I'v experienced the same thing -- lots more spam, mostly penny stock stuff with inline images.
The first thing I did was look at the email source, and write a regex on the single-image include
and put that as a email-blocking test on my spam filter, http://assp.sourceforge.net/ ASSP
this quickly put virtually all of these spams into the blacklist of the filter; after a week I
turned the regex off and ASSP has been blocking it since.
I agree with some of the other posters, however, if this gets escalated again where the
filters start becoming useless, I'll go 100% whitelist here.
I've not found too many hackable webcams. In this instance I'd suggest foregoing hackability for realiability and being able to access the camera's output easily with linux or open source tools. I found the DCS-900 series (particularly the DCS-G900, which is 802.11G) works very well with open-source tools like motion.
I've played with hawking, panasonic, and D-link. I found hawking useless. The panasonic cameras have a GREAT API that works via http requests and is well documented; however I found the cameras would hang every 2-3 weeks. The DCS is pretty basic, but once it's set up it's rock solid, and you can do the interesting stuff on a host elsewhere on the network trivially. The other nice thing (and this is hardware hacking) is that the DCS-9xx series uses a standard lens mount, so you can swap out the lens for a better one or something more suited to your purpose. In my case I swapped it out for a fisheye to provide a much larger area of coverage for site monitoring.
As far as the greater question of hackable hardware, google is definitely your friend. Most devices get torn apart pretty quickly, and if the firmware can be upgraded with a OSS solution it's pretty well known. I've been surprised at some items -- the DSL modem I recieved from the telco, for example, had OSS recipes for opening it up and upgrading parts of the network stack. Also, certain vendors get a rep for being hackable, such as the linksys WAPs mentioned elsewhere in this thread. I usually just start picking a few specific part/model numbers for the hardware I'm interested and google on hacks/firmware upgrades, teardowns, etc.
Your interpretation is incorrect. Most airspace is class E -- which is controlled airspace -- but airspace designations are primarily about visibility minimums and communications requirements, and have little bearing on flight plan or instrument flight status. Classes D, C, and B are all associated with areas in and around airports, with increasing requirements for communications (D,C, and B all require communications with the airport tower, C requires transponder, and B requires transponder and permission to enter the airspace.) So today, even the most restricted airspace (B) you can fly over without a flight plan.
So unless most airspace is declared class B, it's not really an issue. I really don't think the FAA / ATC want to deal with the millions of clearance requests, etc they'd encounter if they did something so drastic.
What's more likely is that they'll swiss-cheese the airspace with temporary flight restrictions (TFRs) around areas where the drones operate. Presumably they could become so numerous as to make private flight planning a bit difficult. Before then, however, there will probably be enough crashes with drones to result in them be forced into small saftey zones. If the Predator is any indication, there will be many, many crashes as UAVs get used more extensively -- which would totally undermine any safety-selling approach that might be tried.
You asked what software we used... so I use vim. Obviously it doesn't fit your criteria, (no WYSIWYG, etc.)
I tried a number of things and it came down to having a decent text editor. Organization is more important than the other fancy features in my situation, and being able to grep a years worth of notes to find something in seconds more than made up for the sacrifice of being text-only.
Seeing as how this exact situation happened to me this week, I can provide some light on the/32 IP address issue.
In my case, I moved a server to a new colo facility. Most facilities have an IP block, and you get assigned an IP from it. Six months or a year ago that IP might have belonged to someone else. For me, it turned out in February a spammer installed a server at the colo, spammed from that server for a single day before the colo ISP turned them off. That IP got listed in Spamhaus; in the beginning of June I was assigned that IP.
So, I ended up with a Spamhaus listing for my mail server's IP address -- and _I_ can't get it removed. Spamhaus expects the colo operator to contact them (which they did on my request) but even there, if the blacklist operator doesn't like the ISP/colo people, they can ignore the request.
Fortunately Spamhaus listened and I got the record for my IP removed. But this showed me it was trivial for a non-spammer to inherit a blacklisted IP. I've added doing DNSBL checks on colo-assigned IP addresses for future moves to prevent any future issues.
Funny, I just researched this in April, and decided on Verizon. Depending on your area, Verizon 1XRTT or EVDO will usually work.
On EVDO I'm had real-world speeds routinely hit better than 300Kbps. A couple interesting things though. One, the service clearly does some sort of optimization for web access -- small files will hit the peak speeds, but the larger the file the slower the download rate. You can actually see this in real time if you used a download client that shows transfer rates -- starts off screaming about after about 50-100K bytes, it "decays" to about 150Kbps -- still way better than dialup or even the 2B ISDN line I've normally used.
The other is the usual warning about overused capacity; I definitely notice lower transfer rates at peak times of day -- but at 2AM I'll see better than 400 Kbps quite often.
Upload bandwidth is quite a bit smaller, around 40-60Kbps.
All of these figures are for EVDO. For the more widely available 1XRTT I've had about 40-60Kbps connections.
Anyway, I'd say it's great for web browsing, fine for email as long as there's no big attachments. Verizon has a lot of scary language in their AUP, but there's some folks who frequent the EVDO message boards that have recorded insane levels of bandwidth use (like 20-30GB/mo) without a peep from verizon so far.
One negative: latency is horrible. In the wee morning hours I'll see ~ 150ms ping times to google; during busy times pings climb to 400+ ms, with an occasional martian at > 1000ms.
"at this time XFCE 4.2 is a better GNOME than GNOME itself."
This comment is dead on.
Gnome seems to have an awful case of the Ahab from Moby Dick syndrome: they seem hell-bent on copying a certain whale of an OS company; duplicating good features is fine, but they seem to be trying too hard to duplicate much of the bad stuff as well. After Gnome 1.4 things seemed to move very much in this direction. Unfortunately one of my favorite Gnome apps, Gnumeric, seems headed this way as well, tacitly endorsing VBA as a scripting language (though kudos for supporting other bindings such as python.) You can see this influence in related projects such as Evolution as well. Presumably there's some sort of pressure to do this in the name of easy transitions, but I still think it's the wrong thing to do.
XFCE definitely seems to be more about a philosophy (use the unix small programs together model, and keeping it simple) than copying, and if they're copying anything it's closer to the Mac feel. In any case, it's not so bent on bloat and copying something with many features many don't want to begin with.
Hopefully we'll still be able to use a lot of the gnome apps under XFCE on future versions of slackware without too much additional work.
OK, this isn't a crack per se, and likely even more illegal -- but in my area at least, the digital stuff goes down occasionally and everyone with multi-mode phones automatically switch back to AMPS. Doesn't seem like it'd be too hard to force digital down intentionally with some willful interference.
Only the nerds among us are going to see the AMPS icon lit and think "oh, I better not say anything secret."
Exactly. The experiments produce a repeatable change on the order of 5 parts per 10,000 on data sets of millions of random trials. Many statistical analysts accept something as being signficiant when the probability due to chance is under 5%, ie P.05.
The studies done by PEAR consistently produce results that are P.05. This could be due to some unknown experimental design flaw, but it's pretty interesting nonetheless.
There are some skeptic counterclaims to be noted, however. Apparently something on the order of half the "good" results came from one key influencer -- but even dismissing these results puts P right at the 0.05 boundary, on millions of trials.
Anyway -- if the results are real it's fascinating, and if not it's a pretty interesting case for experiment design reviews to figure out the source for bias.
I stumbled across this project years ago as I was researching "real" random number generation for encryption work. I found a very peculiar disclaimer from some manufacturers that claimed that the output would not be random is used in Psi research.
From that I found multiple pointers to a book, Margins of Reality, by Jahn and Dunne. It details research done at the Princeton Engineering Anomalies Research lab. They basically run millions of RNG trials with people trying to influence the result, and they get pretty much statistically provable effects, but at a very low level (something like a 5 parts per 10,000 deviation from the norm.) What's freaky is it's so consistent they've gotten to the point that they can tell you which test subject is influencing things by the results. Very freaky stuff.
Anyway, even if you're a die-hard prove-it-to-me science buff, the research results described in the book will really make you ponder how well we understand things, particularly RNGs and rigorous test procedures, if nothing else.
Pay attention to context in these stories, ATI's the lead in standalone, swapping places with Nvidia (just slightly). Overall Intel is still on top.
My company is the source of this data and doesn't release it, but on a regular basis portions get leaked, often to present a particular picture that isn't entirely complete. (Usually the leaks come from someone with a financial interest in driving perceptions on way or the other, and not the graphics companies themselves.)
I have several of the iGesture pads, and love them. I got the keyboard, but frankly it was too far of a stretch for me. I supposed I would have gotten used to it if I'd given it more time -- others have raved about the keyboard as well -- but it's simply a flat lexan surface with a tiny bump on two letters on the home row. There's a very slight roughness to it so that's it's not glass smooth, but that's it.
The problem I had with the keyboard primarily is that I have pretty lazy fingers, and if you don't touch the surface correctly -- that is, if you rest one or two fingers on it -- the keyboard interprets this as typing. It's a zero-pressure interface, and if fact it works best with the lightest of touches. I'm sure with more training it'd get there. The keyboard has a much larger gesture library on it.
It's also somewhat large (I use a happy hacking keyboard along with the iGesture as my primary interface, so everything else is huge in comparison.)
Both the keyboard and the pad can have additional gestures added to them, if you wish. I primarily use the pad for navigating very large spreadsheets and editing text files, and it's makes it a breeze as two fingers it behaves as a mouse, and one finger it sends arrow keystrokes; this feature alone sold me on it.
You mentioned trackpads, but the stuff from fingerworks goes a bit beyond this and supports a gesturing interface. I've used one for about a year and bought several as backups (the thing is incredibly durable) just because I know I'll never be able to live without it again.
I also recall a demo -- sorry, no link -- that used webcam-style cameras to watch eye movements and use that as part of the active window selection process. There was another demo at Intel Developer Forum last year that did something similar, but turned off a notebook's display when you weren't looking at it to save power.
This is a surprisingly political issue (I should know, I count these things among other parts of my job, and hear back about it when something is in the wrong "slot.")
Anyway, the usually accepted answers are 7th or 8th generation. There's some definitely blending of lines between the microarchitectures. So you had Pentium (586) and Pentium Pro, Pentium II did share quite a bit but were also fairly different (Pentium Pro, for example, was actually a multichip module.) P3 shared a bit as well, but P4 is clearly a totally different animal.
So if you count 8086 as first generation, it's 8086 - 1 80286 - 2 80386 - 3 80486 - 4 Pentium - 5 Pentium Pro/PII/P3 - 6 Pentium 4 - 7
If you break it between Pro and PII/P3, then it's 8.
The reason this is political is because AMD also has their generations, which were identical to Intel through about the Pentium time frame, but then became radically different microarchitecturally, so you have claims that company "A" is ahead of company "I" generationally. Then, throw Itanium into the counting, and you have to ask WTF generation it is.
Realistically, there's quite a few more microarchitectural tweaks that go on during a given generation than usually are acknowledged, so the lines get blurred even further -- today's P4 is a fair bit different from the original P4. My opinion is the generational nomenclature has lost almost all meaning. For example, technically Pentium M shares a lot of commonality as Pentium 3, but there's been so many changes to fundamental peices of the architecture that it really qualifies as a new and different animal that in many ways is both ahead and behind P4 -- so calling it an eighth, nineth or sixth-generation CPU can all be argued.
FYI, this is already done in a way, as far as Intel helping to build solar panels goes.
Astropower is a US solar panel manufacturer that gets many of its solar cell silicon wafers through recycling programs with Intel. TI, etc. See here. They basically take bad/test wafers, clean them up, and use them for silicon solar cells.
I'm building a 100% solar home and already have a kW of capacity installed, and went with Astropower for several reasons, the above included.
It was chip supplier in question was IIT (Integrated Information Technologies), they later dropped out of the graphics (and math co-processor business) and retargeted themselves at video, becoming 8X8. Later they evolved from video oriented at videophones to move into VoIP.
The 2D ZD WinBench had a string "The quick brown fox..." it rendered in different colors and sizes using the Windows GDI, and the IIT BIOS embedded it. I believe the parts were still ISA based, so embedded the string in a ROM on the card would actually offer more bandwidth than passing it over the ISA bus did.
In the mid-1990s my company did quite a bit of graphics card testing (still does, but it was much higher profile back then.) It was pretty routine for us to get baked drivers (and there were some very impressive cheats) less routine, but still common was to get a board with a BIOS cheat, which would do anything from altering its own board timings to be out of spec (sort of "overclocked out of the box") to running code that would adjust the PC's heartbeat interrupt to slow the clock ticks to make the board appear faster if benchmarked using the PC's own clock.
In the end the best solution we came up with -- because we worked with a lot of alpha/beta silicon since we tracked chips more than boards -- was to more or less formalize the cheats and what was/wasn't permitted, and also to give the companies that submitted alpha/beta hardware to pull the results before publication, so that if one company pulled a fast one, the others that would be look bad in comparison simply wouldn't be compared; this resulted in a sort of a stalemate of cheating.
The most extreme (but permitted) cheat I ever encountered had the company involved paying over $100,000 to have a custom graphics driver written overnight that incorporated an optimized version of parts of the DirectX rendering engine (this was ~ DX5 era). When they found out their primary competitors pulled their boards from testing, you can imagine they were less than pleased.
The point of all this: a competent testing lab, particularly part of a magazine "shootout," should be well aware that cheating is taking place, and prepared to identify major cheats. Back in the heyday of PC Magazine in the mid-90s, their benchmark people were top notch and the benchmark ran a considerable number of cheating tests to clear out the more bogus attempts.
Oh, and you can be pretty assurred your competitors are doing the same things you are.
Assuming he's licensed (which this guy -- but not lawn chair guy -- is).
The basic rule for right of way for aircraft is the the lesser manuverable craft has the right of way.
So it goes like baloons, airships, airplanes, helicopters.
Also, the "lawn chair guy" is dead, of suicide.
It's definitely NOT an urban legend, I remember when he first did it -- made national news. The story still routinely pops up in pilot magazines.
And to echo the cluster ballooning guy's advice: don't try this at home without training. I'm a licensed airplane pilot, and have crewed on hot air balloons a few times in New Mexico during their annual ballooning orgy. IMHO piloting balloons takes more skill as they're so much less manuverable you need to be considerably better at planning. Figuring out you don't have that skill while airborne is a bad thing.
Heighting the terror factor is that when you're screwed you usually know about it well before the actual you're-screwed event takes place, and get to experience it in slow motion.
I couldn't agree more. XFCE feels like what Gnome used to be in the 1.4 days, fast and light and "just works." The other really nice trait about the XFCE 4 rewrite is that it's a DE built on the unix philosophy of simple, small components. For example, I have no need for a taskbar, and one just comments out that particular module in the startup script. There's still a few areas that could stand a bit of polish, but I've switched from gnome to XFCE 4 (and now 4.2) and am very pleased with the system overall.
It's also worth highlighting that there is a graphical installer available for XFCE 4.2, which made installing it beyond simple.
There was a serious, big-time oversupply of chip sets last quarter. This shortage could likely be the chip set suppliers being cautious -- the last time something like this happened we were swimming in finished notebooks for more than a quarter and it had a fairly negative impact on Intel and AMD's operations for about six months.
I'll add one more Canon vote. While the linux drivers aren't quite up to the level of those for Epson, I've had a couple Epsons die immediately after ink cartridge changes (all using Epson branded carts.) Once the print head is toast, it's over for the entire printer, and there's been a number of reports that any time Epson branded ink is mixed with anything else, the print head clogs -- in my case, it appears the ink formulations probably had changed between carts, as they came from different regions.
Even relatively low-end Canons support replacable print heads (not cheap, but more environmentally friendly than disposable printers.)
As others have mentioned, Canon uses simple, clear, un-chipped ink tanks that are trivial to refill. The Canon i560 I use now is the first inkjet where I don't feel I'm getting screwed on supplies. Print quality under linux is pretty good -- top quality isn't as good as Epson at the moment, but very acceptable, and the cost and reliability of the Canon more and outweighs the slight print quality difference for me, at least.
A very long time, indeed. It's been well documented that the feds sent a robot armed with a shotgun to deliver a "negotiation" telephone to Randy Weaver during the Ruby Ridge incident -- and that was 1992.
Just to be clear since he's from my home state and was a really decent fellow -- the radical right makes Goldwater look like Kennedy because Goldwater was a true conservative and closer to a libertarian than a radical right-winger, in spite of media portrayals.
Barry was actively supporting both gay rights to serve in the military and also abortion rights immediately before he died. I think the Republican party he represented died with him, since the current one in no way resembles the party when Goldwater was running for president.
Tracking dish, knowledge of the frequencies in use, signal generator and amplifier and you're pretty much there.
Of course if they're using DS spread spectrum and they don't have the spreading code, it could be considerably harder, though turning up the power sufficiently would probably desensitise the front end of the satellite enough to stop it from working.
Slashdot Mirror
I'v experienced the same thing -- lots more spam, mostly penny stock stuff with inline images. The first thing I did was look at the email source, and write a regex on the single-image include and put that as a email-blocking test on my spam filter, http://assp.sourceforge.net/ ASSP this quickly put virtually all of these spams into the blacklist of the filter; after a week I turned the regex off and ASSP has been blocking it since. I agree with some of the other posters, however, if this gets escalated again where the filters start becoming useless, I'll go 100% whitelist here.
I've not found too many hackable webcams. In this instance I'd suggest foregoing hackability for realiability and being able to access the camera's output easily with linux or open source tools. I found the DCS-900 series (particularly the DCS-G900, which is 802.11G) works very well with open-source tools like motion.
I've played with hawking, panasonic, and D-link. I found hawking useless. The panasonic cameras have a GREAT API that works via http requests and is well documented; however I found the cameras would hang every 2-3 weeks. The DCS is pretty basic, but once it's set up it's rock solid, and you can do the interesting stuff on a host elsewhere on the network trivially. The other nice thing (and this is hardware hacking) is that the DCS-9xx series uses a standard lens mount, so you can swap out the lens for a better one or something more suited to your purpose. In my case I swapped it out for a fisheye to provide a much larger area of coverage for site monitoring.
As far as the greater question of hackable hardware, google is definitely your friend. Most devices get torn apart pretty quickly, and if the firmware can be upgraded with a OSS solution it's pretty well known. I've been surprised at some items -- the DSL modem I recieved from the telco, for example, had OSS recipes for opening it up and upgrading parts of the network stack. Also, certain vendors get a rep for being hackable, such as the linksys WAPs mentioned elsewhere in this thread. I usually just start picking a few specific part/model numbers for the hardware I'm interested and google on hacks/firmware upgrades, teardowns, etc.
Your interpretation is incorrect. Most airspace is class E -- which is controlled airspace -- but airspace designations are primarily about visibility minimums and communications requirements, and have little bearing on flight plan or instrument flight status. Classes D, C, and B are all associated with areas in and around airports, with increasing requirements for communications (D,C, and B all require communications with the airport tower, C requires transponder, and B requires transponder and permission to enter the airspace.) So today, even the most restricted airspace (B) you can fly over without a flight plan.
So unless most airspace is declared class B, it's not really an issue. I really don't think the FAA / ATC want to deal with the millions of clearance requests, etc they'd encounter if they did something so drastic.
What's more likely is that they'll swiss-cheese the airspace with temporary flight restrictions (TFRs) around areas where the drones operate. Presumably they could become so numerous as to make private flight planning a bit difficult. Before then, however, there will probably be enough crashes with drones to result in them be forced into small saftey zones. If the Predator is any indication, there will be many, many crashes as UAVs get used more extensively -- which would totally undermine any safety-selling approach that might be tried.
You asked what software we used... so I use vim.
Obviously it doesn't fit your criteria, (no WYSIWYG, etc.)
I tried a number of things and it came down to having a decent text editor. Organization is more important than the other fancy features in my situation, and being able to grep a years worth of notes to find something in seconds more than made up for the sacrifice of being text-only.
Seeing as how this exact situation happened to me this week, I can provide some light on the /32 IP address issue.
In my case, I moved a server to a new colo facility. Most facilities have an IP block, and you get assigned an IP from it. Six months or a year ago that IP might have belonged to someone else. For me, it turned out in February a spammer installed a server at the colo, spammed from that server for a single day before the colo ISP turned them off. That IP got listed in Spamhaus; in the beginning of June I was assigned that IP.
So, I ended up with a Spamhaus listing for my mail server's IP address -- and _I_ can't get it removed. Spamhaus expects the colo operator to contact them (which they did on my request) but even there, if the blacklist operator doesn't like the ISP/colo people, they can ignore the request.
Fortunately Spamhaus listened and I got the record for my IP removed. But this showed me it was trivial for a non-spammer to inherit a blacklisted IP. I've added doing DNSBL checks on colo-assigned IP addresses for future moves to prevent any future issues.
Funny, I just researched this in April, and decided on Verizon. Depending on your area, Verizon 1XRTT or EVDO will usually work.
On EVDO I'm had real-world speeds routinely hit better than 300Kbps. A couple interesting things though. One, the service clearly does some sort of optimization for web access -- small files will hit the peak speeds, but the larger the file the slower the download rate. You can actually see this in real time if you used a download client that shows transfer rates -- starts off screaming about after about 50-100K bytes, it "decays" to about 150Kbps -- still way better than dialup or even the 2B ISDN line I've normally used.
The other is the usual warning about overused capacity; I definitely notice lower transfer rates at peak times of day -- but at 2AM I'll see better than 400 Kbps quite often.
Upload bandwidth is quite a bit smaller, around 40-60Kbps.
All of these figures are for EVDO. For the more widely available 1XRTT I've had about 40-60Kbps connections.
Anyway, I'd say it's great for web browsing, fine for email as long as there's no big attachments.
Verizon has a lot of scary language in their AUP, but there's some folks who frequent the EVDO message boards that have recorded insane levels of bandwidth use (like 20-30GB/mo) without a peep from verizon so far.
One negative: latency is horrible. In the wee morning hours I'll see ~ 150ms ping times to google; during busy times pings climb to 400+ ms, with an occasional martian at > 1000ms.
"at this time XFCE 4.2 is a better GNOME than GNOME itself."
This comment is dead on.
Gnome seems to have an awful case of the Ahab from Moby Dick syndrome: they seem hell-bent on copying a certain whale of an OS company; duplicating good features is fine, but they seem to be trying too hard to duplicate much of the bad stuff as well. After Gnome 1.4 things seemed to move very much in this direction. Unfortunately one of my favorite Gnome apps, Gnumeric, seems headed this way as well, tacitly endorsing VBA as a scripting language (though kudos for supporting other bindings such as python.) You can see this influence in related projects such as Evolution as well. Presumably there's some sort of pressure to do this in the name of easy transitions, but I still think it's the wrong thing to do.
XFCE definitely seems to be more about a philosophy (use the unix small programs together model, and keeping it simple) than copying, and if they're copying anything it's closer to the Mac feel. In any case, it's not so bent on bloat and copying something with many features many don't want to begin with.
Hopefully we'll still be able to use a lot of the gnome apps under XFCE on future versions of slackware without too much additional work.
OK, this isn't a crack per se, and likely even more illegal -- but in my area at least, the digital stuff goes down occasionally and everyone with multi-mode phones automatically switch back to AMPS. Doesn't seem like it'd be too hard to force digital down intentionally with some willful interference.
Only the nerds among us are going to see the AMPS icon lit and think "oh, I better not say anything secret."
Exactly. The experiments produce a repeatable change on the order of 5 parts per 10,000 on data sets of millions of random trials. Many statistical analysts accept something as being signficiant when the probability due to chance is under 5%, ie P .05.
.05. This could be due to some unknown experimental design flaw, but it's pretty interesting nonetheless.
The studies done by PEAR consistently produce results that are P
There are some skeptic counterclaims to be noted, however. Apparently something on the order of half the "good" results came from one key influencer -- but even dismissing these results puts P right at the 0.05 boundary, on millions of trials.
Anyway -- if the results are real it's fascinating, and if not it's a pretty interesting case for experiment design reviews to figure out the source for bias.
I stumbled across this project years ago as I was researching "real" random number generation for encryption work. I found a very peculiar disclaimer from some manufacturers that claimed that the output would not be random is used in Psi research.
From that I found multiple pointers to a book, Margins of Reality, by Jahn and Dunne. It details research done at the Princeton Engineering Anomalies Research lab. They basically run millions of RNG trials with people trying to influence the result, and they get pretty much statistically provable effects, but at a very low level (something like a 5 parts per 10,000 deviation from the norm.) What's freaky is it's so consistent they've gotten to the point that they can tell you which test subject is influencing things by the results. Very freaky stuff.
Anyway, even if you're a die-hard prove-it-to-me science buff, the research results described in the book will really make you ponder how well we understand things, particularly RNGs and rigorous test procedures, if nothing else.
Pay attention to context in these stories, ATI's the lead in standalone, swapping places with Nvidia (just slightly). Overall Intel is still on top.
My company is the source of this data and doesn't release it, but on a regular basis portions get leaked, often to present a particular picture that isn't entirely complete. (Usually the leaks come from someone with a financial interest in driving perceptions on way or the other, and not the graphics companies themselves.)
I have several of the iGesture pads, and love them. I got the keyboard, but frankly it was too far of a stretch for me. I supposed I would have gotten used to it if I'd given it more time -- others have raved about the keyboard as well -- but it's simply a flat lexan surface with a tiny bump on two letters on the home row. There's a very slight roughness to it so that's it's not glass smooth, but that's it.
The problem I had with the keyboard primarily is that I have pretty lazy fingers, and if you don't touch the surface correctly -- that is, if you rest one or two fingers on it -- the keyboard interprets this as typing. It's a zero-pressure interface, and if fact it works best with the lightest of touches. I'm sure with more training it'd get there. The keyboard has a much larger gesture library on it.
It's also somewhat large (I use a happy hacking keyboard along with the iGesture as my primary interface, so everything else is huge in comparison.)
Both the keyboard and the pad can have additional gestures added to them, if you wish. I primarily use the pad for navigating very large spreadsheets and editing text files, and it's makes it a breeze as two fingers it behaves as a mouse, and one finger it sends arrow keystrokes; this feature alone sold me on it.
You mentioned trackpads, but the stuff from fingerworks goes a bit beyond this and supports a gesturing interface. I've used one for about a year and bought several as backups (the thing is incredibly durable) just because I know I'll never be able to live without it again.
I also recall a demo -- sorry, no link -- that used webcam-style cameras to watch eye movements and use that as part of the active window selection process. There was another demo at Intel Developer Forum last year that did something similar, but turned off a notebook's display when you weren't looking at it to save power.
This is a surprisingly political issue (I should know, I count these things among other parts of my job, and hear back about it when something is in the wrong "slot.")
Anyway, the usually accepted answers are 7th or 8th generation. There's some definitely blending of lines between the microarchitectures. So you had Pentium (586) and Pentium Pro, Pentium II did share quite a bit but were also fairly different (Pentium Pro, for example, was actually a multichip module.) P3 shared a bit as well, but P4 is clearly a totally different animal.
So if you count 8086 as first generation, it's
8086 - 1
80286 - 2
80386 - 3
80486 - 4
Pentium - 5
Pentium Pro/PII/P3 - 6
Pentium 4 - 7
If you break it between Pro and PII/P3, then it's 8.
The reason this is political is because AMD also has their generations, which were identical to Intel through about the Pentium time frame, but then became radically different microarchitecturally, so you have claims that company "A" is ahead of company "I" generationally. Then, throw Itanium into the counting, and you have to ask WTF generation it is.
Realistically, there's quite a few more microarchitectural tweaks that go on during a given generation than usually are acknowledged, so the lines get blurred even further -- today's P4 is a fair bit different from the original P4. My opinion is the generational nomenclature has lost almost all meaning. For example, technically Pentium M shares a lot of commonality as Pentium 3, but there's been so many changes to fundamental peices of the architecture that it really qualifies as a new and different animal that in many ways is both ahead and behind P4 -- so calling it an eighth, nineth or sixth-generation CPU can all be argued.
OK, so the _used_ to take bad/test wafer... etc.
Crap, they were one of the few US companies really committed to doing solar power right. Guess right != profitable.
FYI, this is already done in a way, as far as Intel helping to build solar panels goes.
Astropower is a US solar panel manufacturer that gets many of its solar cell silicon wafers through recycling programs with Intel. TI, etc. See here. They basically take bad/test wafers, clean them up, and use them for silicon solar cells.
I'm building a 100% solar home and already have a kW of capacity installed, and went with Astropower for several reasons, the above included.
It was chip supplier in question was IIT (Integrated Information Technologies), they later dropped out of the graphics (and math co-processor business) and retargeted themselves at video, becoming 8X8. Later they evolved from video oriented at videophones to move into VoIP.
The 2D ZD WinBench had a string "The quick brown fox..." it rendered in different colors and sizes using the Windows GDI, and the IIT BIOS embedded it. I believe the parts were still ISA based, so embedded the string in a ROM on the card would actually offer more bandwidth than passing it over the ISA bus did.
Hell, not just drivers... BIOS!
In the mid-1990s my company did quite a bit of graphics card testing (still does, but it was much higher profile back then.) It was pretty routine for us to get baked drivers (and there were some very impressive cheats) less routine, but still common was to get a board with a BIOS cheat, which would do anything from altering its own board timings to be out of spec (sort of "overclocked out of the box") to running code that would adjust the PC's heartbeat interrupt to slow the clock ticks to make the board appear faster if benchmarked using the PC's own clock.
In the end the best solution we came up with -- because we worked with a lot of alpha/beta silicon since we tracked chips more than boards -- was to more or less formalize the cheats and what was/wasn't permitted, and also to give the companies that submitted alpha/beta hardware to pull the results before publication, so that if one company pulled a fast one, the others that would be look bad in comparison simply wouldn't be compared; this resulted in a sort of a stalemate of cheating.
The most extreme (but permitted) cheat I ever encountered had the company involved paying over $100,000 to have a custom graphics driver written overnight that incorporated an optimized version of parts of the DirectX rendering engine (this was ~ DX5 era). When they found out their primary competitors pulled their boards from testing, you can imagine they were less than pleased.
The point of all this: a competent testing lab, particularly part of a magazine "shootout," should be well aware that cheating is taking place, and prepared to identify major cheats. Back in the heyday of PC Magazine in the mid-90s, their benchmark people were top notch and the benchmark ran a considerable number of cheating tests to clear out the more bogus attempts.
Oh, and you can be pretty assurred your competitors are doing the same things you are.
Assuming he's licensed (which this guy -- but not lawn chair guy -- is).
The basic rule for right of way for aircraft is the the lesser manuverable craft has the right of way.
So it goes like baloons, airships, airplanes, helicopters.
Also, the "lawn chair guy" is dead, of suicide.
It's definitely NOT an urban legend, I remember when he first did it -- made national news. The story still routinely pops up in pilot magazines.
And to echo the cluster ballooning guy's advice: don't try this at home without training. I'm a licensed airplane pilot, and have crewed on hot air balloons a few times in New Mexico during their annual ballooning orgy. IMHO piloting balloons takes more skill as they're so much less manuverable you need to be considerably better at planning. Figuring out you don't have that skill while airborne is a bad thing.
Heighting the terror factor is that when you're screwed you usually know about it well before the actual you're-screwed event takes place, and get to experience it in slow motion.
I couldn't agree more. XFCE feels like what Gnome used to be in the 1.4 days, fast and light and "just works." The other really nice trait about the XFCE 4 rewrite is that it's a DE built on the unix philosophy of simple, small components. For example, I have no need for a taskbar, and one just comments out that particular module in the startup script. There's still a few areas that could stand a bit of polish, but I've switched from gnome to XFCE 4 (and now 4.2) and am very pleased with the system overall.
It's also worth highlighting that there is a graphical installer available for XFCE 4.2, which made installing it beyond simple.
There was a serious, big-time oversupply of chip sets last quarter. This shortage could likely be the chip set suppliers being cautious -- the last time something like this happened we were swimming in finished notebooks for more than a quarter and it had a fairly negative impact on Intel and AMD's operations for about six months.
I'll add one more Canon vote. While the linux drivers aren't quite up to the level of those for Epson, I've had a couple Epsons die immediately after ink cartridge changes (all using Epson branded carts.) Once the print head is toast, it's over for the entire printer, and there's been a number of reports that any time Epson branded ink is mixed with anything else, the print head clogs -- in my case, it appears the ink formulations probably had changed between carts, as they came from different regions.
Even relatively low-end Canons support replacable print heads (not cheap, but more environmentally friendly than disposable printers.)
As others have mentioned, Canon uses simple, clear, un-chipped ink tanks that are trivial to refill. The Canon i560 I use now is the first inkjet where I don't feel I'm getting screwed on supplies. Print quality under linux is pretty good -- top quality isn't as good as Epson at the moment, but very acceptable, and the cost and reliability of the Canon more and outweighs the slight print quality difference for me, at least.
A very long time, indeed. It's been well documented that the feds sent a robot armed with a shotgun to deliver a "negotiation" telephone to Randy Weaver during the Ruby Ridge incident -- and that was 1992.
Just to be clear since he's from my home state and was a really decent fellow -- the radical right makes Goldwater look like Kennedy because Goldwater was a true conservative and closer to a libertarian than a radical right-winger, in spite of media portrayals.
Barry was actively supporting both gay rights to serve in the military and also abortion rights immediately before he died. I think the Republican party he represented died with him, since the current one in no way resembles the party when Goldwater was running for president.
Anyone remember Captain Midnight and the HBO incident?
Tracking dish, knowledge of the frequencies in use, signal generator and amplifier and you're pretty much there.
Of course if they're using DS spread spectrum and they don't have the spreading code, it could be considerably harder, though turning up the power sufficiently would probably desensitise the front end of the satellite enough to stop it from working.