Bob Metcalfe on Open Source, IPv6, IETF

prostoalex writes "The inventor of Ethernet Bob Metcalfe is interviewed by AlwaysOn on current issues. Metcalfe is known for challenging commonly accepted wisdom and this time he's quite confrontational. On open source and operating systems: "If you look at Windows and Linux, both are based on 25-year-old technology. Windows is sort of a GUI version of the Mac's operating system, and Linux is of course Unix, which stems from 1968. These are both old clunkers. So the question is, Where are the new operating systems likely to come from?" On IPv6 adoption and IETF: "Back when you attended the IETF, you all looked down your noses at the ITU (or I guess it was called CCITT at the time)--the entrenched, corporately manipulated, corrupt, competent standards being embodied in IT. We were the IETF--the swashbuckling, institution-oriented, open people, the rebels. That's changed now. The Internet has arrived, and all of those people are now just like ITU: IETF has become the ITU.""

Comments

Inconsistent Rant

[XorNand]

If you look at Windows and Linux, both are based on 25-year-old technology. Windows is sort of a GUI version of the Mac's operating system, and Linux is of course Unix, which stems from 1968. These are both old clunkers. So the question is, Where are the new operating systems likely to come from?

I guess someone should tell automakers that they should reinvent a mode of transportation from scratch. That four wheels, an engine, and brake and throttle thing is so passé nowadays. Plumbers are going to be pretty upset to learn that using pipes to carry water is so several-centuries ago.

There's no doubt that Mr. Metcalfe is quite bright and has contibuted greatly to the IT world, but I don't understand this rant. If he doesn't see the innovation, I guess he's never compared Slackware '96 to today's distros, or Windows 3.1 to WinXP. Apple certainly can't be ignored here either. Where are the new operating systems likely to come from? I'm going to take a wild guess, and say "probably from the OS's of today." They don't need to be completely rewritten every few years to count as progress. Even the emergence of UNIX itself was evolutionary, not revolutionary.

It's also interesting that he clearly shows a lack of faith in the OSS community, but then digs at the IETF for evolving into elitist and monolithic organization. ::scatches head:: Reading through the article, he doesn't seem to be very consistant with his views.

Re:Inconsistent Rant

[Qzukk]

he doesn't seem to be very consistant with his views.

Actually, I think it's just the opposite. He's very consistent: "Everything sucks."

Thoughts

[AKAImBatman]

Windows is sort of a GUI version of the Mac's operating system, and Linux is of course Unix, which stems from 1968. These are both old clunkers. So the question is, Where are the new operating systems likely to come from?

I hate to break it to Mr. Metcalfe, but most entities lack the resources to do a ground up rewrite of a fully featured Operating System. Simply writing a functional OS isn't the hard part. It's just a platform upon which software will be built. There were hundreds of OSes written between 1960 and 1990. During the '90s, however, computing platforms began to stabalize. Software was written that had a greater than 5 year life span, Operating System began to stabalize on a few "standards" (namely Unix/Vax/CPM derivitives), and massive amounts of time and money were invested into developing these platforms. Now we're standing on the 10,000 ft high towers we call Windows, Linux, and Mac OS X(NextSTEP) and we're looking at how difficult it is to replicate the decades of work that has gone into these systems.

Building a more powerful and "correct" system would mean throwing away software such as OpenOffice, Mozilla, Quickbooks, Photoshop, Acrobat, etc. Software that took decades to build! Thus any future solution based on cutting edge CompSci Technology must either bite the bullet and rewrite these complex apps (good luck) or build in a translation layer that allows them to continue working. Neither choice is very appealing.

The "third road" that is currently being explored is the road of running Virtual Machines on top of today's existing infrastructure. Java, .NET, and Mono are examples of the market attempting to find a way to combine modern technology with the tech of yesterday. Unfortunately, the results are less than stellar. For example, instead of aligning Virtual Memory along object bounds (a natural fit that could be done without hardware support), these systems must contend with the existing 4K VM implementations. Instead of running the protected code in a flat heap (which CAN'T break the memory model!) these systems must contend with the memory indirection that operating systems throw their way. The results of this poor matchup between machine and software is a performance penalty, both real and perceived.

The Virtual Memory swaps more than it should. Object files are not shared. Memory usage is 20% greater than a native program. So on and so forth.

A lot of research has gone into mitigating these issues (with Sun producing some very impressive results!), but it doesn't change the fact that the machine and software are mismatched. That mismatch discourages companies from writing new applications in these managed environments, where they would be free from the bonds of traditional OS designs.

My gut says that a rather major shift in how we use our computer will have to happen before we can truely replace the systems we have today.

I'd like to point out that two major pieces of infrastructure were left out of the Internet when it was being built--largely because it was built by graduate students (and people like graduate students). They left out security and economics. So we have the spam problem (which can be traced directly to the lack of concern for security), and we have IP rules that are in flux because the Internet doesn't have the right tools for monetizing various activities. So we're busily trying to put security and economics into the Internet.

In all honesty, the Internet never would have been as successful as it was if it wasn't for the freedom it provided. Many other networks offered these features, but they were eventually usurped by the Internet.

Hindsight is 20/20. Had the BSD/ARPANET folks attempted to address these issues back when it was created (which would have been ludicrous given its Military intent), their solutions would have likely been wrong. Keep It Simple Stupid. It may not be the best solution, but it's the most effective solution.

P.S. In case of Slashdotting, break glass

Oh the irony

[DarthVeda]

That a site called Always On has been slashdotted.

Re:Clunkers?

[AKAImBatman]

Mac OS X has one foot in the past (FreeBSD) and one foot in the future (Mach). Between the two of them, OS X is a bit more futuristic than its competitors. However, a machine that died before its time was the Symbolics LISP machines. I've never had the opportunity to use one, but my understanding of these machines is that they never needed any of the "modern" processor or software concepts we use today, because the underlying software system (LISP) was incapable of creating the types of memory corruption we try to prevent even today. And if a program blew up, you could actually modify its memory image on the fly and continue its execution.

On top of everything, the hardware architecture was much faster than contemporary computers due to its LISP oriented design. Apparently, a good portion of the LISP language was able to execute directly in the hardware!

At least, that's how I understood it. Sadly, it didn't get much attention outside of academia. :-/

Nothing to see here, move along

[$RANDOMLUSER]

He's just trolling. He has a pathological need to pop up every once in a while and say "You know I invented Ethernet, don't you?"