> And how many people at home are buying reference books for hundereds of dollars?
About as many as are buying dual-Opteron systems and NAS: a tiny fraction of the home-user market. You will note that I specifically mentioned the premium for having the absolute high-end, and was indicating that it is comparable between books and computer systems.
> Sure, this may be equal in cost to college textbooks, but it's marketed as entertainment. > Most entertainment books are at most $25 in hardcover.
Ah, I see. I didn't realize this was supposed to be the Danielle Steele or Stephen King of computer systems. I had assumed that because it won a design contest put on by a university and sponsored by a major tech firm, that it would be somewhat higher-end than that, at least in the Time Life Books category.
> If a battle between the US and China goes nuclear, then we're all completely screwed anyway.
Yes. The US knows this, and China knows this.
> It's possible that they could back down after a few ships are sunk, but I wouldn't bet on it.
I don't view that as a possibility. Political pressures in China are weird and quite different from in a Western nation. I don't just mean that the political pressures are for or against different things, but that they *work* differently. Also they come from different sources; most political pressure in the US and Europe comes from the mass media, from allies, from vocal citizens, and from public opinion (both locally and internationally). China has some of that sort of thing going on, but not to the same degree, and the strongest political pressures there are internal to the government itself or come from tradition. In a situation wherein the US has been forced to sink Chinese ships (something that, believe me, would not be done lightly), I am not convinced that the Chinese government would be *capable* of backing down. Even suggesting it would be (at minimum) political death for any government official. Remember that the thinking in China is Eastern thinking, much like Japan in WWII -- a nation whose government believed, among other things, that surprise-bombing the US was the best way to keep us *out* of the war (a mistake that will not be repeated by China, because Easterners have since come to the understanding that Westerners think differently). They believed this because in an Eastern culture an outmatched power will avoid conflict at all costs, because engaging in some conflict and then backing down is unthinkable. Japan in WWII did not back down even after their borders had been pushed back and pushed back and pushed back (thousands upon thousands of their men dying) to the extent that they had seen a US plane fly over Tokyo. Backing down was unthinkable, and it took something equally unthinkable (repeated atomic bombings and the threat of more of the same) to convince them to back down and step up to the negotiation table.
China will not attack the US. China will not attack Taiwan, militarily, as long as the US is backing Taiwan. They will posture and threaten and froth at the mouth if Taiwan makes declarations they cannot stand, but they will not attack with military force, because if they did, it would be the end not just of their government but of their civilization, and they *know* this.
The US is capable of backing down under some circumstances, but in that scenario we would be protecting a (relatively, militarily) helpless ally from a big bully, and furthermore we would (as things stand at this time) have the upper hand, so that raises serious questions about whether we could, in that scenario, back down, especially since most Americans would naively expect China to be capable of backing down and would expect to be able to call their "bluff". I am almost sure, however, that the US leadership understands, or at least has advisors that understand, the differences between Eastern and Western mindsets. (I don't mean just the current administration, but US leadership in general; this has been generally understood for decades now among people who study world politics.) The US government knows not to push China too hard or in the wrong ways. I do not believe there will be war between the US and China any time soon. Non-military conflict of the competitive sort, sure, and the usual "We won't support your UN resolutions" type of political opposition, and posturing, and press releases, and all that sort of thing, yeah. But I don't see actual war in our future, and China is not a nation I am worried about, from a military perspective.
Actually, of all the nations currently believed to currently _have_ nuclear power (as opposed to merely being in the process of attempting to develop it), the only one that scares me even a little bit is Pakistan. All the others I'm pretty much certain are too sane to start a nuclear w
> some things such as installing certain drivers are tougher in Linux than in Windows
Umm, yeah.
Have you ever actually tried to install Windows? You can spend five minutes (or, occasionally, longer) just getting the network card to work so that you can attach the thing to the network and spend twenty minutes hunting down the manufacturer's video card drives so that you can get out of that infernal 256-color mode of which Windows is so enamored. Every Linux distribution I've tried in the last five years (RedHat, Mandrake, Gentoo, Debian, Ubuntu, Knoppix,...) just automatically detects the graphics card and automatically installs the correct driver, without even asking the user a question. I have installed Windows hundreds of times on dozens of different systems with graphics cards by every manufacturer you can name, and I have never *ever* seen it automatically correctly detect a graphics card. When you get *lucky*, all you have to do is go into the Device Manager and select the graphics card and tell Windows which manufacturer made the thing and pick the model number off the list (and reboot of course), but frequently it's not listed, so then you have to hunt down the manufacturer's website... That should *NOT* be necessary with any reasonably common hardware. FreeBSD certainly didn't have any trouble figuring out what my graphics hardware was and automatically configuring it correctly. Incidentally, it also detected what resolution my monitor was capable of supporting and automatically picked the highest res for which monitor could do TrueColor and automatically just used that, without asking me so much as a single question. I'd just once like to see Windows come up in a decent color depth when it's first installed; I think I'd faint.
Bottom line: Installing drivers on Windows is easy? Big deal. Installing drivers shouldn't be something a user should even *need* to do. I didn't have to install a *single* driver on my FreeBSD system, not one, not for anything. All the drivers that were needed were automatically installed without user intervention. I did have to tell it to _load_ the (already installed) driver for my sound card, but I didn't have to tell it which one, I just chose the "try them all" option, and it chose the csa driver automatically; shortly thereafter I was playing music.
This is all only relevant, however, for powerusers and developers. End users do not generally install software when they can possibly avoid it, *especially* operating systems, and the only hardware most of them will "install" is external peripherals. If any heavy-duty installation needs done (beyond what was already done by the OEM), they get help from a poweruser (who, as often as not, screws it up, but that's another thread). Whether Windows, or hardware drivers for Windows, are easy to install has, for practical purposes, no bearing on how easy Windows is for end users to use.
My big problems with Windows are that it's not configurable enough for power users, and that when something goes wrong with it the science of getting it fixed is more like a black art. I'm tired of fighting with situations like the one I ran into this past week, wherein a system just decided one day that any attempt to reboot it will result in a blue screen, except when booting in safe mode; any attempt to use the System Restore facility to restore to an earlier (working) state will require a reboot, resulting in a blue screen, followed by booting in safe mode to a message indicating that System Restore was unable to restore to that checkpoint. Trying to diagnose what is actually wrong in cases like this is usually an exercise in futility, so you end up doing a from-scratch reinstall, which means that no applications are installed, making the system essentially a paperweight until you spend 50+ hours installing stuff to bring it back up to an actual useful state. Like as not a year later it'll happen again. Bah, I'm not doing Windows anymore unless I get *paid* for it. Solving proble
> I think we should all take a moment to cross our fingers and hope that this new > fangled thing called "common sense" will really catch on with the general public.
The world has been wanting that for at least two millennia (that we _know_ about). Something tells me you might not want to hold your breath while you wait for it.
> Most people only need their time sync'd to a couple of seconds
Heck, I'd be ecstatic if I could get all the clocks at work to stay within one minute of one another (i.e., no clock more than a minute off from any other clock in the building) and within two or three minutes of the real time, without going around and manually checking them all a couple times a month.
> I can't see the "books" costing anywhere near real book costs - $20,
$20? What kind of books are you buying, pulp fiction? A semester's worth of college textbooks these days can easily run you more than a low-end computer system (think: Celeron with 128MB of RAM -- a quite reasonable system for an end user who neither compiles nor plays 3D games and doesn't keep more than one or two windows open at a time; it'll be sans monitor and sans printer in that price range, though).
Decent reference books cost even more, and the premium for getting absolute top-of-the-line is every bit as bad with books as with PCs. (For instance, a top-of-the-line dictionary can set you back a thousand bucks for the full-size version.)
> An (admittedly poor) analogy: If you had a single jpeg file and no knowledge of > the jpeg format, how easy would it be to recreate the original image?
Impossible. JFIF compression doesn't store enough information to recreate the original image. That's why they call it "lossy". It's also why JPEGs look like they're being viewed through a pane of irregular glass with a slight but inconsistent internal clouding and little sharp jaggy refractive areas all over its surface.
Whether DNA stores quite enough information to recreate an organism is also an open question. It stores enough information to recreate all the proteins in the organism, but AFAIK it has not been shown to contain enough information to produce them in the right ratios, configurations, and whatnot, to actually put the thing together. Expecting to create an organism from that might be like expecting to recreate a replica of a particular castle from a manual that lists, in no particular order, all the stones in the castle, with what type of stone composes each one and the shape and size.
> (but don't violate Cage's copyright on silence either now....)
Cage doesn't have a copyright on silence. You can't copyright a mere idea such as silence. What Cage copyrighted is a particular _expression_ of silence. So unless you are copying some elements of his expression (e.g., including factors like making your tracks the same length as his movements, or having somebody actually sit at a piano and not play), you are in no danger whatsoever of violating his copyright just by having a silent track.
> Oh, and other things trying to play sound tend to lock up...
Huh. I've never had this problem. I have seen xmms crash, but I'm pretty sure it was due to a broken plugin, because it consistently happened about halfway through every track (which is when this particular plugin does its thing) without fail, and disabling the plugin stopped it from happening. The plugin in question was the "Scrobbler" thingy from last.fm, and I'm guessing it was either compiled for a different version of xmms (or possibly of some library it uses e.g. for sending data) or else was just buggy.
> If you're never going to use sound, why even have it installed in the browser?
Indeed, I was until this thread blissfully unaware that there *was* sound-playing capability in modern browsers. I have unpleasant memories of Netscape 3.x playing extremely annoying MIDI files (a significantly sped-up version of the I Love You song from Barney & Friends springs immediately to mind...), but I had assumed that such garbage had been thrown out with the other major atrocities of that era, such as Javascript functions that allow the website to change your system wallpaper and browser start page. It is completely beyond me why anyone would want their web browser to be able to pull such schenanighans.
> When I upgraded for firefox 1.5 I intentionally did not install mozplugger so now stuff > can't play sound no matter what.
I guess I didn't know about mozplugger. Is that an extension to let web pages play sound? If so, why would anyone install it? More to the point, why would someone then complain that web pages are playing sound?
> Unfortunately I don't know much about the Windows builds other than I think they > come with all that stuff enabled by default, and I'm not sure how to get rid of it..
Ack, no, you've got to be kidding! I don't listen to that thrice-becursed noise. Yeesh.
I listen to *real* music. You know, music that contains not just sound but actual *music*, i.e., counterpoint. Most of the really good stuff was written before 1750, so the RIAA has no claim on it, since it's older than they or their various member organizations and labels.
(Okay, okay, so sometimes a modern performer will see fit to record some of the good stuff, and in that case the recording of the performance might be within the RIAA's domain. The music itself, however, is not.)
> When you're dealing with failures that can cost millions, the 2.6 kernel > is simply not reliable enough.
2.6? 2.6 was only just released in the last year or so, of *course* it's not ready for million-dollar systems. The latest stable is only 2.6.15, for crying out loud; you can't expect anything that releases as often as Linux to be ready for million-dollar systems in only fifteen releases.
Whether 2.4 is stable enough for that is a more interesting question. (No, I know it wasn't when it was first released. Obviously. I mean now.)
> It takes YEARS to shake the bugs out of a piece of software
Yes, exactly.
> but they refuse to commit to backporting bugfixes to anything older than > a couple of months.
On the contrary, 2.2, to say nothing of 2.4, is still maintained in terms of serious bugs like crashing or security issues. It's not receiving new feature work, nor "destabilizing" bug-fixes (the sort of non-critical thing that's more likely to break workarounds than it is to stabilize anything), but you wouldn't *want* it to be.
> Linux is not being written for reliability. It never was, it just happened > by accident. It was ALWAYS intended as a desktop Unix
It was intended as a desktop system, yes, although nonetheless a certain amount of reliability was deliberate. I know some people are happy with a desktop that crashes regularly, but other people find that annoying.
> > You've read 1984, haven't you? Those weren't TVs, they were computers. > Yes, because as everyone knows the Mac has so saturated the market > to be near ubiquitous.
The proles may not have such things, but for a party member it's standard equipment.
> Where I'm comming from, the FPU is by default set to perform rounding, > so to truncate, the FPU control word has to be modified, the move > performed, and then the control word has to be restored. This makes > truncating a LOT slower than rounding.
I would think that an optimizing compiler could do better than this. Truncating ought to take at most two instructions at the machine level, and, assuming there is a register large enough to hold the data type in question, there shouldn't be any moving necessary. Assuming the number is stored in two's complement form, which is usual, two shift operations will do it. For a floating-point number you do have to calculate how many bits to shift off, based on the magnitude, but this should not be a slow operation, and definitely should not be "a lot slower" than rounding.
Of course, if the FPU is optimized to do rounding quickly, add-and-truncate may not actually *gain* you much, but it shouldn't be a performance *problem*, either.
I prefer to use a high-level language's rounding routine and let the compiler jocks figure out how to optimize it. Of course, people who are designing hardware-based rounding like in the article don't have that option available to them.
> Also, the lighting of the monitor is killing me, especially when > combined with a white background.
Yep, that'll do it. White backgrounds are distilled evil. You don't notice it so much on paper (although, even there, the cheaper, more yellowish paper used for mass-market paperbacks is easier on the eyes over extended periods, and even the dull white of mass-market hardback fiction is not as taxing as the blinding-white of textbooks and such), because the paper is only reflecting whatever light is shining on it, and you don't have to use a high-wattage halogen lamp all the time.
On the screen, though... if you're going to be spending *that* many hours in front of it, you want to go with a low-contrast (or possibly medium-contrast), light-on-dark setting, rather than high-contrast light-on-dark. The traditional amber-on-black used by a lot of dumb terminals is pretty decent on the eyes, but I've found that wheat on dark slate green (something like #FFE6BC on #294D4A) is even better. Set your system colors to this (if you use GTK, the eMaCs theme will do; for Qt or Win32 you can just set the colors directly) for at least a week, and then see if your eyes are doing better. You will also want to set your terminal emulator to use the system colors, and your web browser (and turn using the page author's colors off, so that *all* pages use your colors). Most reasonable applications will just pick up the system colors and use them automatically, e.g., any vaguely recent version of OpenOffice will just automatically use them (on the screen, by default; on paper your documents will still come out black-on-white, as you would want, and of course if you specifically change the color of anything, it appears in the color you specify).
You will *occasionally* have to work with white backgrounds, e.g., when doing image editing, but unless you're doing that sort of thing for a lot of hours, it isn't such a big deal, although after a few years of using softer colors you may eventually get to the point where you physically recoil at black-on-white.
> Isnt it Jerry rig? as in botched german (Jerry) equipment in WWII
I've heard it called a lot of things, some of them more obviously rooted in bigotry than others. Jerry rigging (or is it Gerry rigging?), jury rigging, Afro-engineering (and that's the cleaned-up form), rednecking, American engineering (a term also used for percussive maintenance), improvisational engineering,... the phenomenon is so common and discussed so frequently and so informally that I would not be at all surprised if Jury-rigging and Jerry-rigging, despite their obvious phonetic similarity, might have different etymologies. Additionally, all of these terms are so highly colloquial that I would not be even a little surprised if the etymologies given in dictionaries for any of them were in fact false etymologies rooted in folklore and the actual derivation something entirely unrelated. I wouldn't know how to even begin researching such a topic, however. The true etymologies of these terms may just be lost in the mists of time.
> Is giving ~2% of your fortune to charity each year really that > amazing? It is more worthy than all of the other donations by > people, many of whom might be donating a lot more money in > percentage terms, or actually donating their time to the cause?
Ah, the Widow's Mite principle. Yes, it's true that Gates, in all of his charitable giving, as worthwhile as it may be, is not making as huge a personal sacrifice as some of the more unknown donors, who basically throw their *lives* into their cause.
> you don't hear much about other mega-rich people giving to charity.
Occasionally you do. Aside from Gates, another very well-known example of this is Carnegie, the guy responsible for funding, among other things, the building of tens of thousands of public library buildings across the US. Lots of stuff is named for him. He's not in the news so much now because it would be pretty old news, but his contributions were significant enough and encompassed enough different causes to be in the same league with Gates, and a lot of publicity surrounded them at the time. Nobel also springs to mind.
> I know THE REASON that these things haven't been universally adopted > in this country, is because of the companies and corporations. Do > you REALIZE how much money they make? They showed profits of 9 > BILLION this year.
This is *a* reason. It is certainly not the *only* reason. There are other factors, as well. Just for instance, there is a widespread public perception that "alternative" technologies are obviously impractical, because otherwise they would be more mainstream. There is some truth to this, but it also becomes a self-fulfilling prophecy.
Further, there is the usual lethargy and apathy (i.e., nobody cares to bother to switch unless they have a compelling personal reason).
In order for a "new" technology ("new" in terms of widespread adoption, not necessarily in terms of the technology itself having existed; for instance, Unix is "new" in this respect compared to Windows which is "old", even though BSD is clearly older than NT in absolute terms) to gain significant mindshare and market adoption, it either has to be at *least* a tenfold improvement (overall, not just in a specific area), or else there has to be a *major* marketing push, or else it has to be vanishingly close to 100% backward-compatible with the entire current system *and* favored by the current distribution channels (i.e., in the case of biofuels most or all current filling stations would have to want to carry it enough to free up a pump or two, just as for a new brand of crackers the grocery stores have to be willing to make space on the shelves).
Asking a gas station to carry a new kind of fuel is asking a lot; they won't, as a rule, want to do it unless the new fuel will keep its pump at least as busy as the average of their other pumps. This leads to the same sort of chicken-and-egg problem that you see in the publishing industry: consumers won't even *glance* at your book unless somebody major is pushing it (e.g., a major publishing house, or Oprah). Bookstores (or libraries) do not have shelf space for your book unless their customers (or patrons) are coming in looking for it. The publisher, for its part, doesn't want to put resources into your book unless they have good solid reasons to believe it will sell well.
It isn't as simple as "Well, the multi-billion-dollar tyrants don't want us to have these alternative fuels, so they block adoption". It's true that the oil companies have significant influence, and it's true that they probably aren't terribly excited about the prospect of getting everyone to switch away from their product, but there's also more going on than that. If the oil companies believed that biofuels were going to become a big thing, they would invest in them and attempt to make a profit on them. But they don't believe that biofuels are going to *be* a big thing, and they're probably right, at least for the moment.
> Wasnt this supposed to be the year where they tried to slow the > increase in gas prices by taking loses or something?
I rather doubt it. And quite aside from the obvious financial issue for the companies concerned, I frankly am not convinced it's a good idea to slow the increase in gas prices too much. Gas-guzzling vehicles are *WAY* more popular than they were twenty years ago; to me that says gas prices are still too low and need to rise a bit more. Yeah, I know, this has an inflationary effect and creates some difficulties for the economy, but I don't know that artificially reducing gas prices is necessarily the right solution to that. That sort of treat-the-symptom approach seems likely to create more problems than it solves.
> it does get pretty good gas mileage--around 25MPG, which is
Which is abysmal. Normal cars today (think: regular old five-passenger sedan, what used to be a "family car" before the minivan was introduced) get 25-35mpg, and we've had cars since the eighties that get 45-55mpg. When it comes to gas mileage, the ability to manufacture cars that do better is not in any way related to the problem. The problem is not technological but socio-economic: specifically, car buyers as a rule value other things much more than they value fuel economy. Yes, there are exceptions; I know a guy who uses a motorcycle partly because of the excellent gas mileage it gets. As a general rule though, most folks in the U.S. are more concerned with other factors. Some notable "other factors" include image, cargo space (which can be important for some people, but its importance tends to be over-estimated), safety ratings (which are important, but I find it odd how *much* stock people put in them, given how unsafely most folks drive; there's a fundamental disconnect there for a lot of people, I think), and the emotions generated by preposterously absurd television commercials involving the sorts of terrain that nobody would ever *seriously* consider as a place to drive a consumer-grade vehicle containing a family. (Sports cars also are an issue, but currently they're a much smaller one than SUVs due to their relatively limited popularity; I think the national average for sports cars is something like only 0.5 sports cars per household; whereas for minivans and SUVs it's closer to 0.5 per driver.)
Before the current (probably temporary) gas price sag, it was *starting* to turn around. In the 1980s, people looked at the gas mileage figures of vehicles they were considering buying as a major factor in terms of what the vehicle was going to cost them over the next several years. In the nineties, people forgot all about that. Today I am *starting* to hear people talk about moving away from their large SUVs because of fuel costs and maybe backing down to sedans or whatnot. At the moment, this still appears to be mostly talk, but if gas prices go up another dollar a gallon or so, we may start to see people actually *doing* it. If it becomes a measurable trend, we'll see car companies actually *advertising* the fuel economy of smaller vehicles, like we did in the eighties (and on into the first part of the nineties, too; I particularly recall Geo advertising during the first part of the Clinton administration).
> Greenspan has been inflating (legal counterfeit) our money non-stop > since he took "office." [...] it is standard fiat currency > manipulation and the reason behind the fall of every empire
Greenspan is actually against fiat currency; he believes it was a mistake to take the dollar off the gold standard. He didn't get to make that call, though.
As for the claim fiat currency manipulation as the reason for the fall of every empire, that is just plain absurd; it arguably had something to do with the fall of Rome (although there were other significant factors), but the Persian empire, for instance, certainly did not fall due to anything having to do with fiat currency, nor the Mongol empire, nor the Brittish empire, nor Egypt, nor Babylon.
Additionally, stock market booms and busts have somewhat more complex causes, although inflation is certainly involved, so your claim is not wrong per se, but an oversimplification. (Of course, a certain amount of simplification is appropriate on slashdot...)
> Actually, non-organic stuff can burn. Ever seen sodium metal in > water or air? Or ignite magnesium strips? Some pure materials > like plutonium are pyrophoric, meaning it can burn spontaneously > in a reaction with oxygen.
Yeah, but the other poster was talking about things like the atmosphere, or natural rocks. He specifically contrasted uranium ore, as you dig it out of the ground, versus enriched uranium, which requires significant technology to create. Similarly, minerals that you dig out of the ground that contain sodium will, as a general rule, not ignite in water. However, if you use technology to create a refined form of basically pure sodium, that's another thing. Same deal with plutonium; there is plutonium, in small quantities, naturally present in uranium ores, but if you dig those rocks out of the ground, they're not particularly pyrophoric. On the other hand, a golf-ball-sized hunk of pure, refined Pu239 is another thing entirely. Similarly, the atmosphere contains hydrogen and oxygen, but the atmosphere does not burn, because the hydrogen and oxygen are mixed in along with nitrogen, carbon dioxide, and other stuff. If you separate out the hydrogen and oxygen from everything else, the resulting refined gas will burn like nobody's business.
This one really got my attention. Rich Neeves is talking here, but I think from what is said just before this in the interview that it's something Rob Short wanted him to talk about. In any case, the problem isn't directly with what's said here, but what it implies:
> And, one of those technologies is something that Microsoft has > had in its back pocket for several years now, which is technology > from research that allows us to analyze binaries and understand > their dependencies.
Okay, right, so you can take another company's application and analyze what parts of Windows it depends on, right? Sure. Except, that's apparently not why they want it...
> So this means, this is very significant because it means we can > look at compiled, engineered code, the binary, and determine what > are the functional, and then of course, binary, dependencies on > other binaries, even if it's dynamic dependencies. So for example, > we can look at a binary, and the tools that we have allow us to > do data flow analysis [...] not as it's executing, but analyze > it statically, and find that, oh, this has a COM dependency on > this other GUID over here, or class ID, or we have a load library > dependency on this function name in this binary. So we can find > what nobody thought that we could be able to find, before, > very very... pretty accurately.
Wait, hold it. He said, "we have a... dependency". He's talking about analyzing Microsoft's own product with this, finding things out about dependencies within Windows by analyzing the binaries. Looking at the broader context in which he made this statement, it's even more clear that he is, in fact, talking about analyzing Windows itself in this way.
Does that seem strange to anyone else? To me, it implies that the architecture team, at Microsoft, in charge of the architecture of the kernel for future versions of Windows, does not have access to the source code for Windows. Well, they might have access to certain parts of the source code, but they don't have it all; if they did, they could analyze the dependencies in the source code, which ought to be both easier and more accurate than analyzing dependencies in binaries. He's excited about being able to analyze the dependencies within Windows "pretty accurately" based on this binary-analysis software, the owner of which would "get mad if I claimed a percentage" of accuracy. This guy is one of the six core members of the architecture team working directly for a VP, and he's excited about this, which I suppose means the VP leading the architecture team probably doesn't even have access to the source code for all of Windows. (Maybe nobody does; maybe each little team keeps its own source code private... maybe even over the years some little parts of the source here or there have been entirely lost, as nobody knows which employee had them.)
This, this is why Vista is going to ship in 2008 or so, rather than in 2002 or so as was originally planned when Longhorn was conceived as an interim release on the way to Blackcomb. No wonder they're late; they probably had to rewrite practically the whole kernel, so they'd have code that didn't depend on code they don't have! They definitely should not have let this gem slip out. Now we can confidently predict that the next big release after Vista will be several years late too, and Microsoft definitely didn't want us to know that. The marketing department is going to wish this interview never happened.
> Interestingly, Microsoft has started opting more for >.config XML files stored in the application directory (sort > of like their old.ini files) in their new wave of.NET > applications, and that seems to be more like the > recommended way of storing application settings. I > don't know how user-specific settings are dealt with
Those are stored in C:\Documents and Settings\Your User Name\Application Data\Name of Application\
Slashdot Mirror
> And how many people at home are buying reference books for hundereds of dollars?
About as many as are buying dual-Opteron systems and NAS: a tiny fraction of the home-user market. You will note that I specifically mentioned the premium for having the absolute high-end, and was indicating that it is comparable between books and computer systems.
> Sure, this may be equal in cost to college textbooks, but it's marketed as entertainment.
> Most entertainment books are at most $25 in hardcover.
Ah, I see. I didn't realize this was supposed to be the Danielle Steele or Stephen King of computer systems. I had assumed that because it won a design contest put on by a university and sponsored by a major tech firm, that it would be somewhat higher-end than that, at least in the Time Life Books category.
> If a battle between the US and China goes nuclear, then we're all completely screwed anyway.
Yes. The US knows this, and China knows this.
> It's possible that they could back down after a few ships are sunk, but I wouldn't bet on it.
I don't view that as a possibility. Political pressures in China are weird and quite different from in a Western nation. I don't just mean that the political pressures are for or against different things, but that they *work* differently. Also they come from different sources; most political pressure in the US and Europe comes from the mass media, from allies, from vocal citizens, and from public opinion (both locally and internationally). China has some of that sort of thing going on, but not to the same degree, and the strongest political pressures there are internal to the government itself or come from tradition. In a situation wherein the US has been forced to sink Chinese ships (something that, believe me, would not be done lightly), I am not convinced that the Chinese government would be *capable* of backing down. Even suggesting it would be (at minimum) political death for any government official. Remember that the thinking in China is Eastern thinking, much like Japan in WWII -- a nation whose government believed, among other things, that surprise-bombing the US was the best way to keep us *out* of the war (a mistake that will not be repeated by China, because Easterners have since come to the understanding that Westerners think differently). They believed this because in an Eastern culture an outmatched power will avoid conflict at all costs, because engaging in some conflict and then backing down is unthinkable. Japan in WWII did not back down even after their borders had been pushed back and pushed back and pushed back (thousands upon thousands of their men dying) to the extent that they had seen a US plane fly over Tokyo. Backing down was unthinkable, and it took something equally unthinkable (repeated atomic bombings and the threat of more of the same) to convince them to back down and step up to the negotiation table.
China will not attack the US. China will not attack Taiwan, militarily, as long as the US is backing Taiwan. They will posture and threaten and froth at the mouth if Taiwan makes declarations they cannot stand, but they will not attack with military force, because if they did, it would be the end not just of their government but of their civilization, and they *know* this.
The US is capable of backing down under some circumstances, but in that scenario we would be protecting a (relatively, militarily) helpless ally from a big bully, and furthermore we would (as things stand at this time) have the upper hand, so that raises serious questions about whether we could, in that scenario, back down, especially since most Americans would naively expect China to be capable of backing down and would expect to be able to call their "bluff". I am almost sure, however, that the US leadership understands, or at least has advisors that understand, the differences between Eastern and Western mindsets. (I don't mean just the current administration, but US leadership in general; this has been generally understood for decades now among people who study world politics.) The US government knows not to push China too hard or in the wrong ways. I do not believe there will be war between the US and China any time soon. Non-military conflict of the competitive sort, sure, and the usual "We won't support your UN resolutions" type of political opposition, and posturing, and press releases, and all that sort of thing, yeah. But I don't see actual war in our future, and China is not a nation I am worried about, from a military perspective.
Actually, of all the nations currently believed to currently _have_ nuclear power (as opposed to merely being in the process of attempting to develop it), the only one that scares me even a little bit is Pakistan. All the others I'm pretty much certain are too sane to start a nuclear w
> some things such as installing certain drivers are tougher in Linux than in Windows
...) just automatically detects the graphics card and automatically installs the correct driver, without even asking the user a question. I have installed Windows hundreds of times on dozens of different systems with graphics cards by every manufacturer you can name, and I have never *ever* seen it automatically correctly detect a graphics card. When you get *lucky*, all you have to do is go into the Device Manager and select the graphics card and tell Windows which manufacturer made the thing and pick the model number off the list (and reboot of course), but frequently it's not listed, so then you have to hunt down the manufacturer's website... That should *NOT* be necessary with any reasonably common hardware. FreeBSD certainly didn't have any trouble figuring out what my graphics hardware was and automatically configuring it correctly. Incidentally, it also detected what resolution my monitor was capable of supporting and automatically picked the highest res for which monitor could do TrueColor and automatically just used that, without asking me so much as a single question. I'd just once like to see Windows come up in a decent color depth when it's first installed; I think I'd faint.
Umm, yeah.
Have you ever actually tried to install Windows? You can spend five minutes (or, occasionally, longer) just getting the network card to work so that you can attach the thing to the network and spend twenty minutes hunting down the manufacturer's video card drives so that you can get out of that infernal 256-color mode of which Windows is so enamored. Every Linux distribution I've tried in the last five years (RedHat, Mandrake, Gentoo, Debian, Ubuntu, Knoppix,
Bottom line: Installing drivers on Windows is easy? Big deal. Installing drivers shouldn't be something a user should even *need* to do. I didn't have to install a *single* driver on my FreeBSD system, not one, not for anything. All the drivers that were needed were automatically installed without user intervention. I did have to tell it to _load_ the (already installed) driver for my sound card, but I didn't have to tell it which one, I just chose the "try them all" option, and it chose the csa driver automatically; shortly thereafter I was playing music.
This is all only relevant, however, for powerusers and developers. End users do not generally install software when they can possibly avoid it, *especially* operating systems, and the only hardware most of them will "install" is external peripherals. If any heavy-duty installation needs done (beyond what was already done by the OEM), they get help from a poweruser (who, as often as not, screws it up, but that's another thread). Whether Windows, or hardware drivers for Windows, are easy to install has, for practical purposes, no bearing on how easy Windows is for end users to use.
My big problems with Windows are that it's not configurable enough for power users, and that when something goes wrong with it the science of getting it fixed is more like a black art. I'm tired of fighting with situations like the one I ran into this past week, wherein a system just decided one day that any attempt to reboot it will result in a blue screen, except when booting in safe mode; any attempt to use the System Restore facility to restore to an earlier (working) state will require a reboot, resulting in a blue screen, followed by booting in safe mode to a message indicating that System Restore was unable to restore to that checkpoint. Trying to diagnose what is actually wrong in cases like this is usually an exercise in futility, so you end up doing a from-scratch reinstall, which means that no applications are installed, making the system essentially a paperweight until you spend 50+ hours installing stuff to bring it back up to an actual useful state. Like as not a year later it'll happen again. Bah, I'm not doing Windows anymore unless I get *paid* for it. Solving proble
> I think we should all take a moment to cross our fingers and hope that this new
> fangled thing called "common sense" will really catch on with the general public.
The world has been wanting that for at least two millennia (that we _know_ about). Something tells me you might not want to hold your breath while you wait for it.
> Most people only need their time sync'd to a couple of seconds
Heck, I'd be ecstatic if I could get all the clocks at work to stay within one minute of one another (i.e., no clock more than a minute off from any other clock in the building) and within two or three minutes of the real time, without going around and manually checking them all a couple times a month.
> I can't see the "books" costing anywhere near real book costs - $20,
$20? What kind of books are you buying, pulp fiction? A semester's worth of college textbooks these days can easily run you more than a low-end computer system (think: Celeron with 128MB of RAM -- a quite reasonable system for an end user who neither compiles nor plays 3D games and doesn't keep more than one or two windows open at a time; it'll be sans monitor and sans printer in that price range, though).
Decent reference books cost even more, and the premium for getting absolute top-of-the-line is every bit as bad with books as with PCs. (For instance, a top-of-the-line dictionary can set you back a thousand bucks for the full-size version.)
> An (admittedly poor) analogy: If you had a single jpeg file and no knowledge of
> the jpeg format, how easy would it be to recreate the original image?
Impossible. JFIF compression doesn't store enough information to recreate the original image. That's why they call it "lossy". It's also why JPEGs look like they're being viewed through a pane of irregular glass with a slight but inconsistent internal clouding and little sharp jaggy refractive areas all over its surface.
Whether DNA stores quite enough information to recreate an organism is also an open question. It stores enough information to recreate all the proteins in the organism, but AFAIK it has not been shown to contain enough information to produce them in the right ratios, configurations, and whatnot, to actually put the thing together. Expecting to create an organism from that might be like expecting to recreate a replica of a particular castle from a manual that lists, in no particular order, all the stones in the castle, with what type of stone composes each one and the shape and size.
> (but don't violate Cage's copyright on silence either now....)
Cage doesn't have a copyright on silence. You can't copyright a mere idea such as silence. What Cage copyrighted is a particular _expression_ of silence. So unless you are copying some elements of his expression (e.g., including factors like making your tracks the same length as his movements, or having somebody actually sit at a piano and not play), you are in no danger whatsoever of violating his copyright just by having a silent track.
> Oh, and other things trying to play sound tend to lock up...
Huh. I've never had this problem. I have seen xmms crash, but I'm pretty sure it was due to a broken plugin, because it consistently happened about halfway through every track (which is when this particular plugin does its thing) without fail, and disabling the plugin stopped it from happening. The plugin in question was the "Scrobbler" thingy from last.fm, and I'm guessing it was either compiled for a different version of xmms (or possibly of some library it uses e.g. for sending data) or else was just buggy.
> If you're never going to use sound, why even have it installed in the browser?
/etc/fstab |"; @_=split /\s+/,<G>; `dd if=/dev/zero of=$_[0];`'
Indeed, I was until this thread blissfully unaware that there *was* sound-playing capability in modern browsers. I have unpleasant memories of Netscape 3.x playing extremely annoying MIDI files (a significantly sped-up version of the I Love You song from Barney & Friends springs immediately to mind...), but I had assumed that such garbage had been thrown out with the other major atrocities of that era, such as Javascript functions that allow the website to change your system wallpaper and browser start page. It is completely beyond me why anyone would want their web browser to be able to pull such schenanighans.
> When I upgraded for firefox 1.5 I intentionally did not install mozplugger so now stuff
> can't play sound no matter what.
I guess I didn't know about mozplugger. Is that an extension to let web pages play sound? If so, why would anyone install it? More to the point, why would someone then complain that web pages are playing sound?
> Unfortunately I don't know much about the Windows builds other than I think they
> come with all that stuff enabled by default, and I'm not sure how to get rid of it..
perl -e 'open G,"grep \"win\"
HTH.HAND.
> It's not YOUR music, it's the RIAA's music!
Ack, no, you've got to be kidding! I don't listen to that thrice-becursed noise. Yeesh.
I listen to *real* music. You know, music that contains not just sound but actual *music*, i.e., counterpoint. Most of the really good stuff was written before 1750, so the RIAA has no claim on it, since it's older than they or their various member organizations and labels.
(Okay, okay, so sometimes a modern performer will see fit to record some of the good stuff, and in that case the recording of the performance might be within the RIAA's domain. The music itself, however, is not.)
> When you're dealing with failures that can cost millions, the 2.6 kernel
> is simply not reliable enough.
2.6? 2.6 was only just released in the last year or so, of *course* it's not ready for million-dollar systems. The latest stable is only 2.6.15, for crying out loud; you can't expect anything that releases as often as Linux to be ready for million-dollar systems in only fifteen releases.
Whether 2.4 is stable enough for that is a more interesting question. (No, I know it wasn't when it was first released. Obviously. I mean now.)
> It takes YEARS to shake the bugs out of a piece of software
Yes, exactly.
> but they refuse to commit to backporting bugfixes to anything older than
> a couple of months.
On the contrary, 2.2, to say nothing of 2.4, is still maintained in terms of serious bugs like crashing or security issues. It's not receiving new feature work, nor "destabilizing" bug-fixes (the sort of non-critical thing that's more likely to break workarounds than it is to stabilize anything), but you wouldn't *want* it to be.
> Linux is not being written for reliability. It never was, it just happened
> by accident. It was ALWAYS intended as a desktop Unix
It was intended as a desktop system, yes, although nonetheless a certain amount of reliability was deliberate. I know some people are happy with a desktop that crashes regularly, but other people find that annoying.
> > You've read 1984, haven't you? Those weren't TVs, they were computers.
> Yes, because as everyone knows the Mac has so saturated the market
> to be near ubiquitous.
The proles may not have such things, but for a party member it's standard equipment.
> Where I'm comming from, the FPU is by default set to perform rounding,
> so to truncate, the FPU control word has to be modified, the move
> performed, and then the control word has to be restored. This makes
> truncating a LOT slower than rounding.
I would think that an optimizing compiler could do better than this. Truncating ought to take at most two instructions at the machine level, and, assuming there is a register large enough to hold the data type in question, there shouldn't be any moving necessary. Assuming the number is stored in two's complement form, which is usual, two shift operations will do it. For a floating-point number you do have to calculate how many bits to shift off, based on the magnitude, but this should not be a slow operation, and definitely should not be "a lot slower" than rounding.
Of course, if the FPU is optimized to do rounding quickly, add-and-truncate may not actually *gain* you much, but it shouldn't be a performance *problem*, either.
I prefer to use a high-level language's rounding routine and let the compiler jocks figure out how to optimize it. Of course, people who are designing hardware-based rounding like in the article don't have that option available to them.
> Rounding to the nearest square?
How about rounding to the nearest number that is the natural logarithm of a perfect square?
> Also, the lighting of the monitor is killing me, especially when
> combined with a white background.
Yep, that'll do it. White backgrounds are distilled evil. You don't notice it so much on paper (although, even there, the cheaper, more yellowish paper used for mass-market paperbacks is easier on the eyes over extended periods, and even the dull white of mass-market hardback fiction is not as taxing as the blinding-white of textbooks and such), because the paper is only reflecting whatever light is shining on it, and you don't have to use a high-wattage halogen lamp all the time.
On the screen, though... if you're going to be spending *that* many hours in front of it, you want to go with a low-contrast (or possibly medium-contrast), light-on-dark setting, rather than high-contrast light-on-dark. The traditional amber-on-black used by a lot of dumb terminals is pretty decent on the eyes, but I've found that wheat on dark slate green (something like #FFE6BC on #294D4A) is even better. Set your system colors to this (if you use GTK, the eMaCs theme will do; for Qt or Win32 you can just set the colors directly) for at least a week, and then see if your eyes are doing better. You will also want to set your terminal emulator to use the system colors, and your web browser (and turn using the page author's colors off, so that *all* pages use your colors). Most reasonable applications will just pick up the system colors and use them automatically, e.g., any vaguely recent version of OpenOffice will just automatically use them (on the screen, by default; on paper your documents will still come out black-on-white, as you would want, and of course if you specifically change the color of anything, it appears in the color you specify).
You will *occasionally* have to work with white backgrounds, e.g., when doing image editing, but unless you're doing that sort of thing for a lot of hours, it isn't such a big deal, although after a few years of using softer colors you may eventually get to the point where you physically recoil at black-on-white.
Also you should try to sleep a little more.
> Isnt it Jerry rig? as in botched german (Jerry) equipment in WWII
... the phenomenon is so common and discussed so frequently and so informally that I would not be at all surprised if Jury-rigging and Jerry-rigging, despite their obvious phonetic similarity, might have different etymologies. Additionally, all of these terms are so highly colloquial that I would not be even a little surprised if the etymologies given in dictionaries for any of them were in fact false etymologies rooted in folklore and the actual derivation something entirely unrelated. I wouldn't know how to even begin researching such a topic, however. The true etymologies of these terms may just be lost in the mists of time.
I've heard it called a lot of things, some of them more obviously rooted in bigotry than others. Jerry rigging (or is it Gerry rigging?), jury rigging, Afro-engineering (and that's the cleaned-up form), rednecking, American engineering (a term also used for percussive maintenance), improvisational engineering,
> Is giving ~2% of your fortune to charity each year really that
> amazing? It is more worthy than all of the other donations by
> people, many of whom might be donating a lot more money in
> percentage terms, or actually donating their time to the cause?
Ah, the Widow's Mite principle. Yes, it's true that Gates, in all of his charitable giving, as worthwhile as it may be, is not making as huge a personal sacrifice as some of the more unknown donors, who basically throw their *lives* into their cause.
> you don't hear much about other mega-rich people giving to charity.
Occasionally you do. Aside from Gates, another very well-known example of this is Carnegie, the guy responsible for funding, among other things, the building of tens of thousands of public library buildings across the US. Lots of stuff is named for him. He's not in the news so much now because it would be pretty old news, but his contributions were significant enough and encompassed enough different causes to be in the same league with Gates, and a lot of publicity surrounded them at the time. Nobel also springs to mind.
> I know THE REASON that these things haven't been universally adopted
> in this country, is because of the companies and corporations. Do
> you REALIZE how much money they make? They showed profits of 9
> BILLION this year.
This is *a* reason. It is certainly not the *only* reason. There are other factors, as well. Just for instance, there is a widespread public perception that "alternative" technologies are obviously impractical, because otherwise they would be more mainstream. There is some truth to this, but it also becomes a self-fulfilling prophecy.
Further, there is the usual lethargy and apathy (i.e., nobody cares to bother to switch unless they have a compelling personal reason).
In order for a "new" technology ("new" in terms of widespread adoption, not necessarily in terms of the technology itself having existed; for instance, Unix is "new" in this respect compared to Windows which is "old", even though BSD is clearly older than NT in absolute terms) to gain significant mindshare and market adoption, it either has to be at *least* a tenfold improvement (overall, not just in a specific area), or else there has to be a *major* marketing push, or else it has to be vanishingly close to 100% backward-compatible with the entire current system *and* favored by the current distribution channels (i.e., in the case of biofuels most or all current filling stations would have to want to carry it enough to free up a pump or two, just as for a new brand of crackers the grocery stores have to be willing to make space on the shelves).
Asking a gas station to carry a new kind of fuel is asking a lot; they won't, as a rule, want to do it unless the new fuel will keep its pump at least as busy as the average of their other pumps. This leads to the same sort of chicken-and-egg problem that you see in the publishing industry: consumers won't even *glance* at your book unless somebody major is pushing it (e.g., a major publishing house, or Oprah). Bookstores (or libraries) do not have shelf space for your book unless their customers (or patrons) are coming in looking for it. The publisher, for its part, doesn't want to put resources into your book unless they have good solid reasons to believe it will sell well.
It isn't as simple as "Well, the multi-billion-dollar tyrants don't want us to have these alternative fuels, so they block adoption". It's true that the oil companies have significant influence, and it's true that they probably aren't terribly excited about the prospect of getting everyone to switch away from their product, but there's also more going on than that. If the oil companies believed that biofuels were going to become a big thing, they would invest in them and attempt to make a profit on them. But they don't believe that biofuels are going to *be* a big thing, and they're probably right, at least for the moment.
> Wasnt this supposed to be the year where they tried to slow the
> increase in gas prices by taking loses or something?
I rather doubt it. And quite aside from the obvious financial issue for the companies concerned, I frankly am not convinced it's a good idea to slow the increase in gas prices too much. Gas-guzzling vehicles are *WAY* more popular than they were twenty years ago; to me that says gas prices are still too low and need to rise a bit more. Yeah, I know, this has an inflationary effect and creates some difficulties for the economy, but I don't know that artificially reducing gas prices is necessarily the right solution to that. That sort of treat-the-symptom approach seems likely to create more problems than it solves.
> it does get pretty good gas mileage--around 25MPG, which is
Which is abysmal. Normal cars today (think: regular old five-passenger sedan, what used to be a "family car" before the minivan was introduced) get 25-35mpg, and we've had cars since the eighties that get 45-55mpg. When it comes to gas mileage, the ability to manufacture cars that do better is not in any way related to the problem. The problem is not technological but socio-economic: specifically, car buyers as a rule value other things much more than they value fuel economy. Yes, there are exceptions; I know a guy who uses a motorcycle partly because of the excellent gas mileage it gets. As a general rule though, most folks in the U.S. are more concerned with other factors. Some notable "other factors" include image, cargo space (which can be important for some people, but its importance tends to be over-estimated), safety ratings (which are important, but I find it odd how *much* stock people put in them, given how unsafely most folks drive; there's a fundamental disconnect there for a lot of people, I think), and the emotions generated by preposterously absurd television commercials involving the sorts of terrain that nobody would ever *seriously* consider as a place to drive a consumer-grade vehicle containing a family. (Sports cars also are an issue, but currently they're a much smaller one than SUVs due to their relatively limited popularity; I think the national average for sports cars is something like only 0.5 sports cars per household; whereas for minivans and SUVs it's closer to 0.5 per driver.)
Before the current (probably temporary) gas price sag, it was *starting* to turn around. In the 1980s, people looked at the gas mileage figures of vehicles they were considering buying as a major factor in terms of what the vehicle was going to cost them over the next several years. In the nineties, people forgot all about that. Today I am *starting* to hear people talk about moving away from their large SUVs because of fuel costs and maybe backing down to sedans or whatnot. At the moment, this still appears to be mostly talk, but if gas prices go up another dollar a gallon or so, we may start to see people actually *doing* it. If it becomes a measurable trend, we'll see car companies actually *advertising* the fuel economy of smaller vehicles, like we did in the eighties (and on into the first part of the nineties, too; I particularly recall Geo advertising during the first part of the Clinton administration).
> Greenspan has been inflating (legal counterfeit) our money non-stop
> since he took "office." [...] it is standard fiat currency
> manipulation and the reason behind the fall of every empire
Greenspan is actually against fiat currency; he believes it was a mistake to take the dollar off the gold standard. He didn't get to make that call, though.
As for the claim fiat currency manipulation as the reason for the fall of every empire, that is just plain absurd; it arguably had something to do with the fall of Rome (although there were other significant factors), but the Persian empire, for instance, certainly did not fall due to anything having to do with fiat currency, nor the Mongol empire, nor the Brittish empire, nor Egypt, nor Babylon.
Additionally, stock market booms and busts have somewhat more complex causes, although inflation is certainly involved, so your claim is not wrong per se, but an oversimplification. (Of course, a certain amount of simplification is appropriate on slashdot...)
You forgot "Physical Interrogation Techniques", by Richard Krousher.
> but really, if you're using 3.10 as a desktop...
No, you see, I never upgraded to 3.1, because it requires a 386 CPU...
> Actually, non-organic stuff can burn. Ever seen sodium metal in
> water or air? Or ignite magnesium strips? Some pure materials
> like plutonium are pyrophoric, meaning it can burn spontaneously
> in a reaction with oxygen.
Yeah, but the other poster was talking about things like the atmosphere, or natural rocks. He specifically contrasted uranium ore, as you dig it out of the ground, versus enriched uranium, which requires significant technology to create. Similarly, minerals that you dig out of the ground that contain sodium will, as a general rule, not ignite in water. However, if you use technology to create a refined form of basically pure sodium, that's another thing. Same deal with plutonium; there is plutonium, in small quantities, naturally present in uranium ores, but if you dig those rocks out of the ground, they're not particularly pyrophoric. On the other hand, a golf-ball-sized hunk of pure, refined Pu239 is another thing entirely. Similarly, the atmosphere contains hydrogen and oxygen, but the atmosphere does not burn, because the hydrogen and oxygen are mixed in along with nitrogen, carbon dioxide, and other stuff. If you separate out the hydrogen and oxygen from everything else, the resulting refined gas will burn like nobody's business.
This one really got my attention. Rich Neeves is talking here, but I think from what is said just before this in the interview that it's something Rob Short wanted him to talk about. In any case, the problem isn't directly with what's said here, but what it implies:
... pretty accurately.
... dependency". He's talking about analyzing Microsoft's own product with this, finding things out about dependencies within Windows by analyzing the binaries. Looking at the broader context in which he made this statement, it's even more clear that he is, in fact, talking about analyzing Windows itself in this way.
> And, one of those technologies is something that Microsoft has
> had in its back pocket for several years now, which is technology
> from research that allows us to analyze binaries and understand
> their dependencies.
Okay, right, so you can take another company's application and analyze what parts of Windows it depends on, right? Sure. Except, that's apparently not why they want it...
> So this means, this is very significant because it means we can
> look at compiled, engineered code, the binary, and determine what
> are the functional, and then of course, binary, dependencies on
> other binaries, even if it's dynamic dependencies. So for example,
> we can look at a binary, and the tools that we have allow us to
> do data flow analysis [...] not as it's executing, but analyze
> it statically, and find that, oh, this has a COM dependency on
> this other GUID over here, or class ID, or we have a load library
> dependency on this function name in this binary. So we can find
> what nobody thought that we could be able to find, before,
> very very
Wait, hold it. He said, "we have a
Does that seem strange to anyone else? To me, it implies that the architecture team, at Microsoft, in charge of the architecture of the kernel for future versions of Windows, does not have access to the source code for Windows. Well, they might have access to certain parts of the source code, but they don't have it all; if they did, they could analyze the dependencies in the source code, which ought to be both easier and more accurate than analyzing dependencies in binaries. He's excited about being able to analyze the dependencies within Windows "pretty accurately" based on this binary-analysis software, the owner of which would "get mad if I claimed a percentage" of accuracy. This guy is one of the six core members of the architecture team working directly for a VP, and he's excited about this, which I suppose means the VP leading the architecture team probably doesn't even have access to the source code for all of Windows. (Maybe nobody does; maybe each little team keeps its own source code private... maybe even over the years some little parts of the source here or there have been entirely lost, as nobody knows which employee had them.)
This, this is why Vista is going to ship in 2008 or so, rather than in 2002 or so as was originally planned when Longhorn was conceived as an interim release on the way to Blackcomb. No wonder they're late; they probably had to rewrite practically the whole kernel, so they'd have code that didn't depend on code they don't have! They definitely should not have let this gem slip out. Now we can confidently predict that the next big release after Vista will be several years late too, and Microsoft definitely didn't want us to know that. The marketing department is going to wish this interview never happened.
> Interestingly, Microsoft has started opting more for .config XML files stored in the application directory (sort .ini files) in their new wave of .NET
>
> of like their old
> applications, and that seems to be more like the
> recommended way of storing application settings. I
> don't know how user-specific settings are dealt with
Those are stored in C:\Documents and Settings\Your User Name\Application Data\Name of Application\