I was one of the original developers of the Teradata database system. Man, do I miss it. For big decision support systems, nothing else on the market comes close. It's much easier to administer, and is the ultimate in scalable.
It almost died when NCR/AT&T bought it (they had no idea what it really was), but nowadays NCR has finally recognized it's worth.
Many companies with REALLY LARGE database rely on Teradata. For example, a few years ago I read Wal*Mart had 101 TB in their database (they were always Teradata's biggest customer).
I assume this was a troll, but in case it's not, I'll reply.
First, I never saw anything that said the Xeons wouldn't have unique chip IDs, but perhaps you are right, and they have removed it.
It doesn't matter one way or the other. There was never any controversy over the chip IDs in the server market, which is the market for Xeons. Almost every server CPU has unique IDs. UltraSparcs do, IBM mainframes do, PA-RISC chips do. DataCenter managers don't see any problem with having IDs at all (they use them for asset management). The only issue was with home users and personal workstations, which is not a market for the Xeons.
The new price of these Xeons is half what the older Xeons sold for. Why is cutting their price in half "Exorbitant prices"?
In a typical server, where you might spend $25k and up, $8k for 4 Xeons doesn't seem so bad. It still gives you a system much cheaper than a Sun UltraSparc system.
Many SlashDotter seems to be either ignorant of or just ignoring major issues. For example, who controls COM/DCOM? All MS Apps rely 100% on COM, and the OS relies 100% on COM. The MS APPS group isn't going to port Office to any OS that doesn't have COM, and whatever OS best supports COM (which will be windows) will always be the preferred platform for the apps. If you give joint ownership of COM out, you have a big mess when someone changes it. You'll have the situation where Office runs on old Windows + apps group enhanced com, but apps group enhanced com doesn't work on enhanced windows from OS group.
While the resulting confusing will certainly cause numerous problems for Windows users, much to the joy of Linux advocates everywhere, I thought the DOJ was supposed to be looking out for the consumer.
Or is this really a "stick it to Bill" solution, and nobody cares if the consumers get screwed?
P-III chips have a true random number generator that uses thermal noise to generate the number. Unless you believe the paranoia and think the NSA conviced Intel to subvert their design (making it more expensive to manufacture!), thermal noise is about the best way to get random numbers for cryptography.
Correct me if I'm wrong, but Amazon was always running on Unix (solaris, I think).
They are switching from one Unix web server to another Unix web server (probably because no one trusts that Netscape Enterprise server will be around in a few years).
Well, internal to the Teradata database system, dates were orignally an integer YYMMDD, and when we need to allow for 2000, it was changed so that instead of YY it was CCYY-100, so 1999 was 99, 2000 was 100, just as you mentioned. That way, existing dates stored in the database didn't need updating.
I had my eyes done with PRK (Photoreactive Kerotectomy). Before, I was a -8.75 and -9, with about 1 diopter astigmatism. Now, I am about 0 and -1.5, so I see well enough to not need glasses for driving (20/25 in one eye, 20/35 in the other). I normally wear glasses to correct the remaining astigmatism and nearsightednesss, but can do without.
But I was also one of the unfortunates who had serious side effects from my surgery. My night vision is terrible, with huge halos around any light sources. After many years, the problem has not gone away. I've also lost a little bit of contrast in my vision.
Although they tried to correct my astigmatism, they were not very successful at that.
Overall, I am OK with the results, but it has definitely involved tradeoffs, and once it's done, there is no going back.
If I were to do it again, I would not pick PRK, but rather LASIK. LASIK involves a risky step of cutting a "flap" on the cornea so the laser can be applied to the inner layers of the cornea, rather than the surface as is done for PRK. But LASIK generally results in better vision, and the recovery time is much faster. It was, unfortunately, not available when I had my eyes done.
If you go for LASIK, understand that there are definitely risks. Also, Although many don't need glasses afterwards, some do.
If I remember right, Intel planned to go with FireWire, but the large cost (due to Apple's insistance on large royalties) jacked up the price of the motherboard chipsets enough that the motherboard manufactures told Intel "no way". Even $5 is a huge add-on cost to this market, and if I rememer, FireWire was going to add at least double that.
Competition, yes, but 5 microsofts all with the source means 5x more versions of Windows. How does this help the consumer?
Gee, nifty game x works on Company A's Windows, but the office package I want runs on Company D's windows, and the e-mail package work mandates runs on company B's windows.... Ugh.
Office already runs on the next most popular OS: MacOS. Having Office on MAC didn't cause Windows to lose out. Having Office on some variant of UNIX wouldn't either.
As one of the punishments, this isn't a bad idea. Forcing MS to open the documentation on their many proprietary file formats and APIs would help even the playing field for startup competitors.
While that would hurt Mr Gates, what would it do for the consumers? If the published source code is still MS proprietary, no one can use it in building any new projects without paying MS. If you want to make it Public Domain, or GPL, or BSD, etc., you are talking about the goverment siezing the property of a corporation (and therefor from the shareholders). do you really want the goverment going around siezing private property?
If you broke Microsoft into OS and Apps, you'd have two monopolies instead of one: Company one has a monopoly on desktop OS sales, the other has it on Office application suites. How does the consumer win?
And where would you make the divide? Is IE part of the OS or the applications? Can the OS have an e-mail client (OE) or is that part of the applications?
Which Build of W2K? I used to have this problem, but with the latest build I tried (2128), it installed in about 20 minutes, with almost no user interactions needed during that time.
Microsoft needed Softway to allow NT to acheive POSIX compatibility and UNIX95 branding, which is important for selling into the Federal marketplace.
Softway wasn't making much money on this software,and was in danger of going out of business or dropping the product, both of which were unacceptable to Microsoft.
Yes, the connection from the DSLAM to the ISP is shared, but so is the internet. In my case, The DSLAM is on PacBell's OC-45 ATM network, so the bandwidth is higher than the internet backbones usually carry.
In my case, DSL DSLAM is tied directly to PacBell's OC-45 ATM network. The ISP connects to this network and "bridges" the gap to the internet. But the OC-45 is so much faster than most internet backbones, I'm not worried about the bandwidth.
There are different kinds of DSL. G.Lite uses the same wire pair as the phones, with a filter at each phone to reduce the interference. Regular (old-style?) ADSL uses a splitter at the enterance to the residence, and then separate wires inside. This works much better, but is more costly to install.
Companies can operate at a loss for as long as they have money or liquid assests. When they run out, the company goes backrupt.
Startup companies get money from Venture Capitalists for a share of the company (or borrow money). When it runs out, they need to ask for more, or IPO (or be making a profit).
The whole point behind an IPO was traditionally to get enough money to fund a companies growth (allow them to spend more money than revenue for a while, to grow the business). IPO has instead become a way for investors to get rich quick, although it still serves the purpose of funding the company.
When the money RedHat got in the IPO, plus any reserves they had from before, run out, then RedHat will either need to be profitable, need to find a buyer, or go bankrupt.
So, the IPO money is the "savings account" they are dipping into, and it better not run out before they find some way to earn a living!
I agree with some of your post, but "it's obvious that pretty soon they're going to be making money hand over fist" doesn't fit my understanding. Most of RedHat's profits will come from services (Linux can be copied for free, right?). Growing a services business is slow and difficult (you need to hire many hundreds of linux experts, and find work for all of them).
Linux is a hot buzz word in the market, but most shareholders don't understand RedHat's business.
Worth $7.4B? Only if you believe they will growing at an enormous rate, with revenues growing 100%/quarter, and huge profits to follow.
Some shareholders are believers. Most are subscribing to the "bigger fool" theory that says there is always somebody who will buy the stock for a higher price later even if it is rediculously valued.
RedHat can't possibly live up to the hype and stock price. They are destined to fall, but when? Ahhh... If you only knew, you could make a fortune...
Ummmm... If they keep losing money, eventually they will not have any money to pay employees. When they were private, the could pay employees from money received from VCs or borrowed, and for the near term they have the money from the IPO. When that runs out, they need to be making money, or else!
Slashdot Mirror
I was one of the original developers of the Teradata database system. Man, do I miss it. For big decision support systems, nothing else on the market comes close. It's much easier to administer, and is the ultimate in scalable.
It almost died when NCR/AT&T bought it (they had no idea what it really was), but nowadays NCR has finally recognized it's worth.
Many companies with REALLY LARGE database rely on Teradata. For example, a few years ago I read Wal*Mart had 101 TB in their database (they were always Teradata's biggest customer).
3 TB databases are large, but not unusual.
Most large retailers would have databases this big.
A few years ago I read that Wal*mart had 101 TB.
Most of that is in a Teradata database.
I assume this was a troll, but in case it's not, I'll reply.
First, I never saw anything that said the Xeons wouldn't have unique chip IDs, but perhaps you are right, and they have removed it.
It doesn't matter one way or the other. There was never any controversy over the chip IDs in the server market, which is the market for Xeons.
Almost every server CPU has unique IDs. UltraSparcs do, IBM mainframes do, PA-RISC chips do. DataCenter managers don't see any problem with having IDs at all (they use them for asset management).
The only issue was with home users and personal workstations, which is not a market for the Xeons.
The new price of these Xeons is half what
the older Xeons sold for. Why is cutting their
price in half "Exorbitant prices"?
In a typical server, where you might spend
$25k and up, $8k for 4 Xeons doesn't seem
so bad. It still gives you a system much
cheaper than a Sun UltraSparc system.
Many SlashDotter seems to be either ignorant of or
just ignoring major issues.
For example, who controls COM/DCOM? All
MS Apps rely 100% on COM, and the OS relies
100% on COM. The MS APPS group isn't going to
port Office to any OS that doesn't have
COM, and whatever OS best supports COM (which
will be windows) will always be the preferred platform for the apps.
If you give joint ownership of COM out, you have
a big mess when someone changes it. You'll have
the situation where Office runs on old Windows + apps group enhanced com, but apps group enhanced com doesn't work on enhanced windows from OS group.
While the resulting confusing will certainly cause numerous problems for Windows users, much to the joy of Linux advocates everywhere, I thought
the DOJ was supposed to be looking out for the consumer.
Or is this really a "stick it to Bill" solution, and nobody cares if the consumers get screwed?
P-III chips have a true random number generator
that uses thermal noise to generate the number.
Unless you believe the paranoia and think the NSA conviced Intel to subvert their design (making it more expensive to manufacture!), thermal noise
is about the best way to get random numbers for cryptography.
Correct me if I'm wrong, but Amazon was
always running on Unix (solaris, I think).
They are switching from one Unix web server to another Unix web server (probably because
no one trusts that Netscape Enterprise server will be around in a few years).
Barnes & Nobles on the other hand, uses NT/IIS.
As far as I can see, both work fine.
Well, internal to the Teradata database system,
dates were orignally an integer YYMMDD, and
when we need to allow for 2000, it was changed
so that instead of YY it was CCYY-100, so
1999 was 99, 2000 was 100, just as you mentioned.
That way, existing dates stored in the database didn't need updating.
This was implemented 10 years ago...
I had my eyes done with PRK (Photoreactive Kerotectomy). Before, I was a -8.75 and -9,
with about 1 diopter astigmatism. Now, I am about 0 and -1.5, so I see well enough to not need glasses for driving (20/25 in one eye, 20/35 in the other). I normally wear glasses to correct the remaining astigmatism and nearsightednesss,
but can do without.
But I was also one of the unfortunates who had serious side effects from my surgery. My night vision is terrible, with huge halos around any light sources. After many years, the problem has not gone away. I've also lost a little bit of contrast in my vision.
Although they tried to correct my astigmatism, they were not very successful at that.
Overall, I am OK with the results, but it has definitely involved tradeoffs, and once it's done, there is no going back.
If I were to do it again, I would not pick PRK, but rather LASIK. LASIK involves a risky step of cutting a "flap" on the cornea so the laser can be applied to the inner layers of the cornea, rather than the surface as is done for PRK.
But LASIK generally results in better vision, and
the recovery time is much faster. It was,
unfortunately, not available when I had my eyes done.
If you go for LASIK, understand that there are definitely risks. Also, Although many don't need glasses afterwards, some do.
If I remember right, Intel planned to go with
FireWire, but the large cost (due to Apple's insistance on large royalties) jacked up the price of the motherboard chipsets enough that the motherboard manufactures told Intel "no way".
Even $5 is a huge add-on cost to this market, and if I rememer, FireWire was going to add at least double that.
Competition, yes, but 5 microsofts all with the
source means 5x more versions of Windows.
How does this help the consumer?
Gee, nifty game x works on Company A's Windows,
but the office package I want runs on Company D's windows, and the e-mail package work mandates runs on company B's windows.... Ugh.
Office already runs on the next most popular
OS: MacOS. Having Office on MAC didn't cause
Windows to lose out. Having Office on some
variant of UNIX wouldn't either.
As one of the punishments, this isn't a bad idea.
Forcing MS to open the documentation on their many proprietary file formats and APIs would help even the playing field for startup competitors.
While that would hurt Mr Gates, what would
it do for the consumers? If the published
source code is still MS proprietary, no one can
use it in building any new projects without paying
MS. If you want to make it Public Domain, or GPL, or BSD, etc., you are talking about the
goverment siezing the property of a corporation (and therefor from the shareholders). do you really want the goverment going around siezing private property?
If you broke Microsoft into OS and Apps,
you'd have two monopolies instead of one: Company one has a monopoly on desktop OS sales,
the other has it on Office application suites.
How does the consumer win?
And where would you make the divide? Is IE
part of the OS or the applications?
Can the OS have an e-mail client (OE) or is that part of the applications?
Which Build of W2K? I used to have this problem,
but with the latest build I tried (2128), it installed in about 20 minutes, with almost no user interactions needed during that time.
Microsoft needed Softway to allow NT to acheive
POSIX compatibility and UNIX95 branding,
which is important for selling into the
Federal marketplace.
Softway wasn't making much money on this software,and was in danger of going out of
business or dropping the product, both of which
were unacceptable to Microsoft.
Yes, the connection from the DSLAM to the ISP
is shared, but so is the internet.
In my case, The DSLAM is on PacBell's OC-45 ATM network, so the bandwidth is higher than the
internet backbones usually carry.
In my case, DSL DSLAM is tied directly to
PacBell's OC-45 ATM network. The ISP connects
to this network and "bridges" the gap to the
internet. But the OC-45 is so much faster than
most internet backbones, I'm not worried about
the bandwidth.
And I love my aDSL modem! I get consistant
bandwidth (usually 1 mbit/sec), and a fixed IP
address...
There are different kinds of DSL. G.Lite uses
the same wire pair as the phones, with a filter
at each phone to reduce the interference.
Regular (old-style?) ADSL uses a splitter at
the enterance to the residence, and then separate
wires inside. This works much better, but is
more costly to install.
Companies can operate at a loss for as long
as they have money or liquid assests. When they
run out, the company goes backrupt.
Startup companies get money from Venture Capitalists for a share of the company (or borrow money). When it runs out, they need to ask for more, or IPO (or be making a profit).
The whole point behind an IPO was traditionally to get enough money to fund a companies growth (allow them to spend more money than revenue for a while, to grow the business). IPO has instead become a way for investors to get rich quick, although it still serves the purpose of funding the company.
When the money RedHat got in the IPO, plus any reserves they had from before, run out, then RedHat will either need to be profitable, need to find a buyer, or go bankrupt.
So, the IPO money is the "savings account" they are dipping into, and it better not run out before they find some way to earn a living!
I agree with some of your post, but "it's obvious that pretty soon they're going to be making money hand over fist" doesn't fit my understanding.
Most of RedHat's profits will come from services (Linux can be copied for free, right?). Growing a services business is slow and difficult (you need to hire many hundreds of linux experts, and find work for all of them).
Linux is a hot buzz word in the market, but most shareholders don't understand RedHat's business.
Worth $7.4B? Only if you believe they will growing at an enormous rate, with revenues growing 100%/quarter, and huge profits to follow.
Some shareholders are believers. Most are subscribing to the "bigger fool" theory that says there is always somebody who will buy the stock for a higher price later even if it is rediculously valued.
RedHat can't possibly live up to the hype and stock price. They are destined to fall, but when? Ahhh... If you only knew, you could make a fortune...
Ummmm... If they keep losing money, eventually they will not have any money to pay employees.
When they were private, the could pay employees from money received from VCs or borrowed, and for the near term they have the money from the IPO.
When that runs out, they need to be making money, or else!