Slashdot Mirror

Linux has had some sort of SMP Support since version 2.0, and although this has had it's problems along the way - things are well on their way to improvement. Perhaps it is time for an update on how SMP improvements have progressed.

Linux threads are fine. The following letter may be a good thing to point to. And here is a follow-up.

The basic story is that Linux, unlike most operating systems, does not have a hard distinction between a thread and a process. Instead it has the idea of a context of execution, and with the clone() call a context of execution can reproduce a copy of itself, and decide how much is copied. The last is important, clone() can be anything from spawning another thread to a traditional fork.

This threading model is somewhat different than what most operating systems provide, but it is quite sufficient to provide a full POSIX thread implementation with very good speed on a context-switch between threads or processes. (In fact Linux does a faster context switch between processes than most operating systems do switches between threads on the same hardware.)

By contrast NT has a different problem. NT has tremendous difficulty with processes. The time to create a process is abysmal. Context switches are not cheap. And once you start paging, the paging algorithm has beeen found to literally worse than straight chance!

However NT has a pretty good time on context switches between threads and (on paper) some nice specs for working with them. But NT's threading model is somewhat different from the POSIX model and anyone who is experienced with Microsoft knows that what is on paper and what really happens are not always the same...

Cheers,
Ben Tilly

You probably read something like this misinterpretation of the evidence.

If you actually go and read the articles you will find that what they found evidence for is that all humans have mitochondria that trace back to a single individual several hundred thousand years ago. That means that if you trace us all back on direct matrilineal descent (mother to mother to mother to...) you will eventually arrive at a single person.

What they don't mention is that the scientists expected to find that. Think of direct matrilineal descent as being a bush that constantly branches (women have daughters) and gets pruned (some women have no daughters). Starting from a specific point in time, all that that says is that all of the other branches existing at that point have since been completely pruned. This could happen pretty easily by chance, particularly if you started with a small population that was successful and spread out and replaced other groups.

If this really contradicted evolution, then you would think that courses discussing it might be a little more worried than they are, wouldn't you...

Cheers,
Ben

The last time I checked NPL gave way too much control over to Netscape/AOL to be a seriously considered as a non-profit, volunteer based effort (and I sure wouldn't contribute to any other kind,) now I see that they have an MPL in response to the criticism they received.

Yes, the NPL is indeed another one of those sneaky bastard "open source" licenses.. However, it was finalized sometime in 1998, and I don't believe AOL bought Netscape until 1999, IIRC (I don't know, do I? =P). I'm not too excited about the MPL either, however, although it is better than the NPL.. Being somewhat of a die-hard GPL'er, I'm annoyed with the MPL's incompatibility with it..

Could anyone more proficient in legalese comment on MPL, and the differences between it and L/GPL?

You may want to read On the Netscape Public License by Richard Stallman. It covers the differences between the NPL, MPL, and GPL. The GNU Project Web site is also an excellent source of information with regards to free software in general, the GNU philosophy, and the history of the movement. ;) Ok, so, enough shameless plugs..

And whatever the answer, could someone just please start a GPL browser project? Or is that being worked on already?

There are a couple well-known ones.. And a few others that I can't think of right now.. The first, and most obvious, would be Lynx, but I rather doubt that is what you are looking for.. On the other hand, there is Emacs/W3, which you may find to be of a little more interest.

I couldn't get into the Times, so I went and read the original paper (thanks to Tim Moore for the link).

There are a number of points that stand out:

The technology got in the way of the process, causing frustration and time wasting:

The course providers expected the students to be more net-savvy than they actually were. (Remember, this was 2 years ago.) They neglected to provide either enough instruction or the resources to get answers to support questions. The providers also failed to take into account the limits imposed by slow network connections and the use of e-mail to handle threaded discussions.

What was being taught didn't match the materials available:

The course was taught by a substitute, a graduate student who had audited the course the previous summer and who stepped in after the professor who had designed the course became ill. "It was her first time teaching in the U.S., and by distance education." The original professor had built a web site that included course activities and reading assignments, but the substitute modified these by e-mailing the changes to the students each week. Students complained of dead links on the resource webpages and assignments that didn't seem to match the examples provided.

A multitude of communication problems:

Communication seemed to be strictly limited to e-mail and other written modes (didn't anyone have a telephone?). Slow network links caused problems when the class "visited" a text-based MUD. The instructor and several of the students were not originally English speakers.

"Pedagogical Disconnect":

Students complained about being unable to "know exactly what the instructor want[ed]", and the study's authors say that "instructions could be interpreted in many different ways." It's obvious to me that the instructor was attempting to provide a flexible learning environment, but the students wanted a more exact path to getting the grade.

So what we have is a course with unclear goals, taught by someone who didn't design it and who was unfamiliar with the techniques of distance learning, to students who were unprepared to use the technology, over an inadequate system. It's a wonder to me that anyone got anything out of it.

There's a link to the actual paper at the bottom of the article.

Alan posted a summary of the 2.2.13pre5 patches on the Linux Kernel mailing List last thursday. MandrakeSoft might have considered this prerelease stable enough to include it in their 6.1 tree. They might also be holding their press annoucement until the final 2.2.13 is out and they can rpm-ize it and load it on their ftp site.

I've never installed a pre-xxx kernel before, and I wonder if any of you who did it had experience problems with software that test the minor version number of the hosting kernel. I guess a program based on atoi("13pre5") will not be too happy about the whole idea.

I mean, it just popped up at the top of my /. page when I came to read this afternoon.

Does this have the same selection criteria as regular moderation, or is every (logged-in) /.er seeing this? Every day? Only for a testing period?

Also, as I've just been reading what Hofstadter has to say about recursion and meta-stuff in Godel, Escher, Bach: an Eternal Golden Braid (meta-genies in his case, Chapter V and thereabouts), how long before there's meta-meta-moderation, and so forth?

How deeply into this self-checking are we going to go before the moderation is really no more effectual than the posting? I mean, how many people actually read any given /. comment? Right now, very few. How many will see the effect of a given moderation? A whole lot more... right now. But for how long?

On a separate-but-related note, when I explained /.'s (fairly) new moderation system to a friend, he commented, "That's sort of oddly big-brother, isn't it?"

My only response was "Well, yeah, in a publicly-held sort of way."

Hrm.

I know, we could start an anti-Microsoft (liberal) party! To what extent we allow Microsoft to influence the technology would be a key issue in a cyber government.

But seriously, I don't think they have given much thought to the idea of how government works. Is this serious or is it a government in the same way that Discordianism is a "religion"? How does it go again? : ...among these are Life, Liberty, and the pursuit of Happiness. To secure these rights, Governments are instituted among Men, deriving their just powers from the consent of the governed. Declaration of Independence
This thing won't do that. Nor will it: form a more perfect Union, establish Justice, insure domestic Tranquility, provide for the common defence, promote the general Welfare, and secure the Blessings of Liberty to ourselves and our Posterity

The version that I pointed to may be the latest version, which raises an interesting problem for historians. Because of the speed and ease of which documents can change on the internet, it is nearly impossible to find the "first edition" of a given document.

Stallman's EMACS was brilliant in the 1970s, but today we demand more, specifically Microsoft Word, which can't be written over a weekend, no matter how much Coke you drink.
- Bob Metcalfe

• Emacs runs on pretty much any hardware that can run [Free|Net|Open]BSD, Solaris, SunOS, Ultrix, or Linux.
• Emacs has a free (of course) API called Emacs LISP (or elisp, for short) enabling you to write pretty much any extension you want. Even, as the homepage notes, a web browser that runs inside emacs.
• Emacs has interactive/context sensitive modes for editing a wide variety of documents, including HTML, Lisp, C++, Prolog...

The Constitution :
"We the people of the United States, in order to form a more perfect union..." (you can hear the School House Rock song can't you?)

Declaration of Indep. :
"When in the course of human events...."

It's easy to screw these two up. Anyone remember the print ads for the movie Jefferson in Paris from ~1995? They used "We the people.." in the image, even though Jefferson had nothing to do with the Constitution.

http://linux.u cs.indiana.edu/hypermail/linux/kernel/9812.1/0135. html

It is /proc/cpuinfo and dmesg output from a 14 cpu (6927 bogomips) sparc server with 2gigs of ram and a huge chain of scsi disks. There has also been talk of Sun giving kernel hackers access to bigger systems.

My point? Find a system or a market segment that linux doesn't support, then wait a year or so and I bet you it will support it.

Does anyone seriously think that Monterey is going to be any different than SVR4 or OSF/1?

"people don't need support for Real G2?" It is to laugh. After running a very popular streaming media site for three years, I can tell you *categorically* that the vast majority of users are confused and scared by RealPlayer. Especially the part about downloading and installing it. But even the part about just *using* it throws them.

And people--even I--definitely need support for the encoders and production tools.

Check out FreeExpression, by the way, the new OSS streaming media initiative:

FreeExpression

It's not being used on freespeech.org yet--because it doesn't really work yet--but it will, and it will be.

Joey
www.freespeech.org