Zapping individual neurons won't get you anywhere. (Well, not anywhere useful.) The idea that specific thoughts (or even specific types of thought) are tied to specific neurons (or even specific portions of the cerebrum) has been rather thoroughly debunked. Just about the only thing we do know about the higher levels of how the brain works is that it _doesn't_ divide work up by physical area. Anything you think about, neurons all over your whole brain are involved with it. Beyond that, we're not really sure yet how it works.
We do understand the low levels of how signals are passed from one neuron to another, chemically. But how all that adds up to thought, we don't know yet.
If you take an introductory-level psych course, you'll learn about various major theories of past psychologists about how the brain works. Most of them, up until about the middle of the twentieth century, believed that different areas of the brain are responsible for different types of thought and that individual memories are stored in separate locations. This was the prevailing view, because it seems obvious -- but you'll also learn about the various events (most of them involving accidents of some kind, like the guy who got a metal rod shot into his head and lived, brain tumors, et cetera) that lead gradually toward a near-universal rejection of those ideas. To all appearances, the brain does NOT work that way, despite what many scientists used to think.
Oh, some also would say religion. The press, they would say, was developed to print the Bible. (This is at least partly true.) They would also point to the crusades as a major cause of the development of a lot of technology in Europe.
I personally disagree with this assessment, though it has some validity. However, I believe that the Bible wasn't the *only* thing Gutenberg wanted to print. He printed it first because it was the best-known and most-revered book, but he wanted to print books in general, not _only_ the Bible. As for the crusades, religion was certainly abused to talk many people into them, but I believe they were organised for political, not religious, reasons. Besides, it is an historical coincidence (if there is such a thing) that the crusades drove more tech development than other wars. It is only because they got the warriors to physical locations they hadn't travelled to from Europe recently that the crusades are significant. A religious war between, say, England and France would not have driven any more technological development than any other war. So it was new-found knowledge of a foreign place that drove the new technology -- communication, or transportation, or both. And, of course, war.
I wish I knew where this idea came from so I could debunk it properly. The military I'll grant; war has always been a driving force of technology. The other two I question. Games are driving new technology now, and have been for thirty years or so, but historically that's a blip on the radar.
I would propose a different three things (well, two of them different): war, communications, and entertainment. (If you like, you can group porn and games under the umbrella of entertainment, but I believe other forms of entertainment such as hunting and racing and even literature have been just as responsible for driving technological development as games or porn.) Some would say space, but that was really mostly driven by the miliary, especially at first. Some also would say transportation, which is arguable, but I contend that usually one form or another of communication was driving the development of newer and better forms of transportation, at least until the last century or so. At minimum, transportation and communication have fed off one another, as there's little compelling reason to go somewhere if you don't have information about that place.
Think about communications, though. Writing. Paper (both early forms such as parchment and papyrus as well as modern paper). The printing press. Movable type. Telegraph and phone lines. Radio. Ethernet. TCP/IP. These are the technologies that really matter, the enabling technologies, the ones on which everything else is based, the ones without which none of the others would have happened.
These big three (war, communications, and entertainment) all also feed off one another. Gunpowder was used in fireworks, later weapons. Rapid-loading guns are only possible due to the self-contained cartridge, which is based on the percussion cap developed for... hunting -- a sport, a kind of entertainment. On the other hand, intercontinental communication was driven by better ships, driven by naval needs, for the military -- war. And if you've ever seen Wag the Dog, you'll either chuckle about the notion of entertainment driving war, or shudder and nod, depending on your philosophical position. Either way, it is difficult to argue that any one of these three have not been a major driving force behind the development of technology.
Transportation, as I said, also is arguable. Gaming, in recent years, is true, but I contend that fits under the broader category of entertainment. Maybe I'm naive, but I can't think of any important technology that has been driven to development by porn. Unless you want to say bandwidth, but I'd call that communications, and anyway, I'd say that bandwidth has been driven to a larger extent by games (entertainment) and shopping (entertainment), the military, and, especially in the early years, plain old garden variety communications, like email and usenet.
> You are also assuming that nobody will be lured to write a patch for an > unsolved vulnerability by the thought of large piles of cash, which is > obviously incorrect.
It's also obviously irrelevant. The entity supplying the service will collect the fees whether any given specific patch is released in a timely fashion or not; the person(s) creating the patch has no way to collect any of that money.
If the software is open, the problem will be patched by someone who is motivated to create the patch because they use the software and don't want to be vulnerable. Yes, there are people who might be motivated by the money, but not with the same panicked this-has-to-be-done-YESTERDAY motivation that will possess the admin of a mission-critical system left open by a vulnerability. Also, the people who are most familiar with the software (and therefore best able to patch the issue) are the people who already work with the source code on a regular basis -- so (in the case of open-source stuff) they're obviously already motivated to work on the software in question.
With proprietary software, the only people who could fix the problem would be the employees who are already paid to work on the software, and they'll do it when (or if) management says so. Any money that the company collected from a security patch service would almost surely not make it into the pocket of the employee who fixed the problem. It would be *possible* for a company to have a program that fed bonuses to employees who come out with the first viable fix for vulnerabilities, but it is by no means a foregone conclusion that this would be any more likely to be the case if fees were charged to the end users for an security update service.
Anyway, any possible benefit in reduced response times would be far outweighed by the many, many systems that would go unpatched due to lack of funds for the security update service. The obvious solution to this (for commercial software) is to include an n-year subscription with the price of the software.
For open-source software, I'm thinking a bittorrent-like solution to _reduce_ the bandwidth cost of the update server is preferable to a subscription fee to offset the cost. Incremental binary patches from one specific version to another might potentially be a bandwidth-saving option for some distros, but the bittorrent solution is probably even better and probably also more robust.
Look in the yellow pages under "Salvage". Where I live, there's a little place called CTR just up the road in Crestline, about a five minute drive from here. There may be something similar near where you live.
Anyone can have a bowl, but if you add anything to your bowl, you have to add it back into the main pot as well. So, if you like cheese in your chili and add cheese to yours, you have to put your cheese back into the pot.
There's quite a lot of stuff in the Linux Chili these days. Tomatoes, rice, corn, cheddar, cojack, motzarella, and several other kinds of cheese, at least twenty different kinds of beans and fifteen kinds of peppers (even some Hatari), beef, pork, mutton, vennison, perch,...
> Isn't it funny that nevertheless Microsoft marketing has brainwashed the > masses to the point that they actually believe that WinXP has become more > secure than Win9x?
It is in many respects more *securable*. (More than Win9x, I mean.)
Yes, it's horrible (security-wise) out of the box. NT was, until Marketing got ahold of it, not really designed to be used in its out of the box state. It needs configuring. It was designed to be deployed by an IT staff across desktops in a company or organization -- for that, it doesn't really need to be secure as such out of the box per se (though it would've been nice), or to be installed by OEMs on off-the-shelf systems. OEMs are *supposed* to do that configuring, or some of it, for you. Theoretically. Instead, some of them choose to install a metric butload of cheesy dross, but for that you have to blame the OEM.
As far as non-IT-professional consumers buying boxed copies of WinXP and installing it themselves, well, that's a result of Marketing; it wasn't designed for that. It needs to be installed and configured by a professional.
Of course, in the real world there are hundreds of thousands of WinXP boxes out there that have not been secured at *all*. This is bad. We can only hope Microsoft learned from it and will create a more secure OOTB default configuration for Longhorn desktop. It is IMO noteworthy that earlier versions of NT, due to who bought and used it, did not create the kind of nightmares we've seen with XP, even though in principle they were no more secure OOTB than XP is. Yeah, there were security issues, but nothing like this past fall, in terms of impact on the whole world.
Still, WinXP is not nearly the worst product on the market, security-wise. (That dubious honour clearly belongs to Outlook.)
> I've applied gaussian blurs on images in BeOS (obviously not in photoshop, > but in natively written image editting programs in BeOS), and at most it > has taken a minute and a half to apply the effect.
His point wasn't that this particular operation is necessarily slow, but that it *can* be slower than on other systems without bothering the typical desktop user as much as, say, a delay before display when unminimizing a window. IOW, he was saying that the user's perception of performance is based mostly on things that the BeOS handled well. I tend to agree, though other systems are catching up these days. The most useful feature of BeOS that I still haven't seen in another system is the ability to have different resolutions and color depths on different virtual desktops ("workspaces", the BeOS folks call them). The pervasive multithreading has made its way into other systems, and the abuse of filesystem attributes (e.g., using zero-byte files to store significant amounts of information) was, despite being genuinely innovative, not terribly useful.
Depends. If you're a gamer, 2.6 has better realtime support (than 2.4) for sound and such. It also has SATA support, which only matters if you have any SATA drives. And some other things. ISTR that the O(1) scheduler went in, which could be helpful performance-wise for the computing cluster if it tends to have a lot of processes running at once.
*shrug*. Don't throw out your 2.4 kernels yet. 2.6.0 ends in.0, so obviously some kinks will be worked out over the coming months. But if you've got a non-critical system you can afford to test it out on, such as a desktop, this is a good time to do that. If it breaks things, you can always go back to 2.4.x for a while.
> Keep in mind that all the Pixar movies use rasterization techniques, not > raytracing or radiosity.
This is due to the number of frames they have to render and the time they have to do it in and the hardware they have to do it on. Raytracing takes more time, and they have a deadline to meet if they want to come in under budget. As hardware gets faster and cheaper, you'll *eventually* see things raytraced more and more. I once tried to calculate approximately when good PC hardware would be fast enough to raytrace a 3D FPS game in realtime. I had to make some assumptions, and came out with something like 2020, based on Moore's law, assumptions about increasing expectations of resolution, and some other factors I don't recall now. It's a very rough guess. Also, it assumes that raytracing technology will improve between now and then and that for a FPS game you'd drop some of the more CPU-intensive stuff for performance reasons. (For example, if you had reflective/refractive objects, you wouldn't put them in a room with area lights, nor would you put either of those things in a room with atmospheric effects. Also you'd want to limit the number of light sources in any given room.)
Right now, a good high-end PC can render a frame every several hours if it's complex or every several minutes if it's fairly simple.
For still images, though, nothing beats raytracing except _perhaps_ real photography of real objects (or detailed models), maybe not even that (as it's harder to get exactly the lighting you want with real photography), and the same will be true for animation when the rendering hardware is up to it.
> Can anyone explain what the point of doing raytracing is over quicker > better methods?
There are only quicker methods; there are no known _better_ methods. As for raytracing's advantages, it has three major advantages over faster methods:
1. It looks better. 2. It looks better. 3. It looks better.
If you compare screenshots from 3D games versus raytraced images, the difference is stunning. Good raytraced images *almost* look like photographs. Heck, if you set up the raytracer's virtual camera with a slight blur on the lens, you might be able to pass them off as amateur photos.
Do yourself a favour. Go to the public library, get the books unabbridged
on audiocassette, and listen to them whenever you're in the car for more
than a few moments. Our local public library here
hasthisset
which will do nicely. (The narration is great, except when he tries to sing the songs, which he shouldn't have done, but that's me being picky.)
The books are much better than the movies. Yeah, yeah,
the movies are great. Read the books, and you'll see what I mean.
They're better. Getting them on audio cuts down on the amount of time
you have to devote to it, since you can make double use of time when
your eyes are tied up, such as while driving, or when your hands are
tied up, such as while eating.
The UN is only impotent when it can't agree. Such as in this case. We had several major nations backing us (not least Brittain), and that was enough to prevent the UN from stopping us from doing what we determined we needed to do.
If the UN (excluding the US of course) ever _agrees_ that the US has to get out of Iraq, then for political reasons we'll pretty much have to do that. All the major world powers feel this political pressure. The nations that don't feel it are podunk third-world nations like Iraq -- those are the nations that defy UN sanctions, and yes, somebody (like the US for example) has to step in and do something militarily in those cases. But the larger nations are held in check by the political pressure that the UN does weild.
> Osama bin Laden killed 3000 people in the US. > > Saddam Hussein killed none.
You can't really be enough of an imbecile to believe that killing people in the US is the only measure of a threat, can you? Hussein had been building a war machine in Iraq for two decades. He was an opportunistic tyrant, looking for chances to consolidate power. Some time back, if you recall, he made an attempt at a small neighboring nation. If he hadn't been driven back then, he would have taken a second neighbor and a third by this time. If he hadn't been watched carefully since, he'd have made a second attempt at Kuwait and probably taken it. The man was a serious threat to the political stability (such as it is) of the middle east -- a dangerous and heavily-contested region politically, as much a powder keg as any acreage on the globe, just as likely as the balkans to be the start of a world war.
bin Laden was a threat to the peace of mind of people in the US, and perhaps the lives of a few. Hussein was a threat to humanity.
That is, of course, not what he'll be tried for. Which is fine; it doesn't need to be. As far as any danger to the world goes, we could let him stand trial at the Hague, probably get off easy, and live out his life comfortably in Europe; as long as he doesn't return to power in the middle east, the threat is cared for. What we _can't_ do is put him before a US court and string him up in the fashion he deserves; that would be very policitally bad for the US, and I think Bush knows better. So what we'll probably do is try him in Baghdad for crimes against the Iraqi people. I'm not sure the outcome matters very much, as long as he never comes to power again.
The hard thing is going to be building Iraq up to the point where the next person won't be just as bad or worse. We really need to do approximately what we did in Japan, after WWII, but I don't know if the Bush administration will have enough years in power (even _with_ another four, which is not a thing to take for granted) to make it happen, particularly the way they seem to have been dawdling about it. I would have thought that by now they ought to have districts drawn in a fashion that guarantees enough representatives to at least four major factions that none of them can get jack diddly squat pushed through without support from the other factions, set up an infrastructure for a provisional government, and set a date for the first election. I mean, sure, if you're going to make a first-world nation out of Iraq, you're going to need to install power and phone lines and all that jazz, yeah, but that stuff takes a while; getting a provisional government set up would earn some nice political capital and maybe buy some time to make the rest of it happen.
Then again, depending on how the Hussein trial is conducted, that might buy some political capital and some time too. Still, in Bush's position, I would sure have been a lot quicker about drawing up districts for representation and stuff, the beginnings of a provisional government. It's not like it couldn't be done at the same time as other stuff. Also key is that if you can get a provisional government going, US and UN forces can be used to make sure it stays in power; that doesn't cost nearly as much politically as just running the country directly, like we are basically doing now.
> I believe this is where access falls short. I don't believe the syntax of > access is 100% standard.
Last I knew, no available database has 100% standard syntax. They all differ from the standard in a number of areas.
The problem with Access is that it's becomming obscure. It's not included with most versions of MS Office anymore, for one reason or another (probably because MS wanted to drive sales of their _other_ database offering, SQL Server), and so consequently few desktops have it, so nobody knows it. SQL is used heavily on servers and in backend stuff, and so pretty much every database administrator knows one dialect or another of SQL.
If you're teaching people who are going to work mostly with desktop and office stuff, Access might be appropriate, though frankly the database in MS Works is more widely distributed on the whole, simpler, better known, and adequate for basic office-type needs. If OTOH you're teaching people who are planning to work in IT and administer networks and servers, they're going to have to know SQL in one form or another. This will be true even if they work in an all Microsoft-only environment -- for serious databases MS strongly pushes their SQL Server product. If they work in a heterogenous environment, they'll use Oracle (if the database is mission-critical and the place is rolling in cash) or MySQL or PostgreSQL (for smaller installations) or _potentially_ MS SQL Server -- but the chances of seeing Access in a heterogenous environment are roughly zero. In any non-MS homogenous environment (e.g., everything comes from one specific Unix vendor) the chances of using Access for anything serious are *exactly* zero.
The question then is which SQL implementation to have them know. I would suggest picking one of the four covered by SQL in a Nutshell, and make sure they're aware that there are differences in the other implementations.
Which one to pick? Where are they looking to work? All-Microsoft shops? Teach 'em MS SQL Server. Big enterprises? Oracle. Small businesses with a heterogenous environment? MySQL probably, or perhaps PostgreSQL. If you don't know, just pick one. Knowing one dialect of SQL will make it easy for them to pick up another, so it's not wasted. There are differencesin the details, but the *concepts* are to a large extent the same, at least for the basic stuff you're going to teach in a course.
> It is a full fledged programming language though.
Right, that was my point. Its *purpose* may be page description (a form of markup), but while it is a special-purpose language (and therefore not really a general-purpose language) it nevertheless is a programming language.
The same can be said of Inform (though its purpose is quite different). Zarf once wrote a (small, unpowerful, proof-of-concept-type) lisp interpreter in Inform, for (I assume) hack value, and other things have been done similarly in it (a tetris game, for example), but it is primarily used for just one thing, interactive fiction. Still, there's no question in anybody's mind about whether Inform is a programming language. It's not a *general-purpose* programming language (mostly because the platforms it compiles to don't have good general-purpose I/O, for portability reasons), but it's a programming language -- a special-purpose one. I'd put PostScript in the same category.
As a side note, Zarf once threatened (in jest, I think) to do a work of interactive fiction in PostScript.
If your house is 60F (did you typo that?), everything is working exactly as it should. Maybe if the temperature were _less_ than room temperature, you might start to get concerned, but 60 is the highest you would ever want it to be (unless you're running a nursing home full of little old ladies who eat like birds and so get the shivers if the room drops below body temperature). We set our air conditioner lower than that in the summer.
> Funny aside -- PostScript's processing instructions would probably make it > a good exception to these examples -- it probably _is_ Turing equivalent
I fail to see how that makes it an exception. Sure, PostScript is powerful enough to be classified as a programming language. But it's not really a general-purpose language. So then, it's in the same category with TADS and POV -- a niche programming language, suited for a particular purpose.
I would not classify PostScript as a markup language merely because it is capable of formatting content. Perl has extensive support for formats, and as far as that goes even C has printf, but we don't call them markup languages, because they support variables and flow control, which makes them programming languages. HTML (as such) supports neither, so it's obviously not a programming language. (There is of course ECMA script... but that is generally considered a separate language from HTML, close association notwithstanding.)
I suppose one could argue that PostScript is a markup language because its primary intended purpose is to format content, but IMO that argument falls flat unless you want to consider ECMA script and PHP as markup languages also, which seems clearly wrong to me. Markup languages don't have flow control and variables and conditionals and so on and so forth.
> the ability to have different resolutions on different virtual desktops
BeOS has this (different colour depths too, if you want), and I can tell you, it's pretty nifty. Currently I don't use the multiple-desktops feature of X, because I don't much need to see my wallpaper very often, so having a lot of Windows in front of eachother is okay, and the task list in the Gnome panel is pretty good for keeping track of a lot of windows on one desktop. But I would probably use a second and maybe a third virtual desktop if they could have different resolutions from the first one.
> Actually, there's plenty to say about Windows as a server platform.
Yeah, but he was trying to be *nice*.
> VMS was a cruddy operating system.
From the little experience I have with VMS it seems pretty good to me. Very solid in terms of never going down barring hardware failure. (Unfortunately, hardware failure seems to be depressingly common with the Alpha line. The Vax was a more robust system.)
There are two problems with VMS though. First off, it's weird. By that I mean that there are a lot of things it doesn't do the same way that other systems do them, even when the other major systems (Unix, Windows, Mac) all do them roughly the same way. The second problem with VMS is that it's hard to hire people who know it well.
> Windows NT, being largely derived from VMS
As near as I can determine, this relationship is almost entirely mythological. Early versions of NT ostensibly took some low-level code from MS XENIX, and we're pretty sure that a lot of GUI and API stuff was taken from Windows 3.x, but beyond that it appears to have been mostly new stuff.
If it were even tangentially related to VMS, it would probably be a lot more *different* from the other systems, and a lot more like VMS. If the filesystem for example were taken from VMS, one would expect it to have some of the strange VMS filesystem semantics. (I'm not talking about command line syntax here[1], but stuff like multiple different kinds of text files.) If the kernel were taken from VMS, one would expect NT to have batch queues and other process management features reminiscent of the minicomputer world, but it much more closely resembles Unix. If the networking layer were taken from VMS... err, let's just say it pretty obviously wasn't[2]. In summary, if you've used Unix, NT, and VMS, you start to think of Unix and NT of being fairly similar. If you've used Unix and NT before, VMS will weird you out. It's quite different. Much moreso than Mac, Apple slogans notwithstanding.
The one thing in VMS that's mercifully similar to other systems is the environment variable system.
[1] Though DCL syntax is a little odd, too. But that's a surface thing; the
command-line syntax on any platform can be changed just by installing a
different shell.
[2] The networking stuff in VMS is *weird*. I don't properly understand it,
but it sure as anything is not much like Windows or Unix networking.
> I grow weary of seeing lots of young 20-something applicants fresh out of > school who claim they have excellent coding skills and then proceed to > list about ten different languages including HTML
My resume lists languages in categories according to how well I know them. It says I'm fluent in Perl and lisp and "somewhat familiar" with several other languages. At one point it also said "and have had courses in" yet several more languages, but that's one of the things that's going to get cut next time I have to print the thing and want to get it down to a page, because I've added several technologies (like MySQL) since the last time I did so.
And yeah, I list HTML, and CSS, but not under "programming languages". I list them in under web development, alongside CGI and MySQL.
But I may not be the typical applicant you're talking about; my resume also says I have a little VMS experience...
And I'm gonna have to take MS Office off, because the last time I did anything serious with it was too many versions ago at this point. (Think in terms of a version that ran on Windows 3.1...)
This scales really well to groups of various sizes containing persons of various ages. Just make sure that when you split teams you don't put all of something one one team (e.g., all of the geeks, all of the sports fans, all of the young people, whatever). Split all the demographics across both teams, and it works better.
Here's how to play: Everybody writes names of famous persons on a bunch of little slips of paper, folds them once, and throws them in a big bowl. These can be names of current celebrities, historical figures, literary characters, cartoon characters, whatever, as long as they're sufficiently well-known that there's a decent chance several people in the room know about them.
Then you take turns: a person from the one team, then a person from the other, and then another person from the first team, and so on. You get one minute to see how many you can get, as follows: You draw a slip of paper out, look at it, and then without saying any part of it yourself you must get someone on your team to say the name that's on the paper. If you've never heard of the person, it's too bad: you make a scrunchy face and try to get it some other way. ("Okay, the first name is the same as Ellison's first name, and he's something you make pickles from.") You cannot pass*. When you finish one, you draw another. When time runs out, you put the one you didn't finish back into the bowl without revealing any more about what it was, count how many you got, and add it to your team's score.
This is way more fun than it sounds like. With the right group of people, someone can draw "Marvin K Mooney", "Alan Greenspan", and "Huldrych Zwingly" one after the other. Watching their face can be quite entertaining.
* Exception: In cases of utter illegibility, when you can't make out
the letters at all, a person from the other team can examine the slip
and confirm that it's illegible, and you can skip it.
Slashdot Mirror
Zapping individual neurons won't get you anywhere. (Well, not anywhere useful.)
The idea that specific thoughts (or even specific types of thought) are tied to
specific neurons (or even specific portions of the cerebrum) has been rather
thoroughly debunked. Just about the only thing we do know about the higher
levels of how the brain works is that it _doesn't_ divide work up by physical
area. Anything you think about, neurons all over your whole brain are involved
with it. Beyond that, we're not really sure yet how it works.
We do understand the low levels of how signals are passed from one neuron to
another, chemically. But how all that adds up to thought, we don't know yet.
If you take an introductory-level psych course, you'll learn about various
major theories of past psychologists about how the brain works. Most of them,
up until about the middle of the twentieth century, believed that different
areas of the brain are responsible for different types of thought and that
individual memories are stored in separate locations. This was the prevailing
view, because it seems obvious -- but you'll also learn about the various events
(most of them involving accidents of some kind, like the guy who got a metal
rod shot into his head and lived, brain tumors, et cetera) that lead gradually
toward a near-universal rejection of those ideas. To all appearances, the
brain does NOT work that way, despite what many scientists used to think.
Oh, some also would say religion. The press, they would say, was developed to
print the Bible. (This is at least partly true.) They would also point to the
crusades as a major cause of the development of a lot of technology in Europe.
I personally disagree with this assessment, though it has some validity.
However, I believe that the Bible wasn't the *only* thing Gutenberg wanted to
print. He printed it first because it was the best-known and most-revered book,
but he wanted to print books in general, not _only_ the Bible. As for the
crusades, religion was certainly abused to talk many people into them, but I
believe they were organised for political, not religious, reasons. Besides,
it is an historical coincidence (if there is such a thing) that the crusades
drove more tech development than other wars. It is only because they got the
warriors to physical locations they hadn't travelled to from Europe recently
that the crusades are significant. A religious war between, say, England and
France would not have driven any more technological development than any other
war. So it was new-found knowledge of a foreign place that drove the new
technology -- communication, or transportation, or both. And, of course, war.
> Pornography, military, and gamers.
I wish I knew where this idea came from so I could debunk it properly. The
military I'll grant; war has always been a driving force of technology. The
other two I question. Games are driving new technology now, and have been
for thirty years or so, but historically that's a blip on the radar.
I would propose a different three things (well, two of them different): war,
communications, and entertainment. (If you like, you can group porn and games
under the umbrella of entertainment, but I believe other forms of entertainment
such as hunting and racing and even literature have been just as responsible
for driving technological development as games or porn.) Some would say space,
but that was really mostly driven by the miliary, especially at first. Some
also would say transportation, which is arguable, but I contend that usually
one form or another of communication was driving the development of newer and
better forms of transportation, at least until the last century or so. At
minimum, transportation and communication have fed off one another, as there's
little compelling reason to go somewhere if you don't have information about
that place.
Think about communications, though. Writing. Paper (both early forms such as
parchment and papyrus as well as modern paper). The printing press. Movable
type. Telegraph and phone lines. Radio. Ethernet. TCP/IP. These are the
technologies that really matter, the enabling technologies, the ones on which
everything else is based, the ones without which none of the others would have
happened.
These big three (war, communications, and entertainment) all also feed off one
another. Gunpowder was used in fireworks, later weapons. Rapid-loading guns
are only possible due to the self-contained cartridge, which is based on the
percussion cap developed for... hunting -- a sport, a kind of entertainment.
On the other hand, intercontinental communication was driven by better ships,
driven by naval needs, for the military -- war. And if you've ever seen Wag
the Dog, you'll either chuckle about the notion of entertainment driving war,
or shudder and nod, depending on your philosophical position. Either way, it
is difficult to argue that any one of these three have not been a major driving
force behind the development of technology.
Transportation, as I said, also is arguable. Gaming, in recent years, is true,
but I contend that fits under the broader category of entertainment. Maybe
I'm naive, but I can't think of any important technology that has been driven
to development by porn. Unless you want to say bandwidth, but I'd call that
communications, and anyway, I'd say that bandwidth has been driven to a larger
extent by games (entertainment) and shopping (entertainment), the military,
and, especially in the early years, plain old garden variety communications,
like email and usenet.
> You are also assuming that nobody will be lured to write a patch for an
> unsolved vulnerability by the thought of large piles of cash, which is
> obviously incorrect.
It's also obviously irrelevant. The entity supplying the service will collect
the fees whether any given specific patch is released in a timely fashion or
not; the person(s) creating the patch has no way to collect any of that money.
If the software is open, the problem will be patched by someone who is motivated
to create the patch because they use the software and don't want to be
vulnerable. Yes, there are people who might be motivated by the money, but
not with the same panicked this-has-to-be-done-YESTERDAY motivation that will
possess the admin of a mission-critical system left open by a vulnerability.
Also, the people who are most familiar with the software (and therefore best
able to patch the issue) are the people who already work with the source code
on a regular basis -- so (in the case of open-source stuff) they're obviously
already motivated to work on the software in question.
With proprietary software, the only people who could fix the problem would be
the employees who are already paid to work on the software, and they'll do it
when (or if) management says so. Any money that the company collected from a
security patch service would almost surely not make it into the pocket of the
employee who fixed the problem. It would be *possible* for a company to have
a program that fed bonuses to employees who come out with the first viable fix
for vulnerabilities, but it is by no means a foregone conclusion that this
would be any more likely to be the case if fees were charged to the end users
for an security update service.
Anyway, any possible benefit in reduced response times would be far outweighed
by the many, many systems that would go unpatched due to lack of funds for the
security update service. The obvious solution to this (for commercial
software) is to include an n-year subscription with the price of the software.
For open-source software, I'm thinking a bittorrent-like solution to _reduce_
the bandwidth cost of the update server is preferable to a subscription fee to
offset the cost. Incremental binary patches from one specific version to
another might potentially be a bandwidth-saving option for some distros, but
the bittorrent solution is probably even better and probably also more robust.
Look in the yellow pages under "Salvage". Where I live, there's a little place
called CTR just up the road in Crestline, about a five minute drive from here.
There may be something similar near where you live.
> What the hell is a Linux Chili?
...
Anyone can have a bowl, but if you add anything to your bowl, you have to add
it back into the main pot as well. So, if you like cheese in your chili and
add cheese to yours, you have to put your cheese back into the pot.
There's quite a lot of stuff in the Linux Chili these days. Tomatoes, rice,
corn, cheddar, cojack, motzarella, and several other kinds of cheese, at least
twenty different kinds of beans and fifteen kinds of peppers (even some Hatari),
beef, pork, mutton, vennison, perch,
> Isn't it funny that nevertheless Microsoft marketing has brainwashed the
> masses to the point that they actually believe that WinXP has become more
> secure than Win9x?
It is in many respects more *securable*. (More than Win9x, I mean.)
Yes, it's horrible (security-wise) out of the box. NT was, until Marketing
got ahold of it, not really designed to be used in its out of the box state.
It needs configuring. It was designed to be deployed by an IT staff across
desktops in a company or organization -- for that, it doesn't really need to
be secure as such out of the box per se (though it would've been nice), or to
be installed by OEMs on off-the-shelf systems. OEMs are *supposed* to do
that configuring, or some of it, for you. Theoretically. Instead, some of
them choose to install a metric butload of cheesy dross, but for that you have
to blame the OEM.
As far as non-IT-professional consumers buying boxed copies of WinXP and
installing it themselves, well, that's a result of Marketing; it wasn't
designed for that. It needs to be installed and configured by a professional.
Of course, in the real world there are hundreds of thousands of WinXP boxes
out there that have not been secured at *all*. This is bad. We can only
hope Microsoft learned from it and will create a more secure OOTB default
configuration for Longhorn desktop. It is IMO noteworthy that earlier
versions of NT, due to who bought and used it, did not create the kind of
nightmares we've seen with XP, even though in principle they were no more
secure OOTB than XP is. Yeah, there were security issues, but nothing like
this past fall, in terms of impact on the whole world.
Still, WinXP is not nearly the worst product on the market, security-wise.
(That dubious honour clearly belongs to Outlook.)
> I've applied gaussian blurs on images in BeOS (obviously not in photoshop,
> but in natively written image editting programs in BeOS), and at most it
> has taken a minute and a half to apply the effect.
His point wasn't that this particular operation is necessarily slow, but that
it *can* be slower than on other systems without bothering the typical desktop
user as much as, say, a delay before display when unminimizing a window. IOW,
he was saying that the user's perception of performance is based mostly on
things that the BeOS handled well. I tend to agree, though other systems are
catching up these days. The most useful feature of BeOS that I still haven't
seen in another system is the ability to have different resolutions and color
depths on different virtual desktops ("workspaces", the BeOS folks call them).
The pervasive multithreading has made its way into other systems, and the
abuse of filesystem attributes (e.g., using zero-byte files to store significant
amounts of information) was, despite being genuinely innovative, not terribly
useful.
> I honestly want to know.
.0, so obviously
Depends. If you're a gamer, 2.6 has better realtime support (than 2.4) for
sound and such. It also has SATA support, which only matters if you have
any SATA drives. And some other things. ISTR that the O(1) scheduler went
in, which could be helpful performance-wise for the computing cluster if it
tends to have a lot of processes running at once.
*shrug*. Don't throw out your 2.4 kernels yet. 2.6.0 ends in
some kinks will be worked out over the coming months. But if you've got a
non-critical system you can afford to test it out on, such as a desktop, this
is a good time to do that. If it breaks things, you can always go back to
2.4.x for a while.
> Keep in mind that all the Pixar movies use rasterization techniques, not
> raytracing or radiosity.
This is due to the number of frames they have to render and the time they
have to do it in and the hardware they have to do it on. Raytracing takes
more time, and they have a deadline to meet if they want to come in under
budget. As hardware gets faster and cheaper, you'll *eventually* see things
raytraced more and more. I once tried to calculate approximately when good
PC hardware would be fast enough to raytrace a 3D FPS game in realtime. I
had to make some assumptions, and came out with something like 2020, based
on Moore's law, assumptions about increasing expectations of resolution,
and some other factors I don't recall now. It's a very rough guess. Also,
it assumes that raytracing technology will improve between now and then and
that for a FPS game you'd drop some of the more CPU-intensive stuff for
performance reasons. (For example, if you had reflective/refractive objects,
you wouldn't put them in a room with area lights, nor would you put either
of those things in a room with atmospheric effects. Also you'd want to
limit the number of light sources in any given room.)
Right now, a good high-end PC can render a frame every several hours if
it's complex or every several minutes if it's fairly simple.
For still images, though, nothing beats raytracing except _perhaps_ real
photography of real objects (or detailed models), maybe not even that (as
it's harder to get exactly the lighting you want with real photography), and
the same will be true for animation when the rendering hardware is up to it.
> Can anyone explain what the point of doing raytracing is over quicker
> better methods?
There are only quicker methods; there are no known _better_ methods. As for
raytracing's advantages, it has three major advantages over faster methods:
1. It looks better.
2. It looks better.
3. It looks better.
If you compare screenshots from 3D games versus raytraced images, the difference
is stunning. Good raytraced images *almost* look like photographs. Heck, if
you set up the raytracer's virtual camera with a slight blur on the lens, you
might be able to pass them off as amateur photos.
For some sample raytraced shots, see www.irtc.org
> I don't understand why everyone assumes that people who work with computers
> like fantasy.
I don't particularly like fantasy in general, but LOTR is special. It's of
much higher quality than most fantasy.
Do yourself a favour. Go to the public library, get the books unabbridged on audiocassette, and listen to them whenever you're in the car for more than a few moments. Our local public library here has this set which will do nicely. (The narration is great, except when he tries to sing the songs, which he shouldn't have done, but that's me being picky.)
The books are much better than the movies. Yeah, yeah, the movies are great. Read the books, and you'll see what I mean. They're better. Getting them on audio cuts down on the amount of time you have to devote to it, since you can make double use of time when your eyes are tied up, such as while driving, or when your hands are tied up, such as while eating.
> the U.N. is impotent and pointless
The UN is only impotent when it can't agree. Such as in this case. We had
several major nations backing us (not least Brittain), and that was enough to
prevent the UN from stopping us from doing what we determined we needed to do.
If the UN (excluding the US of course) ever _agrees_ that the US has to get
out of Iraq, then for political reasons we'll pretty much have to do that.
All the major world powers feel this political pressure. The nations that
don't feel it are podunk third-world nations like Iraq -- those are the
nations that defy UN sanctions, and yes, somebody (like the US for example)
has to step in and do something militarily in those cases. But the larger
nations are held in check by the political pressure that the UN does weild.
> Osama bin Laden killed 3000 people in the US.
>
> Saddam Hussein killed none.
You can't really be enough of an imbecile to believe that killing people in
the US is the only measure of a threat, can you? Hussein had been building
a war machine in Iraq for two decades. He was an opportunistic tyrant, looking
for chances to consolidate power. Some time back, if you recall, he made an
attempt at a small neighboring nation. If he hadn't been driven back then,
he would have taken a second neighbor and a third by this time. If he hadn't
been watched carefully since, he'd have made a second attempt at Kuwait and
probably taken it. The man was a serious threat to the political stability
(such as it is) of the middle east -- a dangerous and heavily-contested region
politically, as much a powder keg as any acreage on the globe, just as likely
as the balkans to be the start of a world war.
bin Laden was a threat to the peace of mind of people in the US, and perhaps
the lives of a few. Hussein was a threat to humanity.
That is, of course, not what he'll be tried for. Which is fine; it doesn't
need to be. As far as any danger to the world goes, we could let him stand
trial at the Hague, probably get off easy, and live out his life comfortably
in Europe; as long as he doesn't return to power in the middle east, the
threat is cared for. What we _can't_ do is put him before a US court and
string him up in the fashion he deserves; that would be very policitally bad
for the US, and I think Bush knows better. So what we'll probably do is try
him in Baghdad for crimes against the Iraqi people. I'm not sure the outcome
matters very much, as long as he never comes to power again.
The hard thing is going to be building Iraq up to the point where the next
person won't be just as bad or worse. We really need to do approximately
what we did in Japan, after WWII, but I don't know if the Bush administration
will have enough years in power (even _with_ another four, which is not a
thing to take for granted) to make it happen, particularly the way they seem
to have been dawdling about it. I would have thought that by now they ought
to have districts drawn in a fashion that guarantees enough representatives to
at least four major factions that none of them can get jack diddly squat
pushed through without support from the other factions, set up an
infrastructure for a provisional government, and set a date for the first
election. I mean, sure, if you're going to make a first-world nation out of
Iraq, you're going to need to install power and phone lines and all that jazz,
yeah, but that stuff takes a while; getting a provisional government set up
would earn some nice political capital and maybe buy some time to make the
rest of it happen.
Then again, depending on how the Hussein trial is conducted, that might buy
some political capital and some time too. Still, in Bush's position, I would
sure have been a lot quicker about drawing up districts for representation and
stuff, the beginnings of a provisional government. It's not like it couldn't
be done at the same time as other stuff. Also key is that if you can get a
provisional government going, US and UN forces can be used to make sure it
stays in power; that doesn't cost nearly as much politically as just running
the country directly, like we are basically doing now.
> I believe this is where access falls short. I don't believe the syntax of
> access is 100% standard.
Last I knew, no available database has 100% standard syntax. They all differ
from the standard in a number of areas.
The problem with Access is that it's becomming obscure. It's not included
with most versions of MS Office anymore, for one reason or another (probably
because MS wanted to drive sales of their _other_ database offering, SQL
Server), and so consequently few desktops have it, so nobody knows it. SQL
is used heavily on servers and in backend stuff, and so pretty much every
database administrator knows one dialect or another of SQL.
If you're teaching people who are going to work mostly with desktop and office
stuff, Access might be appropriate, though frankly the database in MS Works is
more widely distributed on the whole, simpler, better known, and adequate for
basic office-type needs. If OTOH you're teaching people who are planning to
work in IT and administer networks and servers, they're going to have to know
SQL in one form or another. This will be true even if they work in an all
Microsoft-only environment -- for serious databases MS strongly pushes their
SQL Server product. If they work in a heterogenous environment, they'll use
Oracle (if the database is mission-critical and the place is rolling in cash)
or MySQL or PostgreSQL (for smaller installations) or _potentially_ MS SQL
Server -- but the chances of seeing Access in a heterogenous environment are
roughly zero. In any non-MS homogenous environment (e.g., everything comes
from one specific Unix vendor) the chances of using Access for anything
serious are *exactly* zero.
The question then is which SQL implementation to have them know. I would
suggest picking one of the four covered by SQL in a Nutshell, and make sure
they're aware that there are differences in the other implementations.
Which one to pick? Where are they looking to work? All-Microsoft shops?
Teach 'em MS SQL Server. Big enterprises? Oracle. Small businesses with
a heterogenous environment? MySQL probably, or perhaps PostgreSQL. If you
don't know, just pick one. Knowing one dialect of SQL will make it easy for
them to pick up another, so it's not wasted. There are differencesin the
details, but the *concepts* are to a large extent the same, at least for the
basic stuff you're going to teach in a course.
> It is a full fledged programming language though.
Right, that was my point. Its *purpose* may be page description (a form of
markup), but while it is a special-purpose language (and therefore not really
a general-purpose language) it nevertheless is a programming language.
The same can be said of Inform (though its purpose is quite different). Zarf
once wrote a (small, unpowerful, proof-of-concept-type) lisp interpreter in
Inform, for (I assume) hack value, and other things have been done similarly
in it (a tetris game, for example), but it is primarily used for just one
thing, interactive fiction. Still, there's no question in anybody's mind
about whether Inform is a programming language. It's not a *general-purpose*
programming language (mostly because the platforms it compiles to don't have
good general-purpose I/O, for portability reasons), but it's a programming
language -- a special-purpose one. I'd put PostScript in the same category.
As a side note, Zarf once threatened (in jest, I think) to do a work of
interactive fiction in PostScript.
If your house is 60F (did you typo that?), everything is working exactly as it
should. Maybe if the temperature were _less_ than room temperature, you might
start to get concerned, but 60 is the highest you would ever want it to be
(unless you're running a nursing home full of little old ladies who eat like
birds and so get the shivers if the room drops below body temperature). We
set our air conditioner lower than that in the summer.
> Funny aside -- PostScript's processing instructions would probably make it
> a good exception to these examples -- it probably _is_ Turing equivalent
I fail to see how that makes it an exception. Sure, PostScript is powerful
enough to be classified as a programming language. But it's not really a
general-purpose language. So then, it's in the same category with TADS and
POV -- a niche programming language, suited for a particular purpose.
I would not classify PostScript as a markup language merely because it is
capable of formatting content. Perl has extensive support for formats, and
as far as that goes even C has printf, but we don't call them markup languages,
because they support variables and flow control, which makes them programming
languages. HTML (as such) supports neither, so it's obviously not a programming
language. (There is of course ECMA script... but that is generally considered
a separate language from HTML, close association notwithstanding.)
I suppose one could argue that PostScript is a markup language because its
primary intended purpose is to format content, but IMO that argument falls
flat unless you want to consider ECMA script and PHP as markup languages
also, which seems clearly wrong to me. Markup languages don't have flow
control and variables and conditionals and so on and so forth.
> You don't "run" an HTML page.
Well, no, of course you can't just run it. You have to compile and link
it first...</rimshot>
> the ability to have different resolutions on different virtual desktops
BeOS has this (different colour depths too, if you want), and I can tell you,
it's pretty nifty. Currently I don't use the multiple-desktops feature of X,
because I don't much need to see my wallpaper very often, so having a lot of
Windows in front of eachother is okay, and the task list in the Gnome panel
is pretty good for keeping track of a lot of windows on one desktop. But I
would probably use a second and maybe a third virtual desktop if they could
have different resolutions from the first one.
> Actually, there's plenty to say about Windows as a server platform.
Yeah, but he was trying to be *nice*.
> VMS was a cruddy operating system.
From the little experience I have with VMS it seems pretty good to me. Very
solid in terms of never going down barring hardware failure. (Unfortunately,
hardware failure seems to be depressingly common with the Alpha line. The
Vax was a more robust system.)
There are two problems with VMS though. First off, it's weird. By that I
mean that there are a lot of things it doesn't do the same way that other
systems do them, even when the other major systems (Unix, Windows, Mac) all
do them roughly the same way. The second problem with VMS is that it's hard
to hire people who know it well.
> Windows NT, being largely derived from VMS
As near as I can determine, this relationship is almost entirely mythological.
Early versions of NT ostensibly took some low-level code from MS XENIX, and
we're pretty sure that a lot of GUI and API stuff was taken from Windows 3.x,
but beyond that it appears to have been mostly new stuff.
If it were even tangentially related to VMS, it would probably be a lot more
*different* from the other systems, and a lot more like VMS. If the
filesystem for example were taken from VMS, one would expect it to have some
of the strange VMS filesystem semantics. (I'm not talking about command line
syntax here[1], but stuff like multiple different kinds of text files.) If
the kernel were taken from VMS, one would expect NT to have batch queues and
other process management features reminiscent of the minicomputer world, but
it much more closely resembles Unix. If the networking layer were taken from
VMS... err, let's just say it pretty obviously wasn't[2]. In summary, if
you've used Unix, NT, and VMS, you start to think of Unix and NT of being
fairly similar. If you've used Unix and NT before, VMS will weird you out.
It's quite different. Much moreso than Mac, Apple slogans notwithstanding.
The one thing in VMS that's mercifully similar to other systems is the
environment variable system.
[1] Though DCL syntax is a little odd, too. But that's a surface thing; the
command-line syntax on any platform can be changed just by installing a
different shell.
[2] The networking stuff in VMS is *weird*. I don't properly understand it,
but it sure as anything is not much like Windows or Unix networking.
> I grow weary of seeing lots of young 20-something applicants fresh out of
> school who claim they have excellent coding skills and then proceed to
> list about ten different languages including HTML
My resume lists languages in categories according to how well I know them.
It says I'm fluent in Perl and lisp and "somewhat familiar" with several
other languages. At one point it also said "and have had courses in" yet
several more languages, but that's one of the things that's going to get cut
next time I have to print the thing and want to get it down to a page, because
I've added several technologies (like MySQL) since the last time I did so.
And yeah, I list HTML, and CSS, but not under "programming languages". I list
them in under web development, alongside CGI and MySQL.
But I may not be the typical applicant you're talking about; my resume also
says I have a little VMS experience...
And I'm gonna have to take MS Office off, because the last time I did anything
serious with it was too many versions ago at this point. (Think in terms of
a version that ran on Windows 3.1...)
> a person from the other team can examine the slip and confirm that
> it's illegible
Or if they think you're just being a wuss about the legibility, they can
wisper to you what it says and you have to get your team to say it.
This scales really well to groups of various sizes containing persons of
various ages. Just make sure that when you split teams you don't put all
of something one one team (e.g., all of the geeks, all of the sports fans,
all of the young people, whatever). Split all the demographics across
both teams, and it works better.
Here's how to play: Everybody writes names of famous persons on a bunch of
little slips of paper, folds them once, and throws them in a big bowl. These
can be names of current celebrities, historical figures, literary characters,
cartoon characters, whatever, as long as they're sufficiently well-known
that there's a decent chance several people in the room know about them.
Then you take turns: a person from the one team, then a person from the
other, and then another person from the first team, and so on. You get one
minute to see how many you can get, as follows: You draw a slip of paper out,
look at it, and then without saying any part of it yourself you must get
someone on your team to say the name that's on the paper. If you've never
heard of the person, it's too bad: you make a scrunchy face and try to get
it some other way. ("Okay, the first name is the same as Ellison's first
name, and he's something you make pickles from.") You cannot pass*. When
you finish one, you draw another. When time runs out, you put the one you
didn't finish back into the bowl without revealing any more about what it
was, count how many you got, and add it to your team's score.
This is way more fun than it sounds like. With the right group of people,
someone can draw "Marvin K Mooney", "Alan Greenspan", and "Huldrych Zwingly"
one after the other. Watching their face can be quite entertaining.
* Exception: In cases of utter illegibility, when you can't make out
the letters at all, a person from the other team can examine the slip
and confirm that it's illegible, and you can skip it.