If you really must use Microsoft Access (or any other non-DirectX product), you could always do so using VMWare Express. While it's slow, it still gives me the "advantages" of Windows (using Access for example), without the disadvantages (rebooting after MS Access crashes).
On my AMD K6-II 500MHz box, the VMware machine runs at speed that I'd expect from a 200MHz Pentium class. Everything else apart from Windows-only stuff, I run native under Linux (licq, Netscape, and friends).
VMWare does not support DirectX, so you can't run most Microsoft Windows games - or any other product that uses DirectX such as Wizards of The Coast's "Dungeons and Dragons Character Generator".
I would suggest that the main reason that Erin Brokovitch got made into a movie is that it's the only time that the legal system has worked for the poor against the rich.
Even then, Erin had to work the system and pull all kinds of stunts - social engineering included - to get the outcome she did. The succeess of the case was more about having the nouse to manipulate people to obtain information, rather than having money to pay lawyers to hide information.
Erin was actually quite rich in terms of motivation, energy, people skills and time.
Why go to the expense of redeveloping the banner ad system, before you even bother giving PayPal a chance? Give the readers a chance to trial different sponsorship/subscription plans, before you write off those methods as "unworthy".
Most people I know are aware of how much it costs (esp. in terms of maintenance work) to keep a web site online. We'd be willing to contribute some amount on an irregular basis. Heck - maybe we'd even be willing to "tip" for good stories. Where's a Jon Katz story when I've got a few spare dollars?;)
Maybe then the editors would realise that I like certain types of article, and hate others (because when they write the ones I like, they get more money).
Slashdot Taking On [H]ard|OCP?
on
Cool Case
·
· Score: 1
For more stuff dealing with overclocking, cool case mods and other assorted soldering-iron-and-rivet-gun fetish stories, you really want to be over at [H]ard|OCP, not here on Slashdork.
Bounce big asteroids off the Earth's magnetic field.
The hard part will be finding electromagnets big enough to do the job, that we can afford to throw away.
Basically: get a nickel/iron asteroid, throw it towards the Earth. As the asteroid-lump approaches, set up a magnetic field around it, making it an asteroid-magnet. The magnetic field should be designed to repel the Earth-magnet.
Then you have to straighten out the orbit once we've reached the desired distance (make the now-elliptical orbit circular again).
I'd do a couple of trial runs on places like Ceres or Mars first...
I use my Palm as a glorified alarm clock. It helps me remember things that I would never have remembered otherwise.
For example, I might set an alarm today that's going to beep at me on Friday night at 17:30, to remind me that I've got a dinner/movie date.
By Friday afternoon, I'll have forgotten the dinner date - it's of no use to remember that I'm seeing "Castaway" with Sarah at Hoyts, when I'm in the middle of debugging a Java application with 100 Classes.
However, as soon as the alarm goes off, I don't even need to look at the Palm - I have already remembered where I'm supposed to be and when.
My PDA has not made me more forgetful - it's served quite well as a trigger for my memory.
Extraction costs could come down a lot - you don't have to be picky about disposal of waste. What you don't want, you leave right there. There's a massive savings right off the bat - let's have a stab in the dark and say about 50%.
Distance is not a great factor. A small delta-V means you get the materials into Earth orbit in a couple of years. Heck, equip the refinery with an ion drive and refine your reaction mass from the asteroid. Constant small force over long time = very large delta-V.
Yes, to make economic gains from asteroid mining, you can't be in the "Get Rich Quick" category. You have to be patient, and willing to pay large up front costs (research, development and launch). The payoff will be cheaper in-orbit materials, cleaner to-surface materials, and much less pollution on the Earth.
The longer you're willing to wait for processed materials to arrive in Earth orbit, the cheaper it becomes to ship them from the NEO, or even "Asteroid Belt" orbit.
The most expensive part of asteroid mining will be developing the robotic asteroid mine. Compared to research and development of this device, launch costs will be negligible. If we can power these things off "dirty" ion-drive fuel (say, iron, nickel or silicon) then we don't have to worry about refueling them.
There are huge economic gains to be made from asteroid mining. The hardest part is the first step.
My company of 30 people has been happily chugging along using Netscape Calendar. This isn't supported anymore, the original product is CS&T CorporateTime.
CS&T have merged (bought? been bought by?) with another company "Lexacom", and they've changed their name to Steltor. Cute little chameleon logo, just like a certain Linux distribution.
Steltor's CorporateTime product provides a whole heap of features, the ones I use most frequently are:
"Resources" representing rooms, cars, computers - these can be booked into meetings, then a "designate" selects which of any conflicting meetings that resource will attend
Off-line operation, similar to Lotus Notes IIRC. You download your agenda to an off-line store, work with it, then next time you connect to the server, your off-line and server-based calendars will be synchronised
"In-Tray" where you can see all the new tasks and appointments that have been added for you
Clients for i386 linux, Windows 9x & NT 4.0, Macintosh
Macintosh and Windows based conduits for your PalmOS or WinCE based palm-top
Some of the features that CorporateTime didn't have in version 4, that I'd have liked were:
No way of saying "I won't be here for this week" - the only option you have is to book your self into a meeting called "On Leave", which starts at 00:00 and goes to 23:59, repeating each day for the period of your leave
No way of saying "Don't EVER accept bookings for meetings before 09:30", the only way around this is to book yourself into a recurring meeting from 00:00 to 09:30 for "Don't Bug Me", location "In Bed"
No way of tying in your schedule with a project plan (done in say Microsoft Project or MicroPlanner X-Pert).
CorporateTime does achieve the goal of managing resources and booking appointments. It's the bits that it doesn't manage that tend to chafe me a little - but most of these are human-to-human communications issues like "don't book me for a meeting without at least 24 hours notice".
Wearable computers are also useful for such things as forward observers (linking the GPS and the laser range finder), more reliable comms links (digital data is more resilient to noise than ananlogue voice) and overlaying tactical information on the soldier's FOV.
Overlays could include waypoint diamonds for nearest friendlies and spotted enemies. Don't you think that would help the poor Grunt and her squad mates stay alive for just a few more minutes?
The same sort of thing happened to us at work. We have a Linux network, with a Solaris box for development under Solaris, a couple of Windows 98 boxes for secretaries (who live on Microsoft Word and MYOB) and assorted routers and appliances.
Then some fool decided he'd install Windows NT, on a box he brought in from home.
He figured there was no Windows NT server already on the network, so his box became Primary Domain Controller (that's a choice the user has to make during install, implying that Microsoft assumes only MCSEs have access to Windows NT install CDs).
For three days we're trying to figure out WTF was wrong with our network, until we figured out that this box "Pyramid" was a machine that we hadn't seen before. We disconnected that office from the network, and everything came back to normal.
The same thing could have happened if it was a Windows network, and the guy had installed Samba on his personal Unix/BSD/Linux box.
Moral of the story - Don't mess with/on other people's networks unless they've okayed it.
Petrol has already reached the level of AUD 1/L - which I think works out to about USD 2.30 per Gallon
There's lots of whingeing and complaining, but nothing's going to happen - petrol excise is the Federal Government's biggest tax income, and petrol taxes are the State Governments' biggest incomes. The petrol stations are already posting signs along the lines of "$0.40/L cost + $0.60/L tax = $1.00/L"
Right now, it'd probably be really nice to own a Honda Insight. Pity Australia is currently too small a market for the Insight (Australia has a population of about 17 million, vs. New York's population of 25 million).
Windows Millenium, n. : 64 months of work to clean up the 32 bit graphical shell for the 16 bit extensions to an 8 bit operating system originally coded for a 4 bit processor, by a 2 bit company that doesn't have one bit of respect for its customers, and will stand for not a bit of competition.
Go have a look at the SETI@Home Downloads page. Tell me there are no processors in there that might have vector-optimised maths units.
Go have a look at the Crusoe Technology page, or the VMWare site. Tell me that it's impossible to use hardware or software to emulate or translate from one instruction set to another. Besides, what's stopping the KrosnoConv "surplus military" stock being military-spec MIPS or SPARC clones? I didn't read the "about the company" bit, so I expected this was just an American company picking up after the excesses of the United States war machine.
Heck, check any Pentium III and tell me that it's impossible to execute another device's instruction set (8088) natively.
The only points that concerned me about the KrosnoConv boards were the Linux-in-Flash claim, and 32Mb of RAM per processor for less than $US200 (either very slow memory, or only 32Mbit perhaps, co-packaged RAM from the old 8086 days). There are projects out there to put Linux in a PC BIOS, or even an LS120. You can get CompactFlash cards, which behave like very small hard drives (either Flash memory pretending to be an IDE drives, or IBM microdrives really being IDE drives). But they're not cheap.
x86 is not the only architecture that SETI@Home supports. Why shouldn't someone produce an add-in card that uses your existing infrastructure? I would still be interested in getting a (cheap!) board full of heavy-maths processors to do hardware encryption for Virtual Private Networks, or even just a heavy duty key server.
I've already got the expensive bits like hard drives, network cards, monitors, cases and memory. I'd actually love to have a "parasitic" processor running its own OS, where I can download software to it, and have it process data that I store for it in my real RAM. Kinda like a multi-processor machine, where one or more processors are especially suited to encryption math.
Don't just spout "that's not how things work". Because with the introduction of technology like that used in the Crusoe, or even older technology like that used in VMWare, or any C64 emulator, you know this is how things work now.
I use an AMD K6-2 500MHz chip on a EPoX EP-MVP3G5 M/B - this uses the VIA Apollo 4 chipset.
It's only the second out of 5 motherboards I've ever had which run Windows 98 without crashing (the last was an Intel Triton III chipset with a Pentium 133).
I have 256Mb of RAM in 2*128Mb DIMMS - one PC100 the other PC133. Even with this mix of DIMMs, things work perfectly.
This system runs Linux nicely (always too slow of course;), and VMWare with Win98 installed as Guest OS is adequate to my needs (running Excel to edit timesheets, CorporateSync for Palm to synchronise my PalmOS device with CorporateTime calendar).
I'm just wondering if maybe you shouldn't just upgrade your M/B, rather than spending all that extra mulah on a new machine?
Use the first as the key, the remainder as a historical archive of the state of our planet.
Future civilisations could learn from our mistakes. Then Archaeology will finally become recognised as the amazing science that it is. Pity the Ancient Romans only discovered lead poisoning, and not the Greenhouse Effect or Faster Than Light travel.
Failing all else, they can figure out why our social models failed, and rework their society into something more stable, that will actually last out the next 10,000 years.
For starters, a properly designed manual (regardless of whether it's printed, in PDF or HTML or whatever) reduces the need for word-based searching. The theory being that the people writing the documentation break it up into "atomic" pieces, and give each piece a useful title. Then you have an index and glossary which cover all the "jargon" words for your product, along with all the useful or moderately useful terms.
In the event that you want to provide a search capability on your on-line documentation, there are self-contained tools available such as Verity's (closed-source, proprietary) Search 97 CD Web Publisher, which is a self-contained web server and search engine. There is Perlfect Search, which is not stand-alone, but could probably be incorporated into a self-contained documentation server.
Failing all else, you give the customer permission to copy the documentation to their own Intranet site, which will usually have some kind of indexing/searching system available.
IMHO, PDF is the worst thing to ever happen to the Internet/World Wide Web. Here is a format that:
can't be re-rendered to fit the screen I'm using,
requires a specific paper size when printing (ie: US documentation usually wants US letter, while British will usually want A4),
dictates that certain fonts are used,
looks ugly on screen, and only looks good on paper, because that's what the document was designed for, and
wastes bandwidth (Internet or Intranet) because you have to download the whole thing to view one page.
One great advantage of XML/HTML documentation is that it can be read using Lynx (after being served up by your web server), or converted to doc format for Newton or Palm devices (or in fact, any format you want).
Wouldn't it be funny if the International Space Station is abandoned because the Russians are cheap stingy (and noisy) bastards, while the USA and China build a cooperative Moon base and start mining?
China supplies the cheap labour, and derives hard Western currency (as well as huge Kudos). The USA gets a well maintained waystation for cooperative attempts to colonise Mars and explore the asteroid belt.
Can't wait until the first electrodynamic tether is used to sling an Earth-orbiting payload into lunar space:) Or what about when the University of Mars becomes a *real* IP address, and we have to re-write a bunch of RFCs?
The cost of interplanetary travel drops tremendously when you don't have to take all of the fuel for the trip off the planet with you.
So you set up a mining operation on the moon. Mine Aluminium and Oxygen. Voila - cheap (if nasty) propellents.
From the Moon, launch your propellents (using any number of exotic techniques from linear accelerators through chemical rockets through orbiting catapults, slingshots or tethers) into low Earth orbit.
Then all the fuel you need to start an interplanetary mission is what it takes to get off the ground and into Low Earth Orbit.
Of course, the catch with a Low-Earth-Orbit tether is getting the payload onto the tether, which will be moving at several hundred km/h. On the moon, the tether could touch the ground at one point, and be rotated at the appropriate speed such that the end-point of the techter isn't moving in relation to the surface. Doing this with a much longer tether in Low Earth Orbit may be possible, but I bags not being the first human to test that ferris-wheel!
The cost of interplanetary travel drops tremendously when you don't have to take all of the fuel for the trip off the planet with you.
So you set up a mining operation on the moon. Mine Aluminium and Oxygen. Voila - cheap (if nasty) propellents.
From the Moon, launch your propellents (using any number of exotic techniques from linear accelerators through chemical rockets through orbiting catapults/slingshots/tet hers) into low Earth orbit.
Then all the fuel you need to start an interplanetary mission is what it takes to get off the ground and into Low Earth Orbit.
Of course, the catch with a Low-Earth-Orbit tether is getting the payload onto the tether, which will be moving at several hundred km/h. On the moon, the tether could touch the ground at one point, and be rotated at the appropriate speed such that the end-point of the techter isn't moving in relation to the surface. Doing this with a much longer tether in Low Earth Orbit may be possible, but I bags not being the first human to test that ferris-wheel!
The decision for Quote.Com to change wasn't only based on the "reliability" of the platform.
Their decision was also based on facts like:
There are more pre-built software components for IIS/SQL server
Things like XML support are very primitive in PERL, for example.
MCSEs are cheaper to hire than Unix admin/programmers
With more, cheaper machines, you can play the "uptime numbers game"
A lot of developers are working on XML support in PERL (there is a Perl/XML FAQ), but you still can't support Unicode. Perl still relies on 8-bit character sets, so we use UTF-8 instead of 16-bit Unicode. Unicode support is neccesary for a complete XML implementation.
You'll also find that MCSEs will be cheaper to hire than Unix programmers. This is partly due to their (general) lack of skills, and partly due to their great abundance. An MSCE course only teaches you how to think the Microsoft Way. I wouldn't trust an MSCE to maintain or write code in C++ or Perl, for example. Without the MFC and a pointy-clicky interface, an MSCE can't function.
However, give the MSCE the MFC and a pointy-clicky interface, and an MSCE can deliver a program faster than a "traditional" developer. The fact that the program inherits all the bugs and mis-features of the MFC is not an issue here. The fact that the program was slapped together without regard for maintenance or robustness is also not an issue here. The issue that Quote.Com chose to focus on was delivery time, not quality of product.
As for the uptime numbers game, it works like this:
If you have 1 Sun server, and you need to upgrade the hardware, you need to shut it down. If it takes 1 hour to shut down, replace the hardware and restart, then you have 1 hour downtime.
If you have 2 Windows NT servers (for the same price as 1 Sun machine), and you need to upgrade the hardware, you need to shut them down. If you do it one machine at a time, and take 4 hours total to replace the hardware, then the server pool still has 0 hours downtime. Windows NT pundits will happily overlook the fact that the individual machines are constantly being overhauled.
In addition, Microsoft introduces the idea of "scheduled downtime". That is - you plan to reboot each machine once a day, to make sure the system remains stable. So twice a day, you have one of your two machines reboot. One machine reboots in the morning, the other in the afternoon. The total downtime of the server pool as a whole is still 0 hours (because you're not counting "scheduled downtime" as "real downtime").
Now combine the MSCE factor with the downtime numbers game factor, and you'll find that you can get away with shoddy code, because when your server crashes, it's not really downtime anymore. The problem of data integrity in your backend database is something for the DBA to worry about. You've got your uptime figures and time-to-delivery figures up there in the top 10. If the DBA complains about data integrity, you sack her and fire someone with a more "can-do" attitude. You don't want slackers in your Microsoft Powered enterprise!
Not quite right. In a population of 1000, you will have approximately 500 of them die before they're nine years old, approximately 500 of them die sometime after nine years, and the rest dying dead on the nine year mark.
So in a population of 154,488 machines, you'll have one dying every hour for the next 18 years (for the original population of 77,244 that works out to one dying every two hours for the next 18 years).
If you believe in statistics, you'll find the figure is more like 80,000 of them dying in the period between 7 years and 11 years, if we're looking at a "bell" curve.
Slashdot Mirror
The catch is that the software that they are selling will take business away from the rest of their company.
Cold Fusion is a Windows product - most of O'Rielly's books are about Unix based stuff.
Check this post for a more elegant description of the problem.
What dynamically types remote interfaces and service descriptions did "people" have before CORBA?
How does CORBA represent a step backwards?
If you really must use Microsoft Access (or any other non-DirectX product), you could always do so using VMWare Express. While it's slow, it still gives me the "advantages" of Windows (using Access for example), without the disadvantages (rebooting after MS Access crashes).
On my AMD K6-II 500MHz box, the VMware machine runs at speed that I'd expect from a 200MHz Pentium class. Everything else apart from Windows-only stuff, I run native under Linux (licq, Netscape, and friends).
VMWare does not support DirectX, so you can't run most Microsoft Windows games - or any other product that uses DirectX such as Wizards of The Coast's "Dungeons and Dragons Character Generator".
I would suggest that the main reason that Erin Brokovitch got made into a movie is that it's the only time that the legal system has worked for the poor against the rich.
Even then, Erin had to work the system and pull all kinds of stunts - social engineering included - to get the outcome she did. The succeess of the case was more about having the nouse to manipulate people to obtain information, rather than having money to pay lawyers to hide information.
Erin was actually quite rich in terms of motivation, energy, people skills and time.
Why go to the expense of redeveloping the banner ad system, before you even bother giving PayPal a chance? Give the readers a chance to trial different sponsorship/subscription plans, before you write off those methods as "unworthy".
Most people I know are aware of how much it costs (esp. in terms of maintenance work) to keep a web site online. We'd be willing to contribute some amount on an irregular basis. Heck - maybe we'd even be willing to "tip" for good stories. Where's a Jon Katz story when I've got a few spare dollars? ;)
Maybe then the editors would realise that I like certain types of article, and hate others (because when they write the ones I like, they get more money).
For more stuff dealing with overclocking, cool case mods and other assorted soldering-iron-and-rivet-gun fetish stories, you really want to be over at [H]ard|OCP, not here on Slashdork.
Bounce big asteroids off the Earth's magnetic field.
The hard part will be finding electromagnets big enough to do the job, that we can afford to throw away.
Basically: get a nickel/iron asteroid, throw it towards the Earth. As the asteroid-lump approaches, set up a magnetic field around it, making it an asteroid-magnet. The magnetic field should be designed to repel the Earth-magnet.
Then you have to straighten out the orbit once we've reached the desired distance (make the now-elliptical orbit circular again).
I'd do a couple of trial runs on places like Ceres or Mars first...
I use my Palm as a glorified alarm clock. It helps me remember things that I would never have remembered otherwise.
For example, I might set an alarm today that's going to beep at me on Friday night at 17:30, to remind me that I've got a dinner/movie date.
By Friday afternoon, I'll have forgotten the dinner date - it's of no use to remember that I'm seeing "Castaway" with Sarah at Hoyts, when I'm in the middle of debugging a Java application with 100 Classes.
However, as soon as the alarm goes off, I don't even need to look at the Palm - I have already remembered where I'm supposed to be and when.
My PDA has not made me more forgetful - it's served quite well as a trigger for my memory.
Extraction costs could come down a lot - you don't have to be picky about disposal of waste. What you don't want, you leave right there. There's a massive savings right off the bat - let's have a stab in the dark and say about 50%.
Distance is not a great factor. A small delta-V means you get the materials into Earth orbit in a couple of years. Heck, equip the refinery with an ion drive and refine your reaction mass from the asteroid. Constant small force over long time = very large delta-V.
Yes, to make economic gains from asteroid mining, you can't be in the "Get Rich Quick" category. You have to be patient, and willing to pay large up front costs (research, development and launch). The payoff will be cheaper in-orbit materials, cleaner to-surface materials, and much less pollution on the Earth.
The longer you're willing to wait for processed materials to arrive in Earth orbit, the cheaper it becomes to ship them from the NEO, or even "Asteroid Belt" orbit.
The most expensive part of asteroid mining will be developing the robotic asteroid mine. Compared to research and development of this device, launch costs will be negligible. If we can power these things off "dirty" ion-drive fuel (say, iron, nickel or silicon) then we don't have to worry about refueling them.
There are huge economic gains to be made from asteroid mining. The hardest part is the first step.
My company of 30 people has been happily chugging along using Netscape Calendar. This isn't supported anymore, the original product is CS&T CorporateTime.
CS&T have merged (bought? been bought by?) with another company "Lexacom", and they've changed their name to Steltor. Cute little chameleon logo, just like a certain Linux distribution.
Steltor's CorporateTime product provides a whole heap of features, the ones I use most frequently are:
Some of the features that CorporateTime didn't have in version 4, that I'd have liked were:
CorporateTime does achieve the goal of managing resources and booking appointments. It's the bits that it doesn't manage that tend to chafe me a little - but most of these are human-to-human communications issues like "don't book me for a meeting without at least 24 hours notice".
Wearable computers are also useful for such things as forward observers (linking the GPS and the laser range finder), more reliable comms links (digital data is more resilient to noise than ananlogue voice) and overlaying tactical information on the soldier's FOV.
Overlays could include waypoint diamonds for nearest friendlies and spotted enemies. Don't you think that would help the poor Grunt and her squad mates stay alive for just a few more minutes?
The same sort of thing happened to us at work. We have a Linux network, with a Solaris box for development under Solaris, a couple of Windows 98 boxes for secretaries (who live on Microsoft Word and MYOB) and assorted routers and appliances.
Then some fool decided he'd install Windows NT, on a box he brought in from home.
He figured there was no Windows NT server already on the network, so his box became Primary Domain Controller (that's a choice the user has to make during install, implying that Microsoft assumes only MCSEs have access to Windows NT install CDs).
For three days we're trying to figure out WTF was wrong with our network, until we figured out that this box "Pyramid" was a machine that we hadn't seen before. We disconnected that office from the network, and everything came back to normal.
The same thing could have happened if it was a Windows network, and the guy had installed Samba on his personal Unix/BSD/Linux box.
Moral of the story - Don't mess with/on other people's networks unless they've okayed it.
We're doing this already in Australia.
Petrol has already reached the level of AUD 1/L - which I think works out to about USD 2.30 per Gallon
There's lots of whingeing and complaining, but nothing's going to happen - petrol excise is the Federal Government's biggest tax income, and petrol taxes are the State Governments' biggest incomes. The petrol stations are already posting signs along the lines of "$0.40/L cost + $0.60/L tax = $1.00/L"
Right now, it'd probably be really nice to own a Honda Insight. Pity Australia is currently too small a market for the Insight (Australia has a population of about 17 million, vs. New York's population of 25 million).
Windows Millenium, n. : 64 months of work to clean up the 32 bit graphical shell for the 16 bit extensions to an 8 bit operating system originally coded for a 4 bit processor, by a 2 bit company that doesn't have one bit of respect for its customers, and will stand for not a bit of competition.
Go have a look at the SETI@Home Downloads page. Tell me there are no processors in there that might have vector-optimised maths units.
Go have a look at the Crusoe Technology page, or the VMWare site. Tell me that it's impossible to use hardware or software to emulate or translate from one instruction set to another. Besides, what's stopping the KrosnoConv "surplus military" stock being military-spec MIPS or SPARC clones? I didn't read the "about the company" bit, so I expected this was just an American company picking up after the excesses of the United States war machine.
Heck, check any Pentium III and tell me that it's impossible to execute another device's instruction set (8088) natively.
The only points that concerned me about the KrosnoConv boards were the Linux-in-Flash claim, and 32Mb of RAM per processor for less than $US200 (either very slow memory, or only 32Mbit perhaps, co-packaged RAM from the old 8086 days). There are projects out there to put Linux in a PC BIOS, or even an LS120. You can get CompactFlash cards, which behave like very small hard drives (either Flash memory pretending to be an IDE drives, or IBM microdrives really being IDE drives). But they're not cheap.
x86 is not the only architecture that SETI@Home supports. Why shouldn't someone produce an add-in card that uses your existing infrastructure? I would still be interested in getting a (cheap!) board full of heavy-maths processors to do hardware encryption for Virtual Private Networks, or even just a heavy duty key server.
I've already got the expensive bits like hard drives, network cards, monitors, cases and memory. I'd actually love to have a "parasitic" processor running its own OS, where I can download software to it, and have it process data that I store for it in my real RAM. Kinda like a multi-processor machine, where one or more processors are especially suited to encryption math.
Don't just spout "that's not how things work". Because with the introduction of technology like that used in the Crusoe, or even older technology like that used in VMWare, or any C64 emulator, you know this is how things work now.
SETI@Home uses Fast Fourier Transforms.
Fast Fourier Transforms on digitised samples can be done using large-integer or arbitrary-precision-number algorithms.
Therefore even Pentiums can do SETI@Home without generating errors.
I use an AMD K6-2 500MHz chip on a EPoX EP-MVP3G5 M/B - this uses the VIA Apollo 4 chipset.
It's only the second out of 5 motherboards I've ever had which run Windows 98 without crashing (the last was an Intel Triton III chipset with a Pentium 133).
I have 256Mb of RAM in 2*128Mb DIMMS - one PC100 the other PC133. Even with this mix of DIMMs, things work perfectly.
This system runs Linux nicely (always too slow of course ;), and VMWare with Win98 installed as Guest OS is adequate to my needs (running Excel to edit timesheets, CorporateSync for Palm to synchronise my PalmOS device with CorporateTime calendar).
I'm just wondering if maybe you shouldn't just upgrade your M/B, rather than spending all that extra mulah on a new machine?
We could create a series of "Rosetta disks".
Use the first as the key, the remainder as a historical archive of the state of our planet.
Future civilisations could learn from our mistakes. Then Archaeology will finally become recognised as the amazing science that it is. Pity the Ancient Romans only discovered lead poisoning, and not the Greenhouse Effect or Faster Than Light travel.
Failing all else, they can figure out why our social models failed, and rework their society into something more stable, that will actually last out the next 10,000 years.
It could simply be the case that AltaVista uses the Last-Modified-Date supplied by the web server it gets the document from.
So if Microsoft's web server reports 21 January, then AltaVista believes it. Why would it do otherwise?
Microsoft may well have prepared the document on that date, but not actually made it available on the Web until just recently.
For starters, a properly designed manual (regardless of whether it's printed, in PDF or HTML or whatever) reduces the need for word-based searching. The theory being that the people writing the documentation break it up into "atomic" pieces, and give each piece a useful title. Then you have an index and glossary which cover all the "jargon" words for your product, along with all the useful or moderately useful terms.
In the event that you want to provide a search capability on your on-line documentation, there are self-contained tools available such as Verity's (closed-source, proprietary) Search 97 CD Web Publisher, which is a self-contained web server and search engine. There is Perlfect Search, which is not stand-alone, but could probably be incorporated into a self-contained documentation server.
Failing all else, you give the customer permission to copy the documentation to their own Intranet site, which will usually have some kind of indexing/searching system available.
IMHO, PDF is the worst thing to ever happen to the Internet/World Wide Web. Here is a format that:
One great advantage of XML/HTML documentation is that it can be read using Lynx (after being served up by your web server), or converted to doc format for Newton or Palm devices (or in fact, any format you want).
Wouldn't it be funny if the International Space Station is abandoned because the Russians are cheap stingy (and noisy) bastards, while the USA and China build a cooperative Moon base and start mining?
:) Or what about when the University of Mars becomes a *real* IP address, and we have to re-write a bunch of RFCs?
China supplies the cheap labour, and derives hard Western currency (as well as huge Kudos). The USA gets a well maintained waystation for cooperative attempts to colonise Mars and explore the asteroid belt.
Can't wait until the first electrodynamic tether is used to sling an Earth-orbiting payload into lunar space
The cost of interplanetary travel drops tremendously when you don't have to take all of the fuel for the trip off the planet with you.
So you set up a mining operation on the moon. Mine Aluminium and Oxygen. Voila - cheap (if nasty) propellents.
From the Moon, launch your propellents (using any number of exotic techniques from linear accelerators through chemical rockets through orbiting catapults, slingshots or tethers) into low Earth orbit.
Then all the fuel you need to start an interplanetary mission is what it takes to get off the ground and into Low Earth Orbit.
Of course, the catch with a Low-Earth-Orbit tether is getting the payload onto the tether, which will be moving at several hundred km/h. On the moon, the tether could touch the ground at one point, and be rotated at the appropriate speed such that the end-point of the techter isn't moving in relation to the surface. Doing this with a much longer tether in Low Earth Orbit may be possible, but I bags not being the first human to test that ferris-wheel!
The cost of interplanetary travel drops tremendously when you don't have to take all of the fuel for the trip off the planet with you.
So you set up a mining operation on the moon. Mine Aluminium and Oxygen. Voila - cheap (if nasty) propellents.
From the Moon, launch your propellents (using any number of exotic techniques from linear accelerators through chemical rockets through orbiting catapults/slingshots/tet hers) into low Earth orbit.
Then all the fuel you need to start an interplanetary mission is what it takes to get off the ground and into Low Earth Orbit.
Of course, the catch with a Low-Earth-Orbit tether is getting the payload onto the tether, which will be moving at several hundred km/h. On the moon, the tether could touch the ground at one point, and be rotated at the appropriate speed such that the end-point of the techter isn't moving in relation to the surface. Doing this with a much longer tether in Low Earth Orbit may be possible, but I bags not being the first human to test that ferris-wheel!
The decision for Quote.Com to change wasn't only based on the "reliability" of the platform.
Their decision was also based on facts like:
A lot of developers are working on XML support in PERL (there is a Perl/XML FAQ), but you still can't support Unicode. Perl still relies on 8-bit character sets, so we use UTF-8 instead of 16-bit Unicode. Unicode support is neccesary for a complete XML implementation.
You'll also find that MCSEs will be cheaper to hire than Unix programmers. This is partly due to their (general) lack of skills, and partly due to their great abundance. An MSCE course only teaches you how to think the Microsoft Way. I wouldn't trust an MSCE to maintain or write code in C++ or Perl, for example. Without the MFC and a pointy-clicky interface, an MSCE can't function.
However, give the MSCE the MFC and a pointy-clicky interface, and an MSCE can deliver a program faster than a "traditional" developer. The fact that the program inherits all the bugs and mis-features of the MFC is not an issue here. The fact that the program was slapped together without regard for maintenance or robustness is also not an issue here. The issue that Quote.Com chose to focus on was delivery time, not quality of product.
As for the uptime numbers game, it works like this:
If you have 1 Sun server, and you need to upgrade the hardware, you need to shut it down. If it takes 1 hour to shut down, replace the hardware and restart, then you have 1 hour downtime.
If you have 2 Windows NT servers (for the same price as 1 Sun machine), and you need to upgrade the hardware, you need to shut them down. If you do it one machine at a time, and take 4 hours total to replace the hardware, then the server pool still has 0 hours downtime. Windows NT pundits will happily overlook the fact that the individual machines are constantly being overhauled.
In addition, Microsoft introduces the idea of "scheduled downtime". That is - you plan to reboot each machine once a day, to make sure the system remains stable. So twice a day, you have one of your two machines reboot. One machine reboots in the morning, the other in the afternoon. The total downtime of the server pool as a whole is still 0 hours (because you're not counting "scheduled downtime" as "real downtime").
Now combine the MSCE factor with the downtime numbers game factor, and you'll find that you can get away with shoddy code, because when your server crashes, it's not really downtime anymore. The problem of data integrity in your backend database is something for the DBA to worry about. You've got your uptime figures and time-to-delivery figures up there in the top 10. If the DBA complains about data integrity, you sack her and fire someone with a more "can-do" attitude. You don't want slackers in your Microsoft Powered enterprise!
Daily Reboots:
Not quite right. In a population of 1000, you will have approximately 500 of them die before they're nine years old, approximately 500 of them die sometime after nine years, and the rest dying dead on the nine year mark.
So in a population of 154,488 machines, you'll have one dying every hour for the next 18 years (for the original population of 77,244 that works out to one dying every two hours for the next 18 years).
If you believe in statistics, you'll find the figure is more like 80,000 of them dying in the period between 7 years and 11 years, if we're looking at a "bell" curve.