Seriously, as much as I'd love to be able to afford a high-end graphics card, I'm more interested in finding the cheapest card which will support Oblivion reasonably. Any recommendations?
It sounds interesting but I'm still frustrated by the lack of support for Ogg format in consumer electronics. I recently had to re-rip a bunch of my CDs into MP3 format in order to take them on a one month car trip. I couldn't find an car CD players that understood Ogg.
The Mech Eng department inherited the old IBM 1401 when the second 360 arrived around 1970. They had a simulation program on the 1401 that was ancient even then - but it worked. It was simpler to take the old 1401 and set up the minimal configuration needed to keep the beast running. This pared down dinosaur was relegated to the most inconvenient back corner of the ops centre.
Now core memory is amazingly reliable. It just doesn't fail. But human judgement certainly does!
Over the years, some clever hack and figured out how to rerun a simulation using the alternate (B) model - it could be set by toggling a single core (bit). A second clever hack actually mapped the bit to the core memory block and lo-and-behold, the bit was on the most nearly exposed core plane! All you had to do was loosen some protective shielding and you could see the magic bit - right there!
Now you could do this from the control panel but there was some odd constraint which meant that you would have to reload the (identical) input dataset for the second run. Needless-to-say, over time, it became almost routine to show off the quaint old box by using a tapered wooden ruler to do the reset.
You had to be careful because each core was only about a millimeter across. And the core modules were not intended to be exposed. But, heh!, the beast was built like a tank and it was only ever used for this one simulation (which meant you never needed to load the program - it just persisted in core indefinitely.) So, the appropriate chassis was typically left with the cover plate off and the magic core was highlighted with whiteout.
The core memory was easier to reach from the side than from the front but as a consequence, you had to stretch while holding the wooden ruler that we used to flip the chronic offender. And this is where the inevitable happened. While trying to pass on this secret technique to a grad student, the two of us were forced into tight quarters. As he leaned past me to get a closer look, he slipped. And as he fell, he grabbed me and pulled me down on top of him. And the wooden ruler you ask? Yes, you guessed it; It ripped a gash through fully half the core plane!
It get's pretty tiresome seeing "linux-powered" this and that. I'm as big a fan of Linux as the next guy, but Linux is just the O/S. Stories like this make it sound like Linux is controlling the car. It's *not*. It's controlled by somebody else's software.
It's not even an embedded Linux story. It's just another neat PC application that could just as easily be running on a Mac.
Mr. Chuck wrote: > If it TRULY sucks, you can leave. If you can > make it a year first, great. > > After that you have a track record that you can > survive at a place for a year. (it's important, > really).
A year is too short. When I'm reviewing resumes, one of the first things I do is evaluate the person's record. If I see a lot of bouncing around, then I'm not interested. It costs a lot of time(=money) and energy to train a new employee and we want to keep them for the long haul.
I know that the climate has been harsh the past few years but that can be taken into account.
This reinforces my firm belief that coders should *never* design GUIs for non-coders.
Much as I despise MS Windows, their tree navigator just plain *works*.
1. As with every other file navigator, this one consumes too much space for the amount of information that it presents. The actual filenames only consume ~24% of the window (yes, enlarging the window will improve the ratio but it is a lot of wasted space)
2. The ability to switch between different view modes requires the user to pull down a list and select an entry. The space consumed by the list pulldown widget is enough to present a small but coherent set of button images.
3. The favourite locations sub-window is full of fluff and borrows too heavily from windows. Remember that one of the beautiful things about Unix is the rational filesystem layout: it is a heirarchy under/. Windows requires this fluff because they separate out all the different filespaces into disjoint areas. We don't need it - this is where a simple tree navigator is a perfect solution.
If you have "favourite" locations, then put them on your desktop and then use drag-and-drop. That way, the favourites are confined to the file navigator - they can be used anywhere.
4. The search dialogue is entirely superfluous. Just type! Use something like Mozilla's searching capability.
5. Font/Line sizes: enough with the white space already. Stop wasting so much space between lines of text. I want to see more information.
Sorry to rain on your parade people but please go back to the drawing board.
TI calculators would be a lot more interesting if they didn't force me to use algebraic entry. A modified OS that let me turn "=" into "enter" would be a big step forward.
In the meantime, I'll stick with my HP-41CX and HP-16C.
I'd rather get my excitement from non-technological pursuits such as rock-climbing and motorcycling. High-tech merely substitutes stress for stimulation.
Whatever happened to technology as a tool rather than an end-until-itself? Do you notice how most of the items in BT's list offer little or no "added-value" - they are merely demonstrations of our technological prowess.
[No, I'm not a luddite. I love my toys and have been programming professionally for a quarter of a century now.]
The US government still hasn't clued in to the fact that The Internet is a global resource. As such, local entities such as governments are going to have very little power to control it.
The proper solution is to get off our butts and start migrating away from SMTP.
In an earlier article http://slashdot.org/article.pl?sid=03/10/19/211820 1&mode=thread&tid=126&tid=185 a near-future AI machine was foreseen as managing a call-centre. It's not too much of a leap to imagine that the world's first commercial application of AI will instead be to sell us more unnecessary junk.
And since I can't tell the AI apart from a human, let alone a sales-droid, I'll be doomed.
Now I feel like slitting my wrists.
1. Type the same thing three times in a row 2. Decide that this task could be replaced by a
shell script. 3. Spend the afternoon perfecting and documenting
the 400 shell/perl/expect/... script so that I
can save 30 seconds a day for the next few
months. 4. Find a better solution on sourceforge
You *can't* switch to another set of domain servers. To permit alternate TLDs (as has already been attempted without success) would lead to a partitioned Internet.
I agree with the original poster - somethings should be regulated and commerce should not be the only motivating factor in all matters of policy.
This isn't much of a workaround since the mistyped DNS name will still resolve. Instead of a no-such-domain response from the resolver, you'll instead get a no-response at the application level. This suggests that the server (website or mailserver for example) exists but is down.
In the case of SMTP traffic, the sender will waste time and bandwidth retrying.
Note also that Mockapetris explicitly intended for wildcarding to be supported in RFC1034 - unfortunately, I don't think he foresaw the crass exploitation of the internet by ICANN 16 years ago.
There will still be physical media, it just that the producers are going to shift the extra cost to the consumer. Supply your own damn cdrw. Print off your own liner notes.Buy your own jewel case.
For rentals - sure. But for music I will still want something to hold in my hand and read while I listen.
Theory doesn't make sense
on
P2P Spam?
·
· Score: 2, Insightful
If this theory were true, then the "test" virii would be much more benign. Since they have been quite noticable, people have been compelled to take steps to close the holes. I would suspect that the next variant will be much less of a nuisance than its predecessors simply because the target market has been substantially reduced.
No, if I was looking for a fun conspiracy theory, I would enjoy suspecting that Microsoft has decided that this is a good time have all their customers tighten up their security.
That's a good point and one that should be considered whenever one patches the source. However some of the patches are trivial and "obviously" safe while others are additions that don't actually require changes to the qmail source itself.
Because of qmail's design, it is very resistent to compromise, even if one of the components is modified.
I believe that the strict partitioning of function in qmail lends itself better to extension than a constantly evolving package such as sendmail.
The DoS problem doesn't lie with qmail itself. That particular issue is best addressed through thresholding which is supported by ucspi-tcp's tcpserver (a replacement for inetd or xinetd).
If you are using ucspi-tcp already, then it is probably as simple as modifying the contents of/var/qmail/control/concurrencyincoming.
ucspi-tcp is not *required* but much of the qmail documentation assumes that you are using it. ucspi-tcp is also written by Dan Berstein (cr.yp.to/ucspi-tcp.html)
What can you do with sendmail that you can't to with qmail? There is a a very large set of mature additions and patches to qmail that permit just about anything you may wish to undertake with your mail server.
On the point of qmail being cumbersome: I disagree - what could be simpler than adding a single line to your rcpthosts file? Maintaining qmail is trivial. However, I'll agree that the author's terse documentation makes it seem quite foreign but compared to sendmail it is positively didactic. There are also many other resources available which supplement the original docs.
Qmail is rock-solid. The best proof I can offer is that fact that no security flaw has been found since 1.03 was released in 1998. The man is a cryptographer and designed it for security.
There is also an enormous amount of support for the product available. Check out qmail.org and cr.yp.to/qmail.html
At the two extremes, there are idiot users and idiot tech support staff. However, the majority of support calls fall somewhere in between where the hurdle to overcome is not knowledge but being able to communicate using the same language. I find that being able to "talk da lingo" is immeasurably helpful.
For example, I have successfully been able to get routed directly to technicians in the CO merely because I have been able to convince the Bell Canada staff that I know what I am talking about.
This leads me to believe that end-users have a certain responsibility to acquire a modicum of background understanding. There will always be incompetent support staff but if an end-user can accurately describe their problem, their chances of getting satisfactory help increase dramatically.
Slashdot Mirror
Seriously, as much as I'd love to be able to afford a high-end graphics card, I'm more interested in finding the cheapest card which will support Oblivion reasonably. Any recommendations?
It sounds interesting but I'm still frustrated by the lack of support for Ogg format in consumer electronics. I recently had to re-rip a bunch of my CDs into MP3 format in order to take them on a one month car trip. I couldn't find an car CD players that understood Ogg.
is the fact that this article is coming from TechZone and not NYTimes or the Washington Post.
Sorry, nobody else here remembers wow and flutter specs from turntables. face it - we're old.
The Mech Eng department inherited the old IBM 1401 when the second 360 arrived around 1970. They had a simulation program on the 1401 that was ancient even then - but it worked. It was simpler to take the old 1401 and set up the minimal configuration needed to keep the beast running. This pared down dinosaur was relegated to the most inconvenient back corner of the ops centre.
Now core memory is amazingly reliable. It just doesn't fail. But human judgement certainly does!
Over the years, some clever hack and figured out how to rerun a simulation using the alternate (B) model - it could be set by toggling a single core (bit). A second clever hack actually mapped the bit to the core memory block and lo-and-behold, the bit was on the most nearly exposed core plane! All you had to do was loosen some protective shielding and you could see the magic bit - right there!
Now you could do this from the control panel but there was some odd constraint which meant that you would have to reload the (identical) input dataset for the second run. Needless-to-say, over time, it became almost routine to show off the quaint old box by using a tapered wooden ruler to do the reset.
You had to be careful because each core was only about a millimeter across. And the core modules were not intended to be exposed. But, heh!, the beast was built like a tank and it was only ever used for this one simulation (which meant you never needed to load the program - it just persisted in core indefinitely.) So, the appropriate chassis was typically left with the cover plate off and the magic core was highlighted with whiteout.
The core memory was easier to reach from the side than from the front but as a consequence, you had to stretch while holding the wooden ruler that we used to flip the chronic offender. And this is where the inevitable happened. While trying to pass on this secret technique to a grad student, the two of us were forced into tight quarters. As he leaned past me to get a closer look, he slipped. And as he fell, he grabbed me and pulled me down on top of him. And the wooden ruler you ask? Yes, you guessed it; It ripped a gash through fully half the core plane!
Oh Crap!
It get's pretty tiresome seeing "linux-powered" this and that. I'm as big a fan of Linux as the next guy, but Linux is just the O/S. Stories like this make it sound like Linux is controlling the car. It's *not*. It's controlled by somebody else's software.
It's not even an embedded Linux story. It's just another neat PC application that could just as easily be running on a Mac.
Mr. Chuck wrote:
> If it TRULY sucks, you can leave. If you can
> make it a year first, great.
>
> After that you have a track record that you can
> survive at a place for a year. (it's important,
> really).
A year is too short. When I'm reviewing resumes, one of the first things I do is evaluate the person's record. If I see a lot of bouncing around, then I'm not interested. It costs a lot of time(=money) and energy to train a new employee and we want to keep them for the long haul.
I know that the climate has been harsh the past few years but that can be taken into account.
Or the document's spelling and grammatical errors are there in order to identify the mole.
This reinforces my firm belief that coders should *never* design GUIs for non-coders.
/. Windows requires this fluff because they separate out all the different filespaces into disjoint areas. We don't need it - this is where a simple tree navigator is a perfect solution.
Much as I despise MS Windows, their tree navigator just plain *works*.
1. As with every other file navigator, this one consumes too much space for the amount of information that it presents. The actual filenames only consume ~24% of the window (yes, enlarging the window will improve the ratio but it is a lot of wasted space)
2. The ability to switch between different view modes requires the user to pull down a list and select an entry. The space consumed by the list pulldown widget is enough to present a small but coherent set of button images.
3. The favourite locations sub-window is full of fluff and borrows too heavily from windows. Remember that one of the beautiful things about Unix is the rational filesystem layout: it is a heirarchy under
If you have "favourite" locations, then put them on your desktop and then use drag-and-drop. That way, the favourites are confined to the file navigator - they can be used anywhere.
4. The search dialogue is entirely superfluous. Just type! Use something like Mozilla's searching capability.
5. Font/Line sizes: enough with the white space already. Stop wasting so much space between lines of text. I want to see more information.
Sorry to rain on your parade people but please go back to the drawing board.
TI calculators would be a lot more interesting if they didn't force me to use algebraic entry. A modified OS that let me turn "=" into "enter" would be a big step forward.
In the meantime, I'll stick with my HP-41CX and HP-16C.
I'd rather get my excitement from non-technological pursuits such as rock-climbing and motorcycling. High-tech merely substitutes stress for stimulation.
Whatever happened to technology as a tool rather than an end-until-itself? Do you notice how most of the items in BT's list offer little or no "added-value" - they are merely demonstrations of our technological prowess.
[No, I'm not a luddite. I love my toys and have been programming professionally for a quarter of a century now.]
The US government still hasn't clued in to the fact that The Internet is a global resource. As such, local entities such as governments are going to have very little power to control it.
The proper solution is to get off our butts and start migrating away from SMTP.
In an earlier article http://slashdot.org/article.pl?sid=03/10/19/211820 1&mode=thread&tid=126&tid=185 a near-future AI machine was foreseen as managing a call-centre. It's not too much of a leap to imagine that the world's first commercial application of AI will instead be to sell us more unnecessary junk.
And since I can't tell the AI apart from a human, let alone a sales-droid, I'll be doomed.
Now I feel like slitting my wrists.
1. Type the same thing three times in a row
2. Decide that this task could be replaced by a
shell script.
3. Spend the afternoon perfecting and documenting
the 400 shell/perl/expect/... script so that I
can save 30 seconds a day for the next few
months.
4. Find a better solution on sourceforge
You *can't* switch to another set of domain servers. To permit alternate TLDs (as has already been attempted without success) would lead to a partitioned Internet.
I agree with the original poster - somethings should be regulated and commerce should not be the only motivating factor in all matters of policy.
This isn't much of a workaround since the mistyped DNS name will still resolve. Instead of a no-such-domain response from the resolver, you'll instead get a no-response at the application level. This suggests that the server (website or mailserver for example) exists but is down.
In the case of SMTP traffic, the sender will waste time and bandwidth retrying.
Note also that Mockapetris explicitly intended for wildcarding to be supported in RFC1034 - unfortunately, I don't think he foresaw the crass exploitation of the internet by ICANN 16 years ago.
There will still be physical media, it just that the producers are going to shift the extra cost to the consumer. Supply your own damn cdrw. Print off your own liner notes.Buy your own jewel case.
For rentals - sure. But for music I will still want something to hold in my hand and read while I listen.
If this theory were true, then the "test" virii would be much more benign. Since they have been quite noticable, people have been compelled to take steps to close the holes. I would suspect that the next variant will be much less of a nuisance than its predecessors simply because the target market has been substantially reduced.
No, if I was looking for a fun conspiracy theory, I would enjoy suspecting that Microsoft has decided that this is a good time have all their customers tighten up their security.
I suppose I should have emphasized the "not" as well as the "required". You do *not* have to use ucspi-tcp with qmail if you don't want to.
That's a good point and one that should be considered whenever one patches the source. However some of the patches are trivial and "obviously" safe while others are additions that don't actually require changes to the qmail source itself.
Because of qmail's design, it is very resistent to compromise, even if one of the components is modified.
I believe that the strict partitioning of function in qmail lends itself better to extension than a constantly evolving package such as sendmail.
I'm not in a position to compare it to Postfix.
The DoS problem doesn't lie with qmail itself. That particular issue is best addressed through thresholding which is supported by ucspi-tcp's tcpserver (a replacement for inetd or xinetd).
/var/qmail/control/concurrencyincoming.
If you are using ucspi-tcp already, then it is probably as simple as modifying the contents of
ucspi-tcp is not *required* but much of the qmail documentation assumes that you are using it. ucspi-tcp is also written by Dan Berstein (cr.yp.to/ucspi-tcp.html)
What can you do with sendmail that you can't to with qmail? There is a a very large set of mature additions and patches to qmail that permit just about anything you may wish to undertake with your mail server.
On the point of qmail being cumbersome: I disagree - what could be simpler than adding a single line to your rcpthosts file? Maintaining qmail is trivial. However, I'll agree that the author's terse documentation makes it seem quite foreign but compared to sendmail it is positively didactic. There are also many other resources available which supplement the original docs.
Qmail is rock-solid. The best proof I can offer is that fact that no security flaw has been found since 1.03 was released in 1998. The man is a cryptographer and designed it for security.
There is also an enormous amount of support for the product available. Check out qmail.org and cr.yp.to/qmail.html
At the two extremes, there are idiot users and idiot tech support staff. However, the majority of support calls fall somewhere in between where the hurdle to overcome is not knowledge but being able to communicate using the same language. I find that being able to "talk da lingo" is immeasurably helpful.
For example, I have successfully been able to get routed directly to technicians in the CO merely because I have been able to convince the Bell Canada staff that I know what I am talking about.
This leads me to believe that end-users have a certain responsibility to acquire a modicum of background understanding. There will always be incompetent support staff but if an end-user can accurately describe their problem, their chances of getting satisfactory help increase dramatically.
Walkmans were around US$400 when first released.