Microsoft is right. Most applications should never have administrator privileges, not even during
installation. It's way past time to tighten the screws.
NVidia may just put a CPU or two in their graphics chips. They already have more transistors in the GPU than Intel has in many of their CPUs. They could license a CPU design from AMD.
A CPU design today is a file of Verilog, so integrating that onto a part with a GPU isn't that big a deal.
The Great Brass Brain, an analog tide predictor. It was built in 1910, and used until 1966, for regular tide predictions.
The Bay Model, a working 1.5 acre model of water flow in San Francisco Bay. Built in 1956, in use until 2000. (You can still visit, but it's not used as a research tool any more.)
SCEPTRON,
a mechanical filter bank of quartz fibres which could record and play spectra onto photographic film. This was trainable as a speech recognition system. Early 1960s.
The Iconarama., the USAF's Etch-A-Sketch. This was one of the first large screen displays, basically a plotter/slide projector combo. It could write, but not erase selectively, so units were used in pairs, allowing a redraw by the unit not projecting, then a lamp switch. 1950s.
We need the FBI Baltimore office taken out of the business of distributing child porn and put on this problem. After ten years of work, they've arrested over 6,000 people.
How many computer criminals have they arrested? The Department of Justice doesn't seem to provide useful statistics, but it looks like the number per year is in the 10-100 range.
This is backwards, given the relative size of the problems.
Part of the problem is that the FBI has a measurement bias against white-collar crime. See the FBI Crime Statistics page. Violent crimes are counted if they are reported; white collar crimes are only counted if there's an arrest.
The music industry doesn't really matter, but the decline of newspaper reporting is important to society.
There are too few reporters out there digging. Bloggers don't help; they're mostly mouthing off, not out gathering unpublished information.
Call your local newspaper and ask them how many full time reporters they have on the street. If they have ten, you're very lucky. There are very few newspapers left with big reporting staffs, and it shows.
News is what someone doesn't want published. All else is publicity.
This matters. Who's going to the city council meeting and reporting when they do something stupid? Who's tracking the state legislature? Who didn't report that FEMA's director was out to lunch before Katrina? What's not being reported now that's going to lead to big trouble in a year or two?
Combustion instability is an old problem with rocket engines. The Saturn V main engine had serious combustion instability problems, which were fixed by trial and error testing. The Apollo booster people had to resort to setting off small bombs inside engines on test stands to induce instability, then trying different patterns of holes in the plates the distributed fuel to find a stable configuration.
The SR-71 engine had serious combustion instability. That, too, was fixed with something of a hack, an automated "sympathetic unstart" system which, when one engine had a stall, would stall the other one, then restart both.
Better simulation tools in that area can't hurt. Not many big supersonic engines are designed any more. As Scott Crossfield pointed out just before he died a few years ago, every aircraft that went significantly over Mach 3 is now in a museum.
The industry is going at this all wrong. There are only a few players left, and they're all crooks. We need a consortium of companies with spam problems to hire Kroll, Blackwater, or one of the other big international security companies to deal with the people behind the problem.
If 5% of the money spent on dealing with spam went into finding the people behind it and making them go away, the problem would go away.
The poll defines "top academic scientist" as a reader of Nature.
Good point. Nature suffers when it goes outside its field. Their articles on biology tend to be excellent, but when they cover computing, the articles are weak, and Nature articles on economics are just laughable.
Most of the "ubiquitous computing" ideas are silly. There's all this information collection, but the systems don't have the actuators or smarts to do much with the information except bother some human.
Something you can buy right now, yet few buildings have, is really good HVAC control. You can get air sensors that sense temperature, humidity, CO, CO2, and particulates. You can get heating units, fans, dampers, and chillers that will talk to a network. You can get control systems that can manage all this to provide an optimal indoor environment as occupants come and go. A system like this will lower HVAC costs. Yet such systems are rare.
We still don't have good cleaning robots. The iRobot Scooba is about as good as it gets, but it's very dumb, frequently gets stuck, and can't refill, clean, or recharge itself.
Most of the "kitchen automation" stuff is just inventory control, not automated cooking.
The "ubiquitous computing" people haven't even been able to deliver a good meeting room automation system, one that gets lights, audio, and projector to play well together.
Educational games come in two basic flavors. In the first, the "education" consists of puzzles which have to be solved to advance gameplay. These are basically drill-and-practice programs with gameplay wrapped around them. If it makes dull drill and practice tolerable, why not? Most such games are trivial. There were many such games in the DOS era; today they're in Flash.
Try Type Type Revolution, which is exactly what you think it is.
The other class of educational games are simulators, designed to teach some skill. The curriculum has to be adapted to the game, rather than the other way round, and teachers hate that.
Simulators are great for teaching complex skills which require coordinating multiple goals, like landing an aircraft or running a business. They're not useful for preparing students to take written tests. So the military likes them; most educators don't.
Decent keyboard, bigger than the OLPC, which is for kid-sized hands.
WiFi, Bluetooth.
Ports: 100baseT, USB, mini-VGA for projectors.
Battery life > 12 hours of useful operation.
CPU speed around 0.5 to 1.0 GHz.
0.5 GB of RAM.
20 GB of disk/flash.
No swapping to disk/flash.
Need not be able to play DVDs or 3D games.
Price around $300.
It looks like machines in that range are going to catch on. It's a good time to be looking at major open source projects and de-bloating them. Things like browser pre-fetch and page image cacheing, which never did that much for performance but bloated memory use, should go. Make sure that installs don't leave unused junk behind.
In a situation like this, HP should recall the product and reissue a replacement product with a new part number to distinguish old product from new product.
More Roland the Plogger blogspam, driving traffic to his useless ad-laden blog. To get around the block on links to his own site, he's now submitting links disguised via "tinyurl".
I took a look through the site, and got to the "Enter Credit Card Number" point without seeing the name and address of the business.
That's a criminal offense if selling into California: Before accepting any payment or processing any debit or credit charge or funds transfer, the vendor shall disclose to the buyer in writing or by electronic means of communication, such as e-mail or an on-screen notice, the vendor's return and refund policy, the legal name under which the business is conducted and, except as provided in paragraph (3), the complete street address from which the business is actually conducted....
(g) Any violation of the provisions of this section is a misdemeanor punishable by imprisonment in the county jail not exceeding six months, by a fine not exceeding one thousand dollars ($1,000), or by both that imprisonment and fine.
Let's see what we can find out.
WHOIS gives us:
Buzzelli, David
Memsen
3604 SE Powell Valley Road
#267
Gresham, Oregon 97080
United States
(503) 667-3136
That's a start. More info is available if you dig.
Here's
my driveway in Google StreetView. Note the license plate. Coverage of this house is very good. The highest resolution images are available in the aerials and in the street views. Plus it's a corner lot, and there's full coverage from both streets.
This is from one of Google's earliest batches of images. I'm only a few miles from Google HQ, and they started by thoroughly covering the nearby areas.
First, this isn't even a final court decision; it's just some preliminary motion practice, during which the claims are narrowed.
Second, Friendfinder went way beyond being a passive conduit as envisioned by the CDA. They took postings made by their customers (presumably ones of the hotter members) and apparently copied them into paid ads placed elsewhere. That's commercial exploitation.
Incidentally, there's no EULA issue because the plaintiff isn't a Friendfinder customer. Someone else put up a bogus ad describing her.
In Yahoo's case, it has particular problems because Yahoo has substantial unconsolidated holdings in other companies such as Yahoo Japan and Alibaba. The value of these companies shows up in the P part of the ratio, but the earnings aren't counted in the E part. The value of these holdings alone would put the value of the company close to the $10 billion number you propose.
Yahoo's earnings statement includes their 44% share of Alibaba. See Yahoo's 10-K filing for 2006, "Notes to Consolidated Financial Statements", Note 4, "Investments in Equity Interests": "The Company records its share of the results of Alibaba and any related amortization expense, one quarter in arrears, within earnings in equity interests on the consolidated statements of income." Yahoo's 34% stake in Yahoo Japan is accounted for similarly.
Total earnings from equity interests for 2006 were $112 million. That's about 15% of Yahoo's net income, and it is included on the consolidated balance sheet.
If we believe Yahoo's forecasts, their stock price has a fair value closer to $40/share, but even coming up short of this doesn't make them very overpriced. They are in a rapidly growing industry and have had double digit revenue growth for many years, so I think they still qualify as a growth company.
Ignore the wishful "forecasts" and look at the actuals. Yahoo's revenue peaked in 2005, and it's been downhill since then. They're not a growth company any more, they're a declining company.
Yahoo's stock is way overpriced. They're a large, mature company, not a growth company. Revenue is down. So they should have a P/E ration in the 10-20 range, like IBM, Microsoft, Oracle, and SAP.
But YHOO has a P/E ratio of 59 today. Which is far, far too high. Their market cap is around $37 billion. Divide that by 4 and you're close to what the company is really worth. Maybe $10 billion.
This is why Microsoft's institutional shareholders are unhappy with the proposed deal. Microsoft is overpaying, and that makes Microsoft less valuable.
Of course, if Microsoft just drops the deal, the bottom falls out of Yahoo stock, and it probably goes down to something closer to what it is really worth.
Google is overpriced too, but not as badly. Their P/E is around $36, while their revenue is flat or declining slightly. The fundamental problem with Google is that all those free services they give away don't make them any money. They've never found a second big moneymaking product.
What? I've never seen an ad in my gmail when i use my phone.
In the early days of Gmail, you had to supply a cell phone number, and your initial password was sent to your cell phone via SMS. One Gmail account per cell phone number. This puts a dent in spamming; you have to keep buying new phone numbers as your old accounts are terminated.
Some free dating sites now do this. I've been bugging the Craigslist people to try it.
If they actually start making the hard parts, they'll run into some of the
tough problems of machine shop work.
A classic problem: how to make a leadscrew more precise than the one you've got.
If you just make a new one using an existing one for positioning, each generation
of copy will be worse than the previous one.
Maudsley solved this problem between 1800 and 1810, with his "screw-originating machine".
This makes a very accurate screw, slowly and in a soft metal. This screw is then used in a thread-cutting lathe to make second generation leadscrews that aren't quite as good. So most screws are "descended" from a reference screw, and are only a few generations removed from it.
There are other ways to approach the problem today, typically using some form of position
feedback separate from the drive mechanism.
For a self-replicating machine that doesn't get precision parts from an external source, position measurement via absolute means, like interferometery, as in a ruling engine, might be necessary.
Right now, the RepRap people are punting on the generation loss problem, because they're only making the easy parts. If they're serious, they'll need to solve it.
Here's the
real site. Look at the picture. The machine can make the white plastic parts. Not the motors, not the leadscrews, not the frame rods, not the belts, not the wiring, and not the control electronics. The parts it is making look like about $10 worth of injection molded plastic - the cheap parts.
Gmail should go back to their old scheme, where you had to have a cell phone to receive your password, and you could only have one gmail account per phone. That would slow the spammers down.
If you don't have a phone, you're probably not a good candidate for an advertiser-supported service anyway.
At the last CES, the general reaction was that there was nothing new and exciting. The combination MP3 player/Taser got some press, but that was about it.
This list looks similar. There's nothing really exciting there. A few new cell phone variants, Sony's answer to the Mac Air, and a piggy bank. (Even the piggy bank isn't original; I've seen cheezy things like that sold for a few dollars in drugstores.)
We're not seeing a laptop that can run for a week on its fuel cell, or a Roomba-like cleaning robot smart enough to deal with corners, or even something as original as the Wii.
Slashdot Mirror
Microsoft is right. Most applications should never have administrator privileges, not even during installation. It's way past time to tighten the screws.
NVidia may just put a CPU or two in their graphics chips. They already have more transistors in the GPU than Intel has in many of their CPUs. They could license a CPU design from AMD. A CPU design today is a file of Verilog, so integrating that onto a part with a GPU isn't that big a deal.
Some better examples:
We need the FBI Baltimore office taken out of the business of distributing child porn and put on this problem. After ten years of work, they've arrested over 6,000 people.
How many computer criminals have they arrested? The Department of Justice doesn't seem to provide useful statistics, but it looks like the number per year is in the 10-100 range.
This is backwards, given the relative size of the problems.
Part of the problem is that the FBI has a measurement bias against white-collar crime. See the FBI Crime Statistics page. Violent crimes are counted if they are reported; white collar crimes are only counted if there's an arrest.
The music industry doesn't really matter, but the decline of newspaper reporting is important to society. There are too few reporters out there digging. Bloggers don't help; they're mostly mouthing off, not out gathering unpublished information.
Call your local newspaper and ask them how many full time reporters they have on the street. If they have ten, you're very lucky. There are very few newspapers left with big reporting staffs, and it shows.
News is what someone doesn't want published. All else is publicity.
This matters. Who's going to the city council meeting and reporting when they do something stupid? Who's tracking the state legislature? Who didn't report that FEMA's director was out to lunch before Katrina? What's not being reported now that's going to lead to big trouble in a year or two?
Combustion instability is an old problem with rocket engines. The Saturn V main engine had serious combustion instability problems, which were fixed by trial and error testing. The Apollo booster people had to resort to setting off small bombs inside engines on test stands to induce instability, then trying different patterns of holes in the plates the distributed fuel to find a stable configuration.
The SR-71 engine had serious combustion instability. That, too, was fixed with something of a hack, an automated "sympathetic unstart" system which, when one engine had a stall, would stall the other one, then restart both.
Better simulation tools in that area can't hurt. Not many big supersonic engines are designed any more. As Scott Crossfield pointed out just before he died a few years ago, every aircraft that went significantly over Mach 3 is now in a museum.
The industry is going at this all wrong. There are only a few players left, and they're all crooks. We need a consortium of companies with spam problems to hire Kroll, Blackwater, or one of the other big international security companies to deal with the people behind the problem.
If 5% of the money spent on dealing with spam went into finding the people behind it and making them go away, the problem would go away.
The poll defines "top academic scientist" as a reader of Nature.
Good point. Nature suffers when it goes outside its field. Their articles on biology tend to be excellent, but when they cover computing, the articles are weak, and Nature articles on economics are just laughable.
Most of the "ubiquitous computing" ideas are silly. There's all this information collection, but the systems don't have the actuators or smarts to do much with the information except bother some human.
Something you can buy right now, yet few buildings have, is really good HVAC control. You can get air sensors that sense temperature, humidity, CO, CO2, and particulates. You can get heating units, fans, dampers, and chillers that will talk to a network. You can get control systems that can manage all this to provide an optimal indoor environment as occupants come and go. A system like this will lower HVAC costs. Yet such systems are rare.
We still don't have good cleaning robots. The iRobot Scooba is about as good as it gets, but it's very dumb, frequently gets stuck, and can't refill, clean, or recharge itself.
Most of the "kitchen automation" stuff is just inventory control, not automated cooking.
The "ubiquitous computing" people haven't even been able to deliver a good meeting room automation system, one that gets lights, audio, and projector to play well together.
Educational games come in two basic flavors. In the first, the "education" consists of puzzles which have to be solved to advance gameplay. These are basically drill-and-practice programs with gameplay wrapped around them. If it makes dull drill and practice tolerable, why not? Most such games are trivial. There were many such games in the DOS era; today they're in Flash. Try Type Type Revolution, which is exactly what you think it is.
The other class of educational games are simulators, designed to teach some skill. The curriculum has to be adapted to the game, rather than the other way round, and teachers hate that. Simulators are great for teaching complex skills which require coordinating multiple goals, like landing an aircraft or running a business. They're not useful for preparing students to take written tests. So the military likes them; most educators don't.
It looks like machines in that range are going to catch on. It's a good time to be looking at major open source projects and de-bloating them. Things like browser pre-fetch and page image cacheing, which never did that much for performance but bloated memory use, should go. Make sure that installs don't leave unused junk behind.
Here's the HP HP security notice. This was discovered in January/February, according to HP, but not announced by them until April.
Where's the recall notice? HP should be recalling these items. Failure to do so immediately is willful negligence.
Here are the part numbers:
They're still for sale on Amazon, for example.
In a situation like this, HP should recall the product and reissue a replacement product with a new part number to distinguish old product from new product.
More Roland the Plogger blogspam, driving traffic to his useless ad-laden blog. To get around the block on links to his own site, he's now submitting links disguised via "tinyurl".
Slashdot covered this laser weeks ago.
I took a look through the site, and got to the "Enter Credit Card Number" point without seeing the name and address of the business.
That's a criminal offense if selling into California: Before accepting any payment or processing any debit or credit charge or funds transfer, the vendor shall disclose to the buyer in writing or by electronic means of communication, such as e-mail or an on-screen notice, the vendor's return and refund policy, the legal name under which the business is conducted and, except as provided in paragraph (3), the complete street address from which the business is actually conducted. ...
(g) Any violation of the provisions of this section is a misdemeanor punishable by imprisonment in the county jail not exceeding six months, by a fine not exceeding one thousand dollars ($1,000), or by both that imprisonment and fine.
Let's see what we can find out.
WHOIS gives us:
Buzzelli, David
Memsen
3604 SE Powell Valley Road
#267
Gresham, Oregon 97080
United States
(503) 667-3136
That's a start. More info is available if you dig.
Consider the open source alternative, OpenMoko No worries about some sudden "change in corporate direction" screwing you over.
The internet is a revolutionary zone, where something completely new could come out at any moment.
It's 2008, not 1998. There are no more dot-com parties.
Here's my driveway in Google StreetView. Note the license plate. Coverage of this house is very good. The highest resolution images are available in the aerials and in the street views. Plus it's a corner lot, and there's full coverage from both streets.
This is from one of Google's earliest batches of images. I'm only a few miles from Google HQ, and they started by thoroughly covering the nearby areas.
Big Google is watching You.
Mod parent up.
First, this isn't even a final court decision; it's just some preliminary motion practice, during which the claims are narrowed.
Second, Friendfinder went way beyond being a passive conduit as envisioned by the CDA. They took postings made by their customers (presumably ones of the hotter members) and apparently copied them into paid ads placed elsewhere. That's commercial exploitation.
Incidentally, there's no EULA issue because the plaintiff isn't a Friendfinder customer. Someone else put up a bogus ad describing her.
In Yahoo's case, it has particular problems because Yahoo has substantial unconsolidated holdings in other companies such as Yahoo Japan and Alibaba. The value of these companies shows up in the P part of the ratio, but the earnings aren't counted in the E part. The value of these holdings alone would put the value of the company close to the $10 billion number you propose.
Yahoo's earnings statement includes their 44% share of Alibaba. See Yahoo's 10-K filing for 2006, "Notes to Consolidated Financial Statements", Note 4, "Investments in Equity Interests": "The Company records its share of the results of Alibaba and any related amortization expense, one quarter in arrears, within earnings in equity interests on the consolidated statements of income." Yahoo's 34% stake in Yahoo Japan is accounted for similarly. Total earnings from equity interests for 2006 were $112 million. That's about 15% of Yahoo's net income, and it is included on the consolidated balance sheet.
If we believe Yahoo's forecasts, their stock price has a fair value closer to $40/share, but even coming up short of this doesn't make them very overpriced. They are in a rapidly growing industry and have had double digit revenue growth for many years, so I think they still qualify as a growth company.
Ignore the wishful "forecasts" and look at the actuals. Yahoo's revenue peaked in 2005, and it's been downhill since then. They're not a growth company any more, they're a declining company.
Read the SEC filings, not the press releases.
Yahoo's stock is way overpriced. They're a large, mature company, not a growth company. Revenue is down. So they should have a P/E ration in the 10-20 range, like IBM, Microsoft, Oracle, and SAP.
But YHOO has a P/E ratio of 59 today. Which is far, far too high. Their market cap is around $37 billion. Divide that by 4 and you're close to what the company is really worth. Maybe $10 billion.
This is why Microsoft's institutional shareholders are unhappy with the proposed deal. Microsoft is overpaying, and that makes Microsoft less valuable.
Of course, if Microsoft just drops the deal, the bottom falls out of Yahoo stock, and it probably goes down to something closer to what it is really worth.
Google is overpriced too, but not as badly. Their P/E is around $36, while their revenue is flat or declining slightly. The fundamental problem with Google is that all those free services they give away don't make them any money. They've never found a second big moneymaking product.
What? I've never seen an ad in my gmail when i use my phone.
In the early days of Gmail, you had to supply a cell phone number, and your initial password was sent to your cell phone via SMS. One Gmail account per cell phone number. This puts a dent in spamming; you have to keep buying new phone numbers as your old accounts are terminated.
Some free dating sites now do this. I've been bugging the Craigslist people to try it.
If they actually start making the hard parts, they'll run into some of the tough problems of machine shop work. A classic problem: how to make a leadscrew more precise than the one you've got. If you just make a new one using an existing one for positioning, each generation of copy will be worse than the previous one.
Maudsley solved this problem between 1800 and 1810, with his "screw-originating machine". This makes a very accurate screw, slowly and in a soft metal. This screw is then used in a thread-cutting lathe to make second generation leadscrews that aren't quite as good. So most screws are "descended" from a reference screw, and are only a few generations removed from it.
There are other ways to approach the problem today, typically using some form of position feedback separate from the drive mechanism. For a self-replicating machine that doesn't get precision parts from an external source, position measurement via absolute means, like interferometery, as in a ruling engine, might be necessary.
Right now, the RepRap people are punting on the generation loss problem, because they're only making the easy parts. If they're serious, they'll need to solve it.
Here's the real site. Look at the picture. The machine can make the white plastic parts. Not the motors, not the leadscrews, not the frame rods, not the belts, not the wiring, and not the control electronics. The parts it is making look like about $10 worth of injection molded plastic - the cheap parts.
Gmail should go back to their old scheme, where you had to have a cell phone to receive your password, and you could only have one gmail account per phone. That would slow the spammers down.
If you don't have a phone, you're probably not a good candidate for an advertiser-supported service anyway.
At the last CES, the general reaction was that there was nothing new and exciting. The combination MP3 player/Taser got some press, but that was about it.
This list looks similar. There's nothing really exciting there. A few new cell phone variants, Sony's answer to the Mac Air, and a piggy bank. (Even the piggy bank isn't original; I've seen cheezy things like that sold for a few dollars in drugstores.)
We're not seeing a laptop that can run for a week on its fuel cell, or a Roomba-like cleaning robot smart enough to deal with corners, or even something as original as the Wii.