Slashdot Mirror
Torvalds ABCNews Transcripts
yuri-g sent us a link to the
transcripts of last nights ABCNews Chat with Linus.
While he doesn't say what Transmeta does, he talks about the
Cult of Linus, Linux on Consumer Electronics, Distributions,
and the ever popular LiGnuX debate. Somewhat lengthy, but
a good summary, with some new info too.
I did some moderation on this section, so I have to post as an AC. That aside, I think that something needs to be said in response to the individual above who posted defamatory remarks about the nuclear power industry.
First, it should be observed that there have not been any problems with the nuclear industry in the US despite well over 50 years of nuclear power generation. Yes, Three Mile Island had a partial meltdown, but the total radiation exposure for the residents of the surrounding area as a result of the meltdown was far less than the total radiation exposure of the residents of Colorado during the same period of time. Read some of the books by Bernard L. Cohen.
I've studied this issue up and down, and the only reasonable conclusion is that with proper safeguards, nuclear power is completely harmless and totally efficient. Note that I said proper safeguard. The Chernobyl disaster is an example of not using proper safeguards. In the US, not only do plants have safeguards in place, but due to the irrational fears encouraged by the anti-nuclear activists, the NRC has actually put in place measures that are overly tight.
The regulations in effect are so bureaucratic today that it takes six different people to change a light bulb at some plants. And that's just to change a light bulb. The control room operators get periodic drug tests, and failure of a drug test results in the operator being fired (not to mention that he/she cannot get any other jobs in the industry). Did you know that eating a poppyseed roll even two weeks before a drug test will make you test positive for opium use? Let's think about that one for a minute....
As for the computers, well, that's a different matter. For the obvious reasons, nuclear plants use stable computing systems. One power plant that I know of uses VAX/VMS machines. We've seen the stability tests for Linux and UNIX versus NT. I haven't seen any stability tests for VAX/VMS, but I imagine that such tests would compare with UNIX. Would you believe that some of these plants actually have Windows- and NT-based PC's? Not in any critical application of course....they're usually used to model the fuel dyanmics or monitor the control systems. Wouldn't Linux be a far more reasonable choice for these systems? Probably.
Did you know that the critical systems in the power plant are almost all mechanical? SCRAMing the plant works by closing an electrical current to the control rods, dropping them into the core. And in the US, the control rods are usually above the core, so gravity can be used if something fails. The pumping mechanisms are mechanical, as is the turbine. When computers don't have direct access to the core, suggesting that computer failure would cause a meltdown (let alone the deaths of "millions" of people) is laughable.
Sometimes I think that people on slashdot need to do a little more research before spouting off. While I haven't actually worked in the nuclear industry, I know someone who does, and have watched him work. He explained to me what he does, and has given me a good feel for the way things work. A couple of weeks ago he caught a human error in a company that will go unmentioned. The error had happened at another plant, (notice that human error is almost always the problem) resulting in the termination of a high-level manager. By finding the error himself, this man that I know was rewarded with an unheard-of lump sum of cash (to the tune of double the average cash reward for finding serious problems).
How many ways do I need to say it? The industry is safe. The computers are not a potential problem even if they are running Windows or NT . And people really need to know their ground before spouting off....
Posted by Mike@ABC:
Never underestimate the power of the Linux community. That was, by far, the most popular chat held in the Technology section. Linus was a true sport in the face of nearly 1,000 questions -- 500 of which were posted during the hour he was on. My thanks to all the slashdot folks who stopped by.
And, just to answer the post above about the nuclear power plants....would you prefer they were run on WinNT? Just food for thought.
Some points are overlooked here:
1) GNU is much more than a set of tools. As a proper noun it is the name that Richard Stallman gave to a free operating system, and as an adjective it describes software produced under the auspicies of the GNU Project, who's primary goal the is realization of a free operating system.
A strong case can be made that anyone who has used the GNU tools to put together a free operating system should pay homage to the GNU Project by adopting the GNU name for the operating system.
Now this does not mean that one has to do this, but it is a nice way to honour Richard Stallman's vision of free software. I encourage those who take position, but am not particularly upset if someone doesn't.
The GPL doesn't require it, though RMS may regret this.
2) Linux properly refers to the O/S/ kernel. Calling the entire operating system "Linux" is somewhat of a mis-nomer. Qualifying Linux by the name of the dustributor is still less than satisfactory. This will become more of a problem as the Hurd matures.
Now, personally I think that the only thing that should be called GNU/Linux, without any chance for argument, would be a "Linux" distribution released by the GNU Project, or released with their blessing. Debian GNU/Linux fits this bill.
Furthermore, if the GNU Project does not retain sufficient control over the GNU moniker, it may be used to describe operating systems that are not completely free. This would be a bad thing (and an insult to RMS).
In Liberty, Rene
Maybe this is a good reason: so the community won't go through crippling emotional spasms if Linus has to give up control over the kernel for some reason.
Similarly, so the rest of the world doesn't think that Linux is going to take a major downturn once Linus is out of it. Confidence in Linux could drop massively - people aren't inclined to go with a computing solution that doesn't seem to have a future.
In practically every major press report about Linux it's described as "the operating system created by Linus Torvalds" - I think this is bad not just for the reasons above, but because it's horribly inaccurate and unfair on the hundreds of others who have contributed. Remember that Linux has one of the largest kernel dev teams in the world and this is one of its strengths. I think the idea that it was developed by one guy in his spare time does more harm than good. (Though it's also good that Linus gets the credit he deserves for starting and managing the whole thing.)
while I was reading the article I came acrose the some of the addresses (that are behind the name of the poster) of the posters:
most are just IP names,
but some are IP addresses, see what I found:
'Mike Armentano from [205.227.43.11]:
Are you still working with Transmeta... and so on'
Name: inet-fw1-o.oracle.com
Address: 205.227.43.11
'The Keeper from [32.97.136.234], at 2:53pm ET
Java seems to be the latest technology that is taking the computer industry by storm. Is there a port of the Java VM and JDK to the Linux platform and where can I get it? How do you feel that Java can benefit Linux and vice versa?'
Name: ss04.nc.us.ibm.com
Address: 32.97.136.234
maybe more interresting:
'ndex from [131.107.3.76], at 2:42pm ET
One of the biggest drawbacks to having a linux only desktop is that there aren't many supported games. This is keeping a lot of people on Win9x whether they like it or not. Are you planning to do anything to promote or facilitate the porting of games to Linux?'
Name: tide76.microsoft.com
Address: 131.107.3.76
(but, they are probably not talking officially,
but then again, the microsoft guy... I dunno, seems like FUD to me (-; ).
New things are always on the horizon
There are some interesting questions and answers on the topic of Linus' position as 'posterboy' for Linux and his role as keeper of the kernel.
What would really happen if (God forbid, crosses himself) Linus got killed in a car crash tomorrow, or some other circumstance arose that precluded him from continuing to manage the kernel?
Sure, Linus says there are lots of people who are capable of overseeing the coordination of kernel development. He names Alan and David as people whom he trusts and I don't think many would argue that they aren't capable of doing the job or that they wouldn't have the respect in the community that Linus enjoys.
But I think the more interesting question might be the non-technical aspects to Linus' leadership and how those shoes could be filled in his sudden absence. Linus makes mention of the 'personality cult' that must go. At first my reaction was "why?", Linus is a popular figure and is admired, nay, revered by the community. How could this be an undesirable thing?
Maybe this is a good reason: so the community won't go through crippling emotional spasms if Linus has to give up control over the kernel for some reason.
What do you think? How would the community cope with the sudden absence of Linus Torvalds?
Just out of interest, did you actually read the interview? Because you're taking the midwife thing out of context.
Linus wasn't calling the "GNU System" the midwife, he was talking about gcc. He was saying that having a good compiler available helped the system to grow... making "midwife" quite a good analogy IMO.
Um, no. RMS calls the kernel Linux, and the entire system GNU/Linux.
Meept or somebody calls it LiGnuX.
And because of BSD/Linux and GNU/Hurd, we may soon have practical reasons to call it GNU/Linux.