Well of course. CAD == Canadian Dollars, USD == United States of Americas Dollars (I'm making this up as I go along), and Pounds == weight, force, gold, or how much you have to lose. I'm not sure what a Euro is yet (some kind of train in a tunnel?)
Canadians are the among the shyest people on the planet, which explains why we cannot agree on a simple one, two, or three letter acronymn. Take any three-letter combination of C,D,N, and A, and we'll hold a referendum (using pen and paper) to decide the winner;-)
Our bills are works of art -- assorted indigenous critters on one side, and an 80+ year queen on the other -- but for some reason the rest of the world prefers dead presidents;-)
When/. points to Yahoo for "news" about Einstein, and the next item is Y-news is "The Worlds Worst Bathrooms", I hope a few E, M, and C squared smite the editors.
Slashdot articles for years, as it actually matters to nerds. Judging from the number and overall quality of the comments to this post, I think many other readers would agree.
I don't have much to contribute to this discussion -- I'm only a Eng-Phys grad whos been writing software for 20+ years (with a 3-week course in Fortran as my only formal training in CS) -- but I do have one question. What is an FTE developer?
MS did not pioneer this technology, but when they (Bill) took/stole/bought the existing tech, and made the deal with IBM, the floodgates opened. The combo of IBM-PC hardware and cheap DOS software made it possible for many people and companies, at the time, to have a kick at the cat.
This deal was not an innovation, or an invention, but it sure as hell changed the way 90+ percent of us look at and use computers today.
I think a single, well educated man, could reconstruct a good part of the tech he grew up in. If you know something is *possible* (like a steam engine, or an electrical generator, or an a-bomb,...), the
hard work is done... all you need is a few (?) believers to help
build the widget.
I'm being a slight prick here, but if you spent 200+ hours learning the guts of *any* distro of Linux, or Windows, or..., you will be an ubergeek (or close to one) in that system. And if you can open the hood, and make your car go over 40 mph with the wipers turned on, share that knowledge.
Disclaimer: Windows, of any flavor, scares the hell out of me -- I'll stick with Slakware.
Microsoft created an equivalent tool for their OS devision? If a program can turn crappy art into a thing of beauty, can't the same thing be done to the code in Windows xxx itself?
I've written RTLinux applications on boxes using the 633 MHz Via cpu, and found the typical interrupt latency was ~10 usec. About the same latency as on a box with a 3.2 GHz P4, which indicates the limitation is not the cpu speed,
but with the interrupt controller circuitry.
I think you have *way* too many zeros in your comment -- a 1Ghz machine runs 1,000 clock cycles in a microsecond (multiply by N for a N-GHz cpu). However, I agree that it sucks that thousands of clock cycles are required to respond to an interrupt, even one generated by the cpu itself (such as by the internal clock).
I've been using the GPL version of RTLinux for years now, in high-speed data logger and control systems in the geophysics and biomedical fields. The problem (difficulty) is not commodity hardware (we've used off-the-shelve P1:P4, AMD, and Via boxes), it's writing the RTLinux drivers, or kernel modules, to communicate with and control the devices, such as a/d cards, serial devices, etc. That is, you can't take an application written for a non-RTLinux system, plop it on an RTLinux box, and expect hard realtime response -- you have to rewrite it.
Thank you Bob (and Nerd TV), for providing a text-transcript of your interview with Andy.
For people without broadband/torrent/mp4/audio
stuff on their computer, text is the only way.
Kinda like books.
I am not a security expert, and know nothing about MACs, and even less (than zero) about NT.
I'm puzzled... if Lampson and Cutler have been working at Microsoft for over a decade, and MS, with the largest army of developers/software testers on the planet, cannot make Windows security easy to use, who can?
"Windows NT had MACs built in from day one.
They are not the same as the VMS implementation-
and for very good reason...
The problem with Windows security has absolutely
nothing to do with lack of security features.
The problem is the exact reverse, the problem
is too many damn features and applications that
can't make use of them."
So, what use are 'security features' if most people cannot understand them, let alone use them?
I think you have (accidentally?) come up with a killer Slashdot app, nodentity.
Imagine: a 'map' button on Slashdot that lets you see where the poor wittle server is located, and where the hits are coming from. Plotted in pretty colours on a large rock sphere, of course.
Shit... I think a real-time display of the Slashdot effect would almost be as funny as the comments;-)
theme in "Contact" is still the best way of communicating (one-way) with alien civilizations via C (the speed, not the language).
Transmit "I Love Lucy", or whatever, using your current method of choice (RF, UHF, X-Rays) to get their attention. Then broadcast a message with the blueprints of the neatest/scariest stuff you can.
If they can decipher the primer, the grad students can do the rest. However, it may take a few eons before we receive a reply, or friendly visit;-)
You've calulated the volume of the chip correctly (apart from using units of mm^2 instead of mm^3), but how in hell did you come up with the volume of a VW?
1m = 1000mm
=> 7,710,952.32 mm^3 = 0.00771 m^3 =~ a 20x20x20 cm cube
I find it hilarious that Wells Fargo, one of the first banks in the US, and the subject/victim of countless stage coach robberies during the Wild West Era, are converting their (small collection of) ATMs to run under Winwhatever.
I suggest that they: 1) post at least two shotgun- wielding gaurds at each station, and 2) get the posses ready, before they deploy.
And of course, talk to Hollywood ASAP... I can see a lot of cool remakes of of Westerns coming to your neighbourhood soon;-)
In 77, as a summer student, I started working with grad students and techs on a DEC minicomputer (Nova). The precurser to the Eclipse (Soul of a New Machine). This hairy monster had a whopping 8K of core memory, a paper tape reader with Basic and Fortan compilers, and a disfunctional 1-Mbyte hard drive (which we fixed that summer... a resister pack went south). It was a very expensive machine, but we could run scientific routines (such as FFTs) on it as fast as the mainframe (IBM 360) on campus. And it was all ours (most of the code included)!
Two years later, the 'cheap' state of the art was the KIM board. A 6502 processor, with a casette reader/writer, a keypad/display, and a assembler. Twas a bitch to work with (because of the tapes), but programming the 6502 chippy was a no-brainer. And it could do the same calcs as the Nova (and the 360) with a bit of programming.
Skip forward several years. S100 computers, CP/M, floppy disks (8 inch, then 5.25 inch) and hard drives. And most important, a C-compiler for the machine. Sorry, I can't remember the name of compiler, but it worked (apart from floating point stuff) on our 8086 S100 machines like a charm.
We then switched from CP/M (and it's multitasking progeny) to MSDOS V1.0. I cursed my superviser (slighty) for the change, but he was right at the time -- CP/M was toast, and MSDOS was to rule.
Eventually, C (and other) compilers for MSDOS came out, and life was fine again. Apart from making backups from a MSDOS 2.0 machine, and restoring them to 3.x/4.x/... machines (Word.docs anyone?).
When MSDOS was written (stolen,...), there were x86 C-compilers, but at the time, you could not write an OS using C -- it was not the right tool.
To me, *Scientific Applications* means number crunching. You can use any (many) programs to view the output of a program, but for heavy duty sci/comp jobs, like CFD computations, transforms using FFTs or Wavelets,..., the bottleneck is in the FPU of the processor, not the GPU.
If you could program this (or any) graphics card to run a FFT, or a SVD, or whatever algorithm you need, with acceptable precision, and it runs X times faster on the card than the native processer (P4/AMD64/...)... Profit!
I've developed several real-time aquisition and control systems using RT-Linux on x86 hardware (using Slackware x.x), and it rocks. 10 KHz rates for a/d and d/a control with ~10 microsecond latency, under *any* kind of load (disk, network, video,...).
Granted, you have to know how to apply a patch to the kernel, (and write a driver for your application), but in the end, it will work 24/7, and all the other linux/gnu/... stuff on top will not know/see the difference in the kernel.
So, if it smells like a linux, looks like a linux, and waddles like a linux, is it a linux?
As a citizen of a sovereign country (Canada), I can do *everything* within it borders that is allowed within its laws. Driving while drunk is a big no-no everywhere, yet the definition of DUI varies (from BC to Nfld).
If I do not visit/trade/deal with anyone in the US, their government should not have any say in my business (other than the recording guys;-(
Disclaimer: I was working at a military base in the US when the towers were hit. I'm glad I had a Canadian passport with me.
Slashdot Mirror
and tech-writers impose the same rules of conduct on themselves, Hell will be frozen.
We are one of the oil-producing/exporting countries. We meddle with ourselves.
Canadians are the among the shyest people on the planet, which explains why we cannot agree on a simple one, two, or three letter acronymn. Take any three-letter combination of C,D,N, and A, and we'll hold a referendum (using pen and paper) to decide the winner ;-)
Our bills are works of art -- assorted indigenous critters on one side, and an 80+ year queen on the other -- but for some reason the rest of the world prefers dead presidents ;-)
When /. points to Yahoo for "news" about Einstein, and the next item is Y-news is "The Worlds Worst Bathrooms", I hope a few E, M, and C squared smite the editors.
Slashdot articles for years, as it actually matters to nerds. Judging from the number and overall quality of the comments to this post, I think many other readers would agree.
I don't have much to contribute to this discussion -- I'm only a Eng-Phys grad whos been writing software for 20+ years (with a 3-week course in Fortran as my only formal training in CS) -- but I do have one question. What is an FTE developer?
Dos x.yy on the IBM-PC.
MS did not pioneer this technology, but when they (Bill) took/stole/bought the existing tech, and made the deal with IBM, the floodgates opened. The combo of IBM-PC hardware and cheap DOS software made it possible for many people and companies, at the time, to have a kick at the cat.
This deal was not an innovation, or an invention, but it sure as hell changed the way 90+ percent of us look at and use computers today.
I'd call it fate
I think a single, well educated man, could reconstruct a good part of the tech he grew up in. If you know something is *possible* (like a steam engine, or an electrical generator, or an a-bomb, ...), the
hard work is done ... all you need is a few (?) believers to help
build the widget.
I'm being a slight prick here, but if you spent 200+ hours learning the guts of *any* distro of Linux, or Windows, or..., you will be an ubergeek (or close to one) in that system. And if you can open the hood, and make your car go over 40 mph with the wipers turned on, share that knowledge. Disclaimer: Windows, of any flavor, scares the hell out of me -- I'll stick with Slakware.
Microsoft created an equivalent tool for their OS devision? If a program can turn crappy art into a thing of beauty, can't the same thing be done to the code in Windows xxx itself?
I've written RTLinux applications on boxes using the 633 MHz Via cpu, and found the typical interrupt latency was ~10 usec. About the same latency as on a box with a 3.2 GHz P4, which indicates the limitation is not the cpu speed, but with the interrupt controller circuitry.
I think you have *way* too many zeros in your comment -- a 1Ghz machine runs 1,000 clock cycles in a microsecond (multiply by N for a N-GHz cpu). However, I agree that it sucks that thousands of clock cycles are required to respond to an interrupt, even one generated by the cpu itself (such as by the internal clock).
I've been using the GPL version of RTLinux for years now, in high-speed data logger and control systems in the geophysics and biomedical fields. The problem (difficulty) is not commodity hardware (we've used off-the-shelve P1:P4, AMD, and Via boxes), it's writing the RTLinux drivers, or kernel modules, to communicate with and control the devices, such as a/d cards, serial devices, etc. That is, you can't take an application written for a non-RTLinux system, plop it on an RTLinux box, and expect hard realtime response -- you have to rewrite it.
Thank you Bob (and Nerd TV), for providing a text-transcript of your interview with Andy. For people without broadband/torrent/mp4/audio stuff on their computer, text is the only way. Kinda like books.
I am not a security expert, and know nothing about
... if Lampson and Cutler have been
...
MACs, and even less (than zero) about NT.
I'm puzzled
working at Microsoft for over a decade, and MS,
with the largest army of developers/software
testers on the planet, cannot make Windows
security easy to use, who can?
"Windows NT had MACs built in from day one.
They are not the same as the VMS implementation-
and for very good reason
The problem with Windows security has absolutely
nothing to do with lack of security features.
The problem is the exact reverse, the problem
is too many damn features and applications that
can't make use of them."
So, what use are 'security features' if most
people cannot understand them, let alone use them?
I think you have (accidentally?) come up with a killer Slashdot app, nodentity.
Imagine: a 'map' button on Slashdot that lets you see where the poor wittle server is located, and where the hits are coming from. Plotted in pretty colours on a large rock sphere, of course.
Shit
the comments
theme in "Contact" is still the best way of
;-)
communicating (one-way) with alien civilizations
via C (the speed, not the language).
Transmit "I Love Lucy", or whatever, using your
current method of choice (RF, UHF, X-Rays) to
get their attention. Then broadcast a message
with the blueprints of the neatest/scariest stuff
you can.
If they can decipher the primer, the grad students
can do the rest. However, it may take a few eons
before we receive a reply, or friendly visit
Shit, can any here do basic problem solving? Dgatwood ... 7,869,876.85 is the number of
**chips** in a VW ... not the number of LOCS!
You've calulated the volume of the chip
;-)
correctly (apart from using units of mm^2
instead of mm^3), but how in hell did you
come up with the volume of a VW?
1m = 1000mm
=> 7,710,952.32 mm^3 = 0.00771 m^3
=~ a 20x20x20 cm cube
I think your VW shrank in the rain/sun cycle
I find it hilarious that Wells Fargo, one of the
... I can ;-)
first banks in the US, and the subject/victim of
countless stage coach robberies during the Wild
West Era, are converting their (small collection
of) ATMs to run under Winwhatever.
I suggest that they: 1) post at least two shotgun-
wielding gaurds at each station, and 2) get the
posses ready, before they deploy.
And of course, talk to Hollywood ASAP
see a lot of cool remakes of of Westerns coming
to your neighbourhood soon
I'll raise you one ...
... a
.docs anyone?).
...), there were
In 77, as a summer student, I started working with
grad students and techs on a DEC minicomputer
(Nova). The precurser to the Eclipse (Soul of a
New Machine). This hairy monster had a whopping
8K of core memory, a paper tape reader with Basic
and Fortan compilers, and a disfunctional 1-Mbyte
hard drive (which we fixed that summer
resister pack went south). It was a very
expensive machine, but we could run scientific
routines (such as FFTs) on it as fast as the
mainframe (IBM 360) on campus. And it was all
ours (most of the code included)!
Two years later, the 'cheap' state of the art was
the KIM board. A 6502 processor, with a casette
reader/writer, a keypad/display, and a assembler.
Twas a bitch to work with (because of the tapes),
but programming the 6502 chippy was a no-brainer.
And it could do the same calcs as the Nova (and
the 360) with a bit of programming.
Skip forward several years. S100 computers, CP/M,
floppy disks (8 inch, then 5.25 inch) and hard
drives. And most important, a C-compiler for the
machine. Sorry, I can't remember the name of
compiler, but it worked (apart from floating
point stuff) on our 8086 S100 machines like a
charm.
We then switched from CP/M (and it's multitasking
progeny) to MSDOS V1.0. I cursed my superviser
(slighty) for the change, but he was right at
the time -- CP/M was toast, and MSDOS was to rule.
Eventually, C (and other) compilers for MSDOS came
out, and life was fine again. Apart from making
backups from a MSDOS 2.0 machine, and restoring
them to 3.x/4.x/... machines (Word
When MSDOS was written (stolen,
x86 C-compilers, but at the time, you could not
write an OS using C -- it was not the right tool.
Now?
In what language is Win 9x/200x/XP/... written?
To me, *Scientific Applications* means number ..., the bottleneck is
... Profit!
crunching. You can use any (many) programs to
view the output of a program, but for heavy duty
sci/comp jobs, like CFD computations, transforms
using FFTs or Wavelets,
in the FPU of the processor, not the GPU.
If you could program this (or any) graphics card
to run a FFT, or a SVD, or whatever algorithm
you need, with acceptable precision, and it runs
X times faster on the card than the native
processer (P4/AMD64/...)
I've developed several real-time aquisition ...).
and control systems using RT-Linux on x86
hardware (using Slackware x.x), and it rocks.
10 KHz rates for a/d and d/a control with ~10
microsecond latency, under *any* kind of load
(disk, network, video,
Granted, you have to know how to apply a patch
to the kernel, (and write a driver for your
application), but in the end, it will work 24/7,
and all the other linux/gnu/... stuff on top
will not know/see the difference in the kernel.
So, if it smells like a linux, looks like a
linux, and waddles like a linux, is it a linux?
As a citizen of a sovereign country (Canada),
;-(
I can do *everything* within it borders that
is allowed within its laws. Driving while
drunk is a big no-no everywhere, yet the definition of DUI varies (from BC to Nfld).
If I do not visit/trade/deal with anyone in the US, their government should not have any say
in my business (other than the recording guys
Disclaimer: I was working at a military base
in the US when the towers were hit. I'm glad
I had a Canadian passport with me.
The folks at Canopy are now suing each other (see the Register for details). That is justice.