialect, then if you wanted to sell what you created, you spent another $some amount per copy sold) and that had no hooks for anything network-ish, because there were no standards in the DOS ecosystem for that.
Linux changed everything. Suddenly, you could pick up commodity i386 hardware and actually do network stuff with it in Unix-y ways. Even in the early days when Linux was unstable, incomplete, and a bear to install/configure, it made things possible for small shops or independent developers/creators that had simply been prohibitive in every practical way just a year earlier.
As a result, the Unix networking ways—thanks in many ways directly to Linux—would eventually become the industry standard form of networking (TCP/IP over ethernet) that we take for granted today—but in no way was history certain to end up this way. We could just have well been tossing the equivalent of glorified FidoNet payloads today.
Without Linux, GNU, and BSD, it's no stretch to say that we may not have had an Internet today in any way that we'd recognize, and certainly Linux has been the most visible and most widely distributed amongst the three.
Much more than the work by Berners-Lee, Linus Torvalds invented the future that we live in.
STOP . AMERICA . NOW
Linus Torvalds did not "invent" Linux. He implemented a POSIX kernel, working from basic UNIX standards and preexisting hardware (the 80386 MMU). UNIX was an invention. Linux was "just" an implementation. As it grew, there were various inventions going into it. But Linux "as such" was not an invention.
In contrast, Torvalds did basically invent Git. Its shape and functionality, as opposed to what Linux started with, were not predetermined.
Git is really a gift from BitKeeper. If BitKeeper had chosen to not be dicks, everybody would still be using it for Linux kernel development. Hell, BitKeeper doesn't even put pricing information on their web site, you need to 'request' it [but you know it ain't cheap if they say it costs $$$ from a range of zero $ to $$$$].
Sleep your way to a whiter smile...date a dentist!
We should not diminish the importance of Linux. But it's clear that Git is much more important today. Linux is wonderful, but its a commodity for most people. It doesn't matter that Android is based on Linux. It's awesome but most people don't care. It's just a technicality.
It's "just a technicality" in the sense that Android might not exist if Linux hadn't existed; saying that it's less relevant because people don't know it's there is like saying that ARM isn't all that important because most people don't know they have ARM processors in their smartphones. Git is even less directly relevant to most people, as they're not developers.
What Linus did by creating GitHub is of tremendously much more importance if you look at how well it brings open source developers together.
Presumably you meant "by creating Git"; as far as I know, he no more created GitHub than he created Android, even if GitHub uses Git and Android uses Linux.
"In some ways Git can be seen as his more important contribution" - thus spake someone who was not there in the early 90's, and who takes free software and OS competition for granted. In other words, someone who is naive beyond words.
The change brought by Git is insignificant next to that brought by Linux. Utterly insignificant.
I'll bite :-)
I used csv and subversion back in the day, switched to hg, and now switched to git. I manage a smallish project with 5 or so contributors and contribute to some other projects.
Git/hg vs csv/svn is all about distributed vs centralized. With git/hg, you learn to love branching and merging, and commit as often as needed.
Git vs hg is more subtle, but I am strongly in the git camp now.
In my perception, hg et al are about lines of code. You contribute code and the code is checked in. git is all about commits. Your work is in commits, and commits can be rebased, squashed, amended, etc until they are just right to express your contribution. Git is not so much about communicating with yourself about how you got to your code; git is about communicating to the rest of the team what you are contributing. In a sense, you are not (just) writing code, you are writing a commit history.
That said, what I miss in git is the "version history" of commits. I would like to see some sort of "is-based-on" link between the 'final' commit and the commits it is amended, rebased, and/or squashed from. I would love to be able to 'expand' a final commit to see the history that went into it, because now you are sometimes choosing between commit elegance and keeping track of development history (aka in the choice to amend a silly type you choose elegance; in the choice to -no-ff merge a branch you choose history).
I'm not going to say the kids need to leave the lawn, but saying Git can be seen in any way as Torvalds' more important contribution is speaking from ignorance. The people who say there were other OSes that could've filled the same role, but then list off prices for each, are ignorant too.
Linux was free and freely available.
I went from installing it out of the back of a book and from some odd company named Yggdrasil's ftp server, to installing it for a multi-million dollar enterprise fail-over solution.
I went from twiddling values for "drums" to get my hard disks recognized, to it upgrading itself unattended on a phone in my pocket.
Git got to where it was because Torvalds mandated it for Linux contributions. Linux, and the rest of the world, would be fine if Git didn't exist. There were and are plenty of free revision control systems out there. No one can say the same for Linux.
You can tell whether or not someone was actually there by whether or not they mention things like "Minix" in a list of viable operating systems.
I was part of a project at the time that needed real networking and a real Unix development environment. We spent four months working to find an alternative, then shelled out for a series of early Sparc pizza boxes. SS2 boxes maybe? As I recall, we got four at nearly $15k each that ate up a huge chunk of our budget.
Two years later, we had liquidated them and were doing all of the same stuff on Linux with cheap 486 boxes and commodity hardware, and using the GNU userland and toolchain. People here talk about GNU as predating Linux while forgetting that prior to Linux, the only place to run it was on your freaking Sparcstation (or equivalent—but certainly not under Minix), which already came with a vendor-supported userland. GNU starts to be interesting exactly when Linux becomes viable.
All in all, the change was bizarrely cool and amazing. We were like kids in a candy store—computing was suddenly so cheap as to almost be free, rather than the single most expensive non-labor cost in a project.
STOP . AMERICA . NOW