It's been a long wait since the first three book of TAOCP came out (in the 80's I suppose). I first saw volume 1 in 1973, so for me it's been an even longer wait...
To get an ungrounded motherboard these days takes lots of work! No, all it takes is plugging the power cord into a socket that doesn't have the ground pin connected. This could be the user's fault (connecting with a 2-wire extension cord), or an electrician that didn't properly connect grounds in the walls. Yeah, electricians are trained and licensed and all that, but it doesn't mean they actually do the job right. 10 years ago I bought an ancient house from a bank, and among the paperwork turned over was a certificate that the wiring had just been inspected by an electrician. Soon after I moved in, I borrowed this little plug with 3 indicator lights from where I work, and in 5 minutes I found 3 sockets with no ground wire on the third pin and one with hot and neutral reversed!
Both situations are theoretically a safety hazard with things like toasters and lamps, but only if something is defective in the appliance you are using besides. For electronics, however, no ground means that surge suppressors don't work because they've got nowhere to dump the surge. There are also ways you can build electronic devices such that reversed hot and neutral will cause a 120V high-current zap when connected to another device, but I don't think UL would approve of such a design...
Of course, when the original electrician was that careless, I also worry about the one thing you cannot detect once everything's put together: mixing up neutral and ground. These are connected together out where the wiring comes into the house, but large appliances can draw enough current as to get several volts drop in the neutral line. What that means is that when your computer is connected to that miswired socket, the return current from your washing machine (say) might just take a shortcut through your motherboard!
You mean Windows Update will infect -- excuse me, update -- your system without verifying you actually own Windows? Or does it silently send that ID # you have to type in when installing Windows? I don't know if a court would consider that as equivalent to McA's username/password checking, but it looks to me like either it's prior art or else what's left of McA's patent is so narrow that it would be easy to work around -- just don't call the ID check anything like "username" or "password".
Not that I WANT M$ to find a workaround so they can sell Windows X(tra) P(hucked) with the feature that it has to phone home for authorization every time you change your hardware configuration. It would be much more fun if M$ had to either beat this in court (establishing a precedent that software patents truly have to be novel and non-obvious), pay out about $100 billion, or forget about combining anti-"piracy" with remote system maintenance.
No, it is possible, if your XML document had really long-worded tags. And consisted of mostly tags and not much actual data. Er, I think you could be a little more precise than that: it's possible if the document was one 199-byte tag and 1 byte data, assuming that you've agreed on a set of no more than 256 tags. Or you could have up to 65,536 tags, but then 2 bytes would just send the tag.
I couldn't find anything that really explained how ANS.1 works, and the specs appear to require payment, but from the apparently more knowledgeable posts on/. it appears that it substitutes binary numbers for tags and other repeated parts of messages. The substitution table is fixed in advance and it is assumed that both sender and receiver already have it. So it is only effective if the format is pretty much pre-defined and highly repetitive. Satellite telemetry is a good example. E.g., it might turn "Temperature of engine 2 nozzle, zone 4 = 65" into 2.4.6.5. Or ANS could do a pretty good job of compressing stock market prices by replacing those long corporation names with a short code -- but the exchanges long ago assigned short text codes...
LZW (*zip) compression also uses a substitution table, but in LZW most substitutions are not predefined. The software adds to the table as needed while processing a particular file, and puts each new substitution in the compressed file. So it's flexible; if you are compressing an XML files and someone uses a new tag, word, or phrase repeatedly, LZW will just assign a new code to that string, send the full string once (per file), and every subsequent use only requires the code.
In summary, 200 bytes to 2 bytes is B.S. or a contrived case -- about all you can do in 2 bytes is identify one string previously agreed upon, and if you ever might have to send a free-form message (even an update to the table of pre-defined strings) you're going to need at least one byte just to ID the message type. But if you have a large set of large files that are quite repetitive in both content and format, it might be possible to pre-define a substitution table for the whole set and get 100 to 1 lossless compression. But that's going to work with XML on the web only if you browse just one site whose contents meet the repetitiveness criteria...
By the way, I have seen 98% (50-1) compression using PKZIP. This was on AutoCAD DXF files, which is a remarkably bloated ASCII format representing CAD drawings. And it takes several megabytes before the compression becomes that good. You might get over 90% compression on XML if the files are big enough, but you really shouldn't put that much on one web page.
"blatantly obvious" doesn't even begin to cover it. The McAfee patent was filed Dec 8, 1998. Didn't Windows 97 come with a feature to update itself? Sure seems like prior art, aside from little details of the implementation -- which you probably wouldn't want to copy anyhow.
"The McAfee.com inventors of this patent include Chandrasekar Balasubramaniam, CTO & Vice President of Engineering; Babu Katchapalayam, Director of Engineering; Core Products; and Ravi Kannan, Director of Engineering, Infrastructure Services."
In the usual american corporate structure, no one above the lowest level of engineering management is ever going to get time to invent anything. Directors and VP's certainly can never get out of meetings long enough to do any engineering, if the titles actually indicate a high position in the company rather than being 2nd and 3rd tier in a tiny 3 or 4 tier organization. 8-) So do those two "directors" have nothing to direct but themselves, or is putting their boss's name on the patent possibly fraudulent?
I think you can divide the world into two categories...people who play 3D games on their computers and people who don't.
You forgot engineers, video editors, etc. Speed does count there. However, I used to do the same sort of engineering jobs on a 386 -- just had to schedule my work so autorouting would run overnight, if I picked the right settings. 500 MHz will autoroute a similar board in 10 minutes, so scheduling is a little easier, and I can try more different settings to try to get the best routing. 1.4GHz might or might not get it done in 5 minutes (CPU cycles aren't the only limiting factor!). So if I was doing that everyday, a few hundred extra for the fastest system would improve productivity be cost effective. Two 500MHz boxes and a keyboard-monitor-mouse switch would improve productivity far more...
Anyway, routing circuit boards is not my main job, and 500MHz is essentially instantaneous for everything else I do. But I remember when home computers were 8-bitters running games like the text-only D&D, and engineers got $50,000 Unix workstations, or else just a dumb terminal and a logon to the mainframe. Now the gameplayers need more power than most engineers...
Sounds like you ran into Matrix Mismanagement. That's not how it should go, but it takes pretty good managers to bring it off correctly.
The easiest example to understand of proper matrixing is the way we used to organize the factory here. We assemble circuit boards for about a dozen customers. So we used to have 5 "customer" teams, each covering one big customer or 3 or 4 little ones. But then there were also three groups of production automation (surface-mount, through-hole auto insert, and wave solder) that were shared by all the teams. So the operators of those machines had multiple bosses: their own production manager, and the customer team leaders. But we kept the lines of authority fairly clear because the production managers would allocate times on their machine and do the performance reviews for their people.
For engineering or coding, it's hard to keep the lines of authority that clear. You have various "functional" areas (e.g. mech design, electronic design, software, test, usability experts, marketing experts) and you draw people from these to form a project team. If the primary authority stays with the functional managers, then there is little motivation for people to commit themselves to a team. If real authority goes to the project leaders, then people have to contend with performance reviews written by someone who doesn't understand their specialty, and also some wind up splitting time between teams and with real trouble keeping all their bosses happy. If the managers are competent and get along with each other, they can make it work, but it's awfully easy for it to turn into a snake-pit.
Microsoft Project is probably not the answer. It is useful when used properly, but most managers can't. They toss in arbitrary numbers, then play around until the timeline ends when they want the job done -- and it doesn't have a damned thing to do with reality. To make it work, he's got to work very hard to get real numbers in there, identify ALL the dependencies, and (hardest of all) then he's got to accept the results even when they say that stuff he promised in 1 month is going to take 3...
If you do it right, the chart does show points where work can be done in parallel, and critical parts where a little extra resources can speed up the project quite a lot. However, with the typical management gloss over the details, the software is probably going to think that "If one woman can produce one baby in 9 months, then 9 women can produce one baby in 1 month." (Fred Brooks, The Mythical Man-Month) This results in completely unrealistic schedules.
I have worked on projects where project management software was critical. But if your organization is small enough that coders are usually doing multiple projects, then chances are the projects aren't complex enough to really benefit from it. That is, if you've only got 1/2 a coder to put on it and it's 500 hours of coding, then it's going to take half a year if you get him started right away, and longer than that if you mess around planning how you're going to allocate his 20 hours a week first...
And sometimes pilots will fly smack into a mountain, in daylight and in clear air, because they were too busy with all those instruments and radios to look out the windscreen. Nobody's fault but the system designers. 8-(
However, most modern piloting tasks don't require deep thinking. Try doing all that, or even half of that, plus taking a star sight and working out your position with sliderule and nav tables... That's why large airplanes used to have a four man flight crew -- two to fly, one to watch the multiple engines, and one to navigate -- and small airplanes didn't use to have that many distractions. E.g., Lindbergh could center the controls and take a star sight without worrying about running into anything over the Atlantic ocean.
The lost time usually isn't from when you _choose_ to switch tasks, it's from the interrupts, except when you are thrashing the interrupt stack in order to avoid doing real work... 8-)
There some very good parallels between brain work and computing in the 60's. A good CPU would often overrun the IO devices, and since the CPU cost half a million they didn't like wait cycles. So when it had to wait for IO, it would switch to another task, until that had to wait also, etc. The analogy is to you working on one job until you find that you need information from someone else to proceed, writing e-mail to him, then switch to another job until the reply comes back. This (task switching when on hold) improves productivity. On the other hand, when you get a phone call about some project you aren't even working on today, or have to stop coding to go to a meeting about parking spaces, you lose productivity. Likewise, interrupt-driven task switching tends to reduce the number of jobs finished per hour, and only became common when the CPU's became fast and cheap enough that you could afford to waste cycles.
Since the human brain isn't getting any faster, any situation where you are frequently interrupted is going to reduce the amount or quality of work completed. Note also that there are major and minor context switches, and the cost difference is much larger than the difference between switching processes and threads. Switching to a different part of the same project requires re-loading "registers" (short term memory), but the major context of the project stays the same. Switching to something I put away last week will probably require skimming through some of the documents to remember where I left off and to refresh my memory of the overall structure.
Patents were originally a very good idea that would promote the progress of technology. You invented something, you sent the patent office a full description and a working model, and if they hadn't seen anything like it before you could then get the exclusive rights to make it for 17 years -- but the description went into the public domain for other inventors to build upon. Now we have people that never created a working copy, let alone went into production, filing patents on things they don't know how to build (hence, clearly without much useful info in the patent), then selling it to lawyers who use them to harass businesses that do bring new technology to market.
1) Working models were originally required. The USPTO stopped routinely requiring this late in the 19th century because they ran out of storage space for all the inventions. (I think the examiner _can_ still require a working model, but it's only invoked when they really don't believe it can work -- and obviously they're quite a bit too credulous.) We need to bring this requirement back, at least to some extent -- the patent office can't store physical models, but they should ask for them, take some photos of the thing running (if it does), then ship it back. And of course software patent applications (if allowed at all) should come with all source code and a working executable on CD-R; if they can store all the paper in an application, they can certainly store these. Without this requirement, every real inventor is hostage to the Lemelson technique: look at what people need, guess what will be invented to fill that need, file a vague, broad application, then sue the real inventor later.
2) Market it or lose it: If you don't build your invention or license it to someone who can, after 2 or 3 years it should become public domain. (I thought this was once the law.) You do not promote the "progress of science and the useful arts" by letting people just sit on a good idea for 20 years. (The same thing should apply to copyrights, and to items that are removed from the market for long periods. If Disney wants to keep it's copyright on Snow White, let them either keep the videotape in stores, get the movie into theaters every couple of years, or post the MPEG on the web with a "download for non-commercial purposes only" notice. Not hold the damn thing off the market for a generation.)
3) Tighten up the definition of "invention." Under the rules of 1960, one could not patent a mathematical algorithm (LZW compression, public key encryption), or a thing of nature (any naturally occurring gene).
Not that I find checkboxes that hard to operate or anything. It sounds like a whole lot of people didn't see that checkbox, or else it was labeled with something that didn't actually tell you what is was going to do. Judges appointed by Republicans, or by that psuedo-Republican Clinton, may not agree, but I certainly don't see a checkbox marked with something like "enhance web links" as adequate notice, let alone something buried in a few pages of lawyerly gobbledygook.
If I buy a book I can highlight (or doodle, or rip out, pages for that matter). It becomes my book. I can also hire a person to buy some books, make random drawings, and rip out pages before I read it. I'd agree, except that from the reports I've seen (this is one thing I _don't_ intend to check out for myself), KaZaa neither makes it clear that their program has this "feature" before you install it, nor does it make the opt-out easy to find. So it's not like hiring a person to modify your books -- it's like hiring someone to install paneling, and he also stamps advertisements inside your books.
I would have no issues at all with software where the download page starts with a statement such as "To pay for making this free software available, we included XXX, which will modify your internet browser by adding paid advertisements. After downloading and installing this, you will see yellow underlines under certain words while browsing the web. These are links to advertisers' web pages. The links are not from the web sites where you see them and are not endorsed by the creators of the web sites."
Or to follow up on my analogy above: "we'll install paneling free, but we're going to stamp advertisements all over everything, not just the paneling." I wouldn't go for it, but if someone thinks that's an acceptable deal...
Monitor emissions can do a lot of spying, but they will not reveal your password (unless you're using software so braindead it displays the password when you type it). Not that a shielded, encrypted keyboard would be full protection for your passwords, either. It might defeat a physical or distant (Tempest-type) tap on the keyboard cable, but if they break in, they can load a keystroke monitor program that will snag the keys after they're decoded. Remember, the keys have to be decoded before they are presented to the OS to either handle itself or pass on to various applications. It might even be possible to decode your motherboard's radio emissions to tell when it's processing a keystroke and pick that up...
tracking the innocent people wastes inordinate amounts of FBI time, making it easier for criminals to slip under the radar. So? I remember when the FBI spent inordinate effort tracking a few, mostly harmless, political radicals, while claiming that the Mafia didn't exist so they didn't have to put any resources into fighting organized crime. Anything that makes it easier for them to do this sort of thing is BAD.
Note however that this case involves a fine technical point: they got a warrant, but it wasn't the right kind of warrant for a wiretap, and this does seem to be a wiretap as far as the technology goes. On the other hand, wiretaps are especially limited because when you tap a phone, you are tapping two persons, and often one of them is not a suspect. Tapping the keyboard cable doesn't involve an innocent third party. I can see a court reasonably going either way on this one...
Anyway, the big problem is that the penalties are backwards. Cops don't go to jail for illegal searches. DA's don't get disbarred for fooling some judge into approving a warrant that doesn't really cover what they are going to do. It's pretty unlikely that the judge who signs a warrant on insufficient grounds will even get a reprimand. ALL THAT is what should happen when the cops and prosecutors step over the line -- not tossing out the evidence.
the number of 2 and 3 sigma effects that disappear when more data is taken, or old results that feed into their calculations are updated, or theorists calculate the next order correction... well, that number greatly exceeds 6 results in 1000. (Laughter) Let's see, you spin a cobweb of mathematics that predicts a new "particle" might have certain characteristics, including breaking down in less than a femtosecond. You predict how it is going to break down and affect other particles that stay around long enough to be detected, then you try to get a few flickers in the detectors that sort of match. And you do a statistical analysis based on your theory that indicates those flickers probably confirmed your theory. Seems a bit of circular reasoning is involved... But at least the treatment of statistics by physicists beats the heck out of that in social science.
I switched from physics to engineering over 20 years ago, because the only parts of the field that were still making new theories were running into a wall when it came to verifying them. Particle physicists need a multi-billion dollar machine running for weeks to get a few flickers on the detectors. But they're in much better shape than cosmologists, who've got no way to run experiments at all, and who are perhaps drawing far too sweeping conclusions from too few observations. I salute you guys that are still hanging in there.
I've heard two extreme positions here -- and as usual, both of them are wrong. Really, it depends. The understudy idea is a good one, but there are two things that can block it from working. One is that sometimes there is only enough work in that area for one person. The other is if the "irreplaceable" person won't cooperate in training a potential replacement. Or the situation I saw once, where we had one PCB designer, not only about 1500 hrs/yr of PCB design work (he also had other skills that were in use the rest of the year), _and_ he was working at making himself really irreplaceable. I'm rather of a generalist; before we hired "joe", I was doing a little PCB design (not very well), and I obviously would have been on tap to fill in if joe had received a job offer from somewhere at 25% more... But one day when joe was on vacation, a customer called up with a question about one of joe's projects, and I found that he hadn't archived the design files for that -- or most of the other projects. After about six months of struggling to get him to document his work, we finally decided it would be best to get a backup of his hard-drives, fire him, and let me get to work trying to sort things out -- because the longer we waited, the worse that job would get!
And of course, a big part of the problem is that management is often so incapable of judging programmer performance on an individual basis, that it has little choice but to fall back on simplistic indicators, like age.
I'd say that's the main problem. The 10% or 20% of good programmers have been carrying the rest all along, and the managers never noticed. And the bad ones get worse as they get older, while the aces are likely to be in business for themselves by the time they are 40. But if you can't judge competence in the first place, chances are that your 10 new H1-B hires include 8 or 9 duds, and 1 or 2 that could be good programmers if someone translated the problem to be solved into Hindi...
The H1-B system is entirely skewed in the programmer's favor because it protects them from foreign competition. Change "H1-B" to "green card" and you're exactly right. Back in the 20's when immigration restrictions were first passed, people were flat out honest about their motives: (1) they were racists, and (2) they wanted to keep foreign workers out and raise wages for American laborers. I hope we've pretty much got over #1, but #2 is why the labor unions still want to restrict immigration.
But H1-B creates the opposite distortion: to give corporations a limited group of foreign high-tech workers at lower than competitive wages. The conditions amount to indentured servitude: H1-B workers generally cannot change jobs for higher pay, so their wage rates have nothing to do with the free market. Even though according to the law, H1-B's are supposed to get the same pay as Americans, there really isn't such a thing as a "standard rate" for programmers; each job is different, and so it is very easy to manipulate things so as to substantially underpay the H1-B's. The complicated paperwork requirements favor large corporations. Basically, their campaign contributions have bought them a special uncompetitive labor market -- as long as it stays small enough not to give the labor unions a big campaign issue.
As for the labor unions, note that they still have the immigration restrictions that count for them: Mexicans can come here temporarily to do farm work that is so hard and so low paid their members would rather collect unemployment, and H1-B's aren't competing with union workers either.
It's easy enough to purify. And, unlike the water in most cities, you aren't starting with the sewage outflow from the next city up the river... But actually, if drinking water is an issue, the spacecraft wasn't rigged right for a long mission. The crew will produce more than enough water as long as they've got food and oxygen.
The digestible parts of food are mostly chains of HCOH units. Your body burns that with O2 to produce CO2 + H2O. Some of the water is excreted and some evaporates from the lungs, skin, etc. The air system in the cabin has to capture that evaporated water before the humidity gets so high the instruments fog up. So I think you'd get enough drinking water from the dehumidifier, if the designers pay a bit of attention to the materials and arrangements so it didn't get contaminated. But if you want a shower, that's going to be in recycled sewage...
If you read the whole article, it's not just erosion that they're looking at (in fact, the erosion in that picture is from wind), but rather a waffle-like pattern that they think comes from something melting or evaporating out from under the surface. Mars isn't quite cold enough to get much frozen CO2 (except at the poles in winter), so it's probably water ice. Two less likely possibilities that some chemical reaction peculiar to the martian environment produced very large quantities of some other substance which can freeze and melt or sublimate at Martian temperatures to cause those potholes, are that some other mechanism entirely, which is not geologically significant here, produced features that just happen to look like potholes (uneven erosion by swirls in the wind, the footprints of invisible Martian elephant herds). I think about a 75% chance they have the right interpretation and those potholes are the tracks of ice deposits which have evaporated. (They are definitely not indication of ice being in that spot now, but probably the water went back underground somewhere else.)
That's good enough odds to do more studies and try to pick the right spot to send a robot drilling rig to look for ice. But certain other proposals like sending out a manned expedition with one-way fuel and the equipment to electrolyze water into rocket fuel will have to wait until the robots actually find ice that is there NOW, rather than the tracks of evaporated ice deposits.
Yes, the old sailing ships are one of many counter-examples to those worries about space crews going bonkers from "isolation." It's perhaps not the best one because (1) the smallest crews were around 20 men, considerably larger than any space mission currently under consideration, (2) ships rarely went more than 6 months between landfall, and (3) it wasn't at all uncommon for the early explorers to murder a bunch of natives as soon as they landed... 8-/ But seriously, sailing ships didn't have radio and therefore were much more isolated than a space mission would be. The military/exploration vessels were also much more crowded, at least until crewmen started dying of scurvy, etc. And modern nuclear submarines, which have radios but usually aren't allowed to use them, are also more isolated than a spaceship.
If it's anything like the Apollo missions, what will drive them crazy won't be isolation, it will be the CNN cameras watching them for two years!
Color steerable-beam CRT's have to run the beam at a few thousand volts to get it to stay focused as it is scanned around the screen. This gives the electrons enough energy to create soft x-rays when they hit the phosphors. So they have to melt lead oxide into the front glass, and also put some lead around the sides.
At any rate, this system shouldn't need lead shielding. If I understand it, it forms many little beams (one per pixel), which are not steered and travel less than an inch. So it should only need a few volts to operate, and the most powerful photons that can form will be ultraviolet, which is easily stopped by many kinds of lead-free glass or by the opaque back cover.
IMHO, lead from electronics is far from my greatest worry about landfills. (I am not a chemist, but I do know a few things.) Metallic lead only dissolves in acids, and lead glass from CRT's should be even less soluble. There are also a few ounces of 63% lead solder in a typical computer, but people have lived with much larger quantities of lead for millenia and usually avoided poisoning themselves. There is at least 2,000 years of history for lead pipes for drinking water (from liners for Roman aquaducts starting before 100BC to American city water systems in the early 1900's). Chances are these lead pipes haven't yet been all replaced in some of our cities -- and in other cities, they ripped them out and tossed them into the nearest landfill... But you won't find lead in the water coming out of a lead pipe unless the water is far more acidic than drinking water ought to be. Landfills might be more likely to have acid groundwater, but they ought to be ensuring that acid water doesn't leak out anyhow -- and when the acidity is neutralized, the lead compounds drop out of solution and don't travel any further. I'd worry more about the more unusual metals being touted as replacements for lead in solder -- some of them have never been widely used before, and there are probably things we don't know and won't know until too many people have been exposed to them for 20 years...
Slashdot Mirror
It's been a long wait since the first three book of TAOCP came out (in the 80's I suppose). I first saw volume 1 in 1973, so for me it's been an even longer wait...
Both situations are theoretically a safety hazard with things like toasters and lamps, but only if something is defective in the appliance you are using besides. For electronics, however, no ground means that surge suppressors don't work because they've got nowhere to dump the surge. There are also ways you can build electronic devices such that reversed hot and neutral will cause a 120V high-current zap when connected to another device, but I don't think UL would approve of such a design...
Of course, when the original electrician was that careless, I also worry about the one thing you cannot detect once everything's put together: mixing up neutral and ground. These are connected together out where the wiring comes into the house, but large appliances can draw enough current as to get several volts drop in the neutral line. What that means is that when your computer is connected to that miswired socket, the return current from your washing machine (say) might just take a shortcut through your motherboard!
I do my own house wiring now...
Not that I WANT M$ to find a workaround so they can sell Windows X(tra) P(hucked) with the feature that it has to phone home for authorization every time you change your hardware configuration. It would be much more fun if M$ had to either beat this in court (establishing a precedent that software patents truly have to be novel and non-obvious), pay out about $100 billion, or forget about combining anti-"piracy" with remote system maintenance.
I couldn't find anything that really explained how ANS.1 works, and the specs appear to require payment, but from the apparently more knowledgeable posts on /. it appears that it substitutes binary numbers for tags and other repeated parts of messages. The substitution table is fixed in advance and it is assumed that both sender and receiver already have it. So it is only effective if the format is pretty much pre-defined and highly repetitive. Satellite telemetry is a good example. E.g., it might turn "Temperature of engine 2 nozzle, zone 4 = 65" into 2.4.6.5. Or ANS could do a pretty good job of compressing stock market prices by replacing those long corporation names with a short code -- but the exchanges long ago assigned short text codes...
LZW (*zip) compression also uses a substitution table, but in LZW most substitutions are not predefined. The software adds to the table as needed while processing a particular file, and puts each new substitution in the compressed file. So it's flexible; if you are compressing an XML files and someone uses a new tag, word, or phrase repeatedly, LZW will just assign a new code to that string, send the full string once (per file), and every subsequent use only requires the code.
In summary, 200 bytes to 2 bytes is B.S. or a contrived case -- about all you can do in 2 bytes is identify one string previously agreed upon, and if you ever might have to send a free-form message (even an update to the table of pre-defined strings) you're going to need at least one byte just to ID the message type. But if you have a large set of large files that are quite repetitive in both content and format, it might be possible to pre-define a substitution table for the whole set and get 100 to 1 lossless compression. But that's going to work with XML on the web only if you browse just one site whose contents meet the repetitiveness criteria...
By the way, I have seen 98% (50-1) compression using PKZIP. This was on AutoCAD DXF files, which is a remarkably bloated ASCII format representing CAD drawings. And it takes several megabytes before the compression becomes that good. You might get over 90% compression on XML if the files are big enough, but you really shouldn't put that much on one web page.
"blatantly obvious" doesn't even begin to cover it. The McAfee patent was filed Dec 8, 1998. Didn't Windows 97 come with a feature to update itself? Sure seems like prior art, aside from little details of the implementation -- which you probably wouldn't want to copy anyhow.
In the usual american corporate structure, no one above the lowest level of engineering management is ever going to get time to invent anything. Directors and VP's certainly can never get out of meetings long enough to do any engineering, if the titles actually indicate a high position in the company rather than being 2nd and 3rd tier in a tiny 3 or 4 tier organization. 8-) So do those two "directors" have nothing to direct but themselves, or is putting their boss's name on the patent possibly fraudulent?
You forgot engineers, video editors, etc. Speed does count there. However, I used to do the same sort of engineering jobs on a 386 -- just had to schedule my work so autorouting would run overnight, if I picked the right settings. 500 MHz will autoroute a similar board in 10 minutes, so scheduling is a little easier, and I can try more different settings to try to get the best routing. 1.4GHz might or might not get it done in 5 minutes (CPU cycles aren't the only limiting factor!). So if I was doing that everyday, a few hundred extra for the fastest system would improve productivity be cost effective. Two 500MHz boxes and a keyboard-monitor-mouse switch would improve productivity far more...
Anyway, routing circuit boards is not my main job, and 500MHz is essentially instantaneous for everything else I do. But I remember when home computers were 8-bitters running games like the text-only D&D, and engineers got $50,000 Unix workstations, or else just a dumb terminal and a logon to the mainframe. Now the gameplayers need more power than most engineers...
The easiest example to understand of proper matrixing is the way we used to organize the factory here. We assemble circuit boards for about a dozen customers. So we used to have 5 "customer" teams, each covering one big customer or 3 or 4 little ones. But then there were also three groups of production automation (surface-mount, through-hole auto insert, and wave solder) that were shared by all the teams. So the operators of those machines had multiple bosses: their own production manager, and the customer team leaders. But we kept the lines of authority fairly clear because the production managers would allocate times on their machine and do the performance reviews for their people.
For engineering or coding, it's hard to keep the lines of authority that clear. You have various "functional" areas (e.g. mech design, electronic design, software, test, usability experts, marketing experts) and you draw people from these to form a project team. If the primary authority stays with the functional managers, then there is little motivation for people to commit themselves to a team. If real authority goes to the project leaders, then people have to contend with performance reviews written by someone who doesn't understand their specialty, and also some wind up splitting time between teams and with real trouble keeping all their bosses happy. If the managers are competent and get along with each other, they can make it work, but it's awfully easy for it to turn into a snake-pit.
If you do it right, the chart does show points where work can be done in parallel, and critical parts where a little extra resources can speed up the project quite a lot. However, with the typical management gloss over the details, the software is probably going to think that "If one woman can produce one baby in 9 months, then 9 women can produce one baby in 1 month." (Fred Brooks, The Mythical Man-Month) This results in completely unrealistic schedules.
I have worked on projects where project management software was critical. But if your organization is small enough that coders are usually doing multiple projects, then chances are the projects aren't complex enough to really benefit from it. That is, if you've only got 1/2 a coder to put on it and it's 500 hours of coding, then it's going to take half a year if you get him started right away, and longer than that if you mess around planning how you're going to allocate his 20 hours a week first...
Just wait until they toss 50 hours of high priority work at you every week, then ask why the lower priority work is all overdue...
However, most modern piloting tasks don't require deep thinking. Try doing all that, or even half of that, plus taking a star sight and working out your position with sliderule and nav tables... That's why large airplanes used to have a four man flight crew -- two to fly, one to watch the multiple engines, and one to navigate -- and small airplanes didn't use to have that many distractions. E.g., Lindbergh could center the controls and take a star sight without worrying about running into anything over the Atlantic ocean.
There some very good parallels between brain work and computing in the 60's. A good CPU would often overrun the IO devices, and since the CPU cost half a million they didn't like wait cycles. So when it had to wait for IO, it would switch to another task, until that had to wait also, etc. The analogy is to you working on one job until you find that you need information from someone else to proceed, writing e-mail to him, then switch to another job until the reply comes back. This (task switching when on hold) improves productivity. On the other hand, when you get a phone call about some project you aren't even working on today, or have to stop coding to go to a meeting about parking spaces, you lose productivity. Likewise, interrupt-driven task switching tends to reduce the number of jobs finished per hour, and only became common when the CPU's became fast and cheap enough that you could afford to waste cycles.
Since the human brain isn't getting any faster, any situation where you are frequently interrupted is going to reduce the amount or quality of work completed. Note also that there are major and minor context switches, and the cost difference is much larger than the difference between switching processes and threads. Switching to a different part of the same project requires re-loading "registers" (short term memory), but the major context of the project stays the same. Switching to something I put away last week will probably require skimming through some of the documents to remember where I left off and to refresh my memory of the overall structure.
1) Working models were originally required. The USPTO stopped routinely requiring this late in the 19th century because they ran out of storage space for all the inventions. (I think the examiner _can_ still require a working model, but it's only invoked when they really don't believe it can work -- and obviously they're quite a bit too credulous.) We need to bring this requirement back, at least to some extent -- the patent office can't store physical models, but they should ask for them, take some photos of the thing running (if it does), then ship it back. And of course software patent applications (if allowed at all) should come with all source code and a working executable on CD-R; if they can store all the paper in an application, they can certainly store these. Without this requirement, every real inventor is hostage to the Lemelson technique: look at what people need, guess what will be invented to fill that need, file a vague, broad application, then sue the real inventor later.
2) Market it or lose it: If you don't build your invention or license it to someone who can, after 2 or 3 years it should become public domain. (I thought this was once the law.) You do not promote the "progress of science and the useful arts" by letting people just sit on a good idea for 20 years. (The same thing should apply to copyrights, and to items that are removed from the market for long periods. If Disney wants to keep it's copyright on Snow White, let them either keep the videotape in stores, get the movie into theaters every couple of years, or post the MPEG on the web with a "download for non-commercial purposes only" notice. Not hold the damn thing off the market for a generation.)
3) Tighten up the definition of "invention." Under the rules of 1960, one could not patent a mathematical algorithm (LZW compression, public key encryption), or a thing of nature (any naturally occurring gene).
Not that I find checkboxes that hard to operate or anything. It sounds like a whole lot of people didn't see that checkbox, or else it was labeled with something that didn't actually tell you what is was going to do. Judges appointed by Republicans, or by that psuedo-Republican Clinton, may not agree, but I certainly don't see a checkbox marked with something like "enhance web links" as adequate notice, let alone something buried in a few pages of lawyerly gobbledygook.
I would have no issues at all with software where the download page starts with a statement such as "To pay for making this free software available, we included XXX, which will modify your internet browser by adding paid advertisements. After downloading and installing this, you will see yellow underlines under certain words while browsing the web. These are links to advertisers' web pages. The links are not from the web sites where you see them and are not endorsed by the creators of the web sites."
Or to follow up on my analogy above: "we'll install paneling free, but we're going to stamp advertisements all over everything, not just the paneling." I wouldn't go for it, but if someone thinks that's an acceptable deal...
Monitor emissions can do a lot of spying, but they will not reveal your password (unless you're using software so braindead it displays the password when you type it). Not that a shielded, encrypted keyboard would be full protection for your passwords, either. It might defeat a physical or distant (Tempest-type) tap on the keyboard cable, but if they break in, they can load a keystroke monitor program that will snag the keys after they're decoded. Remember, the keys have to be decoded before they are presented to the OS to either handle itself or pass on to various applications. It might even be possible to decode your motherboard's radio emissions to tell when it's processing a keystroke and pick that up...
Note however that this case involves a fine technical point: they got a warrant, but it wasn't the right kind of warrant for a wiretap, and this does seem to be a wiretap as far as the technology goes. On the other hand, wiretaps are especially limited because when you tap a phone, you are tapping two persons, and often one of them is not a suspect. Tapping the keyboard cable doesn't involve an innocent third party. I can see a court reasonably going either way on this one...
Anyway, the big problem is that the penalties are backwards. Cops don't go to jail for illegal searches. DA's don't get disbarred for fooling some judge into approving a warrant that doesn't really cover what they are going to do. It's pretty unlikely that the judge who signs a warrant on insufficient grounds will even get a reprimand. ALL THAT is what should happen when the cops and prosecutors step over the line -- not tossing out the evidence.
I switched from physics to engineering over 20 years ago, because the only parts of the field that were still making new theories were running into a wall when it came to verifying them. Particle physicists need a multi-billion dollar machine running for weeks to get a few flickers on the detectors. But they're in much better shape than cosmologists, who've got no way to run experiments at all, and who are perhaps drawing far too sweeping conclusions from too few observations. I salute you guys that are still hanging in there.
I've heard two extreme positions here -- and as usual, both of them are wrong. Really, it depends. The understudy idea is a good one, but there are two things that can block it from working. One is that sometimes there is only enough work in that area for one person. The other is if the "irreplaceable" person won't cooperate in training a potential replacement. Or the situation I saw once, where we had one PCB designer, not only about 1500 hrs/yr of PCB design work (he also had other skills that were in use the rest of the year), _and_ he was working at making himself really irreplaceable. I'm rather of a generalist; before we hired "joe", I was doing a little PCB design (not very well), and I obviously would have been on tap to fill in if joe had received a job offer from somewhere at 25% more... But one day when joe was on vacation, a customer called up with a question about one of joe's projects, and I found that he hadn't archived the design files for that -- or most of the other projects. After about six months of struggling to get him to document his work, we finally decided it would be best to get a backup of his hard-drives, fire him, and let me get to work trying to sort things out -- because the longer we waited, the worse that job would get!
I'd say that's the main problem. The 10% or 20% of good programmers have been carrying the rest all along, and the managers never noticed. And the bad ones get worse as they get older, while the aces are likely to be in business for themselves by the time they are 40. But if you can't judge competence in the first place, chances are that your 10 new H1-B hires include 8 or 9 duds, and 1 or 2 that could be good programmers if someone translated the problem to be solved into Hindi...
But H1-B creates the opposite distortion: to give corporations a limited group of foreign high-tech workers at lower than competitive wages. The conditions amount to indentured servitude: H1-B workers generally cannot change jobs for higher pay, so their wage rates have nothing to do with the free market. Even though according to the law, H1-B's are supposed to get the same pay as Americans, there really isn't such a thing as a "standard rate" for programmers; each job is different, and so it is very easy to manipulate things so as to substantially underpay the H1-B's. The complicated paperwork requirements favor large corporations. Basically, their campaign contributions have bought them a special uncompetitive labor market -- as long as it stays small enough not to give the labor unions a big campaign issue.
As for the labor unions, note that they still have the immigration restrictions that count for them: Mexicans can come here temporarily to do farm work that is so hard and so low paid their members would rather collect unemployment, and H1-B's aren't competing with union workers either.
The digestible parts of food are mostly chains of HCOH units. Your body burns that with O2 to produce CO2 + H2O. Some of the water is excreted and some evaporates from the lungs, skin, etc. The air system in the cabin has to capture that evaporated water before the humidity gets so high the instruments fog up. So I think you'd get enough drinking water from the dehumidifier, if the designers pay a bit of attention to the materials and arrangements so it didn't get contaminated. But if you want a shower, that's going to be in recycled sewage...
That's good enough odds to do more studies and try to pick the right spot to send a robot drilling rig to look for ice. But certain other proposals like sending out a manned expedition with one-way fuel and the equipment to electrolyze water into rocket fuel will have to wait until the robots actually find ice that is there NOW, rather than the tracks of evaporated ice deposits.
If it's anything like the Apollo missions, what will drive them crazy won't be isolation, it will be the CNN cameras watching them for two years!
At any rate, this system shouldn't need lead shielding. If I understand it, it forms many little beams (one per pixel), which are not steered and travel less than an inch. So it should only need a few volts to operate, and the most powerful photons that can form will be ultraviolet, which is easily stopped by many kinds of lead-free glass or by the opaque back cover.
IMHO, lead from electronics is far from my greatest worry about landfills. (I am not a chemist, but I do know a few things.) Metallic lead only dissolves in acids, and lead glass from CRT's should be even less soluble. There are also a few ounces of 63% lead solder in a typical computer, but people have lived with much larger quantities of lead for millenia and usually avoided poisoning themselves. There is at least 2,000 years of history for lead pipes for drinking water (from liners for Roman aquaducts starting before 100BC to American city water systems in the early 1900's). Chances are these lead pipes haven't yet been all replaced in some of our cities -- and in other cities, they ripped them out and tossed them into the nearest landfill... But you won't find lead in the water coming out of a lead pipe unless the water is far more acidic than drinking water ought to be. Landfills might be more likely to have acid groundwater, but they ought to be ensuring that acid water doesn't leak out anyhow -- and when the acidity is neutralized, the lead compounds drop out of solution and don't travel any further. I'd worry more about the more unusual metals being touted as replacements for lead in solder -- some of them have never been widely used before, and there are probably things we don't know and won't know until too many people have been exposed to them for 20 years...