My wife's tablet is less comfortable when sitting on the couch than my laptop. However, the tablet is much easier to read from when laying down or walking.
All of my datapoints (all 3, that is) for VSS were multi-platform or multi-site. I guess if you have less than half the flexibility cvs, the interface will be easier.;-) FWIW, I've never actually seen or used VSS.
I've been watching subversion. Are you using it for critical data? If not, would you feel confident recommending it for critical data? Don't worry, I won't hunt you down if I decided to use subversion and have problems -- after all, you're anonymous. =-)
I don't think they'd keep Visual Sourcesafe if they had access to Clearcase. I've met some anti-open-source-and-Free-software people (and regular people, too) that use cvs because Visual Sourcesafe is so bad.
That said, I'm not real impressed by Clearcase either. But I've never heard of it being so bad that users preferred cvs.
I liked the short answers. They reminded me of Shatner's delivery during dramatic moments. Can't you imagine him sitting down and thinking, in his trademark stacatto speech,
I think the whole premise of XXX was to lighten the genre. So many things in the film were completely crazy. Like the solar-powered jet-propelled hydrofoil/submarine. That's gotta be the best bad idea I've ever heard of.
My hope for ISS is that it is well-managed by the time it is completed, and well-enough managed until then that it isn't abandoned. Unfortunately, all of my (current) research can be done on earth, and it's unlikely they'll ask me to serve as a mission specialist any time soon.;-)
As far as military in space goes, I'm hoping we'll have a stronger Corp of Engineers presence than Airborne presence. No offense intended to Airborne personell.
While writing that, I wondered how people would take it. What I had in mind is that NASA is trying to be (more-or-less) patient. I believe it is more important to move together than to move as fast as possible.
And when people do distinguish between Socialism and Communism, why don't they distinguish between Marx, Lenin, Stalin, Tito, Ceausescu, Mao Zedong and Zemin?
Every comment I've read yet examines the game design and execution to determine why games fail. I expect that this is only 50% of the story. I believe the other half comes from the publication structure in the game industry.
I am told it is hugely impractical for a (regular?) game compnay to finance its own games. This is partly because of the crazy amounts of Hollywood-style glitz and polishing that the market pays for these days. The result is that game companies get "loans" from game publishers like Activision or Electronic Arts to complete the games.
At this point, the publisher is more-or-less in control. The publisher can cancel the game or change its budget. If the game is released, the game company has to pay back the publisher. Part of the deal assigns some portion of the game copmany's royalties to publisher. In the end, the game company can have a very successful product but barely break even (remind anyone of recorded music publication, or book publication?).
And that previous paragraph described a "good" situation. Imagine that the game company has crappy management and doesn't handle the narrow margins well; that the publisher decides to cancel the project; that the publisher goes bankrupt; that the publisher doesn't effectively market the game. I'm sure there are many more bad scenarios than good.
At least at one point in 2001, NASA estimated America's portion of the ISS cost to be US$100bn (i.e. how much America will have spent when the station is completed). Let's suppose nasa is wrong, and that it is actually triple that, US$300bn.
GW Bush's propsed 2003 military budget is US$378bn, which is something like US$43bn more than last year.
And what do we have to show for our military spending? We successfully (?) bombed Serbia during peace-time. The Pentagon couldn't even protect itself from relatively slow-moving passenger aircraft, even when given a 30 minute warning. We bombed the hell out of Afghanistan, including first aid warehouses and wedding parties, and it appears that terrorist organizations still have the upper hand.
At least with the space station there are many nations *talking* and *cooperating* to at least some extent. That is, ISS does much more to make friends than the B2 stealth bomber does.
Why do we spend so much money to protect ourselves from enemies when making friends is so cheap? I think the ISS is a damn good investment.
I don't think we'll know what the Space Station is for until we're done with it, which won't be for many years. I like to think of our space efforts, in general, as
1) Research Investements 2) Engineering Investments 3) Inspirational Exploration 4) Inspirational Art 5) Occasionally Profitable
and for the space station in particular,
6) The one place Americans have restrained themselves and not taken "unilateral action".
1) Most people happy with Windows don't care much about crashes.
2) Most people happy with GNU/Linux don't care much about installation or desktops.
I know which compromise I'd rather make. =-)
The above summary is woefully inadequate, of course. There are things like functionality, application support, customer support, performance, etc. to consider. All the same, I think the above points provide a useful contrast.
I haven't used Windows since about 1999, and have better stomach health because of it; so I can't comment from personal experience. But from what I hear from my wife, a software developer, Win2K still costs her a lot of time in reboots and lockups. Note that she is using almost exclusively Microsoft software and certified drivers at work (in fact, all might be), but she still has lots of problems. The machine appears to be in fine condition.
But her machine crashes less than it did with Nt and certain combinations of service packs! Woohoo!
That last bit was sarcasm. And it might not be true.
Why do people (including Microsoft) always feel the need to point out just how much better the current(-ish) version is than the old one? In one year, you'll still say "There's your problem right there, Windows 2003 is much more stable than Windows 2000".
If someone shot you last week, but only stabbed you this week, you don't have to thank them.
You may not believe me when I say this, but I'm really glad you showed me that constant. I try not to be outwardly pedantic, because it puts people off (or maybe it's my bad breath). But I shudder when I have to hardcode something like 8 bits per byte.
-Paul Komarek
[Apologies to Monty Python for stealing their punch line]
Any idea how DOC or other solid-state IDE interfaces compare?
Not particularly relevant (hence I unchecked the +1 bonus =-), but one of Cray's old vector machines had static ram for main memory with an incredible fabric between ram and cpus. If I remember correctly, it was a "butterfly" network (5 years after I learned about it, I don't much remember it anymore =-).
Good to hear about the wall-brick PSUs for Mini-ITX systems; that TK fanless ATX psu is expensive. Also good about the net-booting.
I would definitely miss the serial bios console. I say this as an admin of 7 headless macines (4 Alpha boxes, 2 x86, one net4501). It is a real pain to drag monitors around to back of racks when you need to fiddle with them. Not that the Soekris has it, but remote management consoles rock; even if the kernel wedges, you can still power cycle remotely.
I'm really hoping you're going to reply with "such-and-such Mini-ITX board has a serial bios". If you tell me it has a remote management console, I'll know you're lying. =-)
Let's do a comparison, then:
Gateway, Router, Wirless Access Point, Firewall with DMZ iface (which is precisely why I bought a board with three ifaces):
$232+shipping: Soekris net4501 with case and psu $60+shipping: D-Link DWL-520 ------- About $300.
I'm using nfs root right now, but might use an 8MB cf card soon. I have two that came free with digital cameras, but new ones seem to be about $4.
How much would a similar 3-iface Mini-ITX system cost? I'm not trying to prove anything here, just trying to find out if I should have waited before buying the Soekris setup for $300.
Both previous courts disagree with you. In fact, the appealate court was unanimous that MSFT is a monoplist. You're probably not considering the exact same market as the courts were, which includes a certain time period.
My "new technology" comment was definitely off-base. I really meant that the C3 was very new.
I'm not sure how to interpret the cpu power draw. I really want to know about system power, which is what I was measuring on the Soekris (with a regular VOM, not a power meter).
I agree that many applications want "PC" ports. The Soekris isn't being marketed or sold into those markets. By the same token, the Mini-ITX C3 stuff is tremendous overkill for VPN WAN routers.
The Mini-ITX stuff I've seen (not much, and only on the web) doesn't seem to have bios-over-serial and built-in net-booting, which are very nice (if not essential) for "embedded" operation. I noted that VIA's suggested power supplies either had fans, except for that TK fanless ATX supply (which costs $150). I'm using a wall brick for my net4501.
I guess what I'm trying to say is that the price, power, and engineering considerations for a Mini-ITX system *must* include the net-booting eth card and other "accessories", the power supply that probably has moving parts, and the additional heat output.
I think I may use a C3 system when I get around to my fanless mailserver and Zope box, and I'll be willing to pay for all the extra equipment (including that TK supply). But the C3 isn't needed for serving nfs at home or acting as a DSL firewall. Since I want my gateway, nfs server, and web server to run on separate cpus (admin and security reasons), I'm trying to by the least expensive box for each task. The net4501 ended up an excellent choice for my network tasks, including gateway, router, wireless router, and firewall.
I help run a small academic lab of Alphas (and one lonely x86 =-) for my research group. We see tremendously bursty usage patterns. I wouldn't be surprised if we were worse than 50% idle on average. These machines are also used by some folks off-campus and one other department on campus.
Few jobs take weeks, some take days, and most take a few hours. But when conference deadlines hit, we're slammed. Every cpu running full boar all the time. For these usage patterns renting time on a supercomputer might be very economical, even though our jobs are small-ish. Not only wouldn't we have to purchase expensive 64-bit big-memory machines that are doing something less than 50% of the time, but we wouldn't have to maintain them either.
It seems unlikely that IBM, the one IT company whose stock held steady (actually grew slightly) after the.com bubble burst, is on its way out. IBM is a *very* smart company, with a lot of existing customers. It seems unlikely that they'd offer metered CPU hours now unless their current customers asked for it.
IBM may be big and slightly slow, but as a company it is smart and sees far. Their finger is on a lot of (customer) pulses.
Slashdot Mirror
My wife's tablet is less comfortable when sitting on the couch than my laptop. However, the tablet is much easier to read from when laying down or walking.
-Paul Komarek
All of my datapoints (all 3, that is) for VSS were multi-platform or multi-site. I guess if you have less than half the flexibility cvs, the interface will be easier. ;-) FWIW, I've never actually seen or used VSS.
-Paul Komarek
I've been watching subversion. Are you using it for critical data? If not, would you feel confident recommending it for critical data? Don't worry, I won't hunt you down if I decided to use subversion and have problems -- after all, you're anonymous. =-)
-Paul Komarek
I don't think they'd keep Visual Sourcesafe if they had access to Clearcase. I've met some anti-open-source-and-Free-software people (and regular people, too) that use cvs because Visual Sourcesafe is so bad.
That said, I'm not real impressed by Clearcase either. But I've never heard of it being so bad that users preferred cvs.
-Paul Komarek
I liked the short answers. They reminded me of Shatner's delivery during dramatic moments. Can't you imagine him sitting down and thinking, in his trademark stacatto speech,
"I must...complete...this interview."
-Paul Komarek
I think the whole premise of XXX was to lighten the genre. So many things in the film were completely crazy. Like the solar-powered jet-propelled hydrofoil/submarine. That's gotta be the best bad idea I've ever heard of.
-Paul Komarek
The B-52 is also equipped with a sextant and other astral navigation gear. Part of the reason is survivability during a nuclear war.
-Paul
Besides, LORAN stations have super-cool strobes on their four antennas. In a fog, those strobes double their coolness.
-Paul Komarek
My hope for ISS is that it is well-managed by the time it is completed, and well-enough managed until then that it isn't abandoned. Unfortunately, all of my (current) research can be done on earth, and it's unlikely they'll ask me to serve as a mission specialist any time soon. ;-)
As far as military in space goes, I'm hoping we'll have a stronger Corp of Engineers presence than Airborne presence. No offense intended to Airborne personell.
-Paul Komarek
While writing that, I wondered how people would take it. What I had in mind is that NASA is trying to be (more-or-less) patient. I believe it is more important to move together than to move as fast as possible.
-Paul Komarek
And when people do distinguish between Socialism and Communism, why don't they distinguish between Marx, Lenin, Stalin, Tito, Ceausescu, Mao Zedong and Zemin?
-Paul Komarek
Every comment I've read yet examines the game design and execution to determine why games fail. I expect that this is only 50% of the story. I believe the other half comes from the publication structure in the game industry.
I am told it is hugely impractical for a (regular?) game compnay to finance its own games. This is partly because of the crazy amounts of Hollywood-style glitz and polishing that the market pays for these days. The result is that game companies get "loans" from game publishers like Activision or Electronic Arts to complete the games.
At this point, the publisher is more-or-less in control. The publisher can cancel the game or change its budget. If the game is released, the game company has to pay back the publisher. Part of the deal assigns some portion of the game copmany's royalties to publisher. In the end, the game company can have a very successful product but barely break even (remind anyone of recorded music publication, or book publication?).
And that previous paragraph described a "good" situation. Imagine that the game company has crappy management and doesn't handle the narrow margins well; that the publisher decides to cancel the project; that the publisher goes bankrupt; that the publisher doesn't effectively market the game. I'm sure there are many more bad scenarios than good.
-Paul Komarek
At least at one point in 2001, NASA estimated America's portion of the ISS cost to be US$100bn (i.e. how much America will have spent when the station is completed). Let's suppose nasa is wrong, and that it is actually triple that, US$300bn.
GW Bush's propsed 2003 military budget is US$378bn, which is something like US$43bn more than last year.
And what do we have to show for our military spending? We successfully (?) bombed Serbia during peace-time. The Pentagon couldn't even protect itself from relatively slow-moving passenger aircraft, even when given a 30 minute warning. We bombed the hell out of Afghanistan, including first aid warehouses and wedding parties, and it appears that terrorist organizations still have the upper hand.
At least with the space station there are many nations *talking* and *cooperating* to at least some extent. That is, ISS does much more to make friends than the B2 stealth bomber does.
Why do we spend so much money to protect ourselves from enemies when making friends is so cheap? I think the ISS is a damn good investment.
-Paul Komarek
I don't think we'll know what the Space Station is for until we're done with it, which won't be for many years. I like to think of our space efforts, in general, as
1) Research Investements
2) Engineering Investments
3) Inspirational Exploration
4) Inspirational Art
5) Occasionally Profitable
and for the space station in particular,
6) The one place Americans have restrained themselves and not taken "unilateral action".
-Paul Komarek
I think the take-home part of the lesson is:
1) Most people happy with Windows don't care much about crashes.
2) Most people happy with GNU/Linux don't care much about installation or desktops.
I know which compromise I'd rather make. =-)
The above summary is woefully inadequate, of course. There are things like functionality, application support, customer support, performance, etc. to consider. All the same, I think the above points provide a useful contrast.
-Paul Komarek
I haven't used Windows since about 1999, and have better stomach health because of it; so I can't comment from personal experience. But from what I hear from my wife, a software developer, Win2K still costs her a lot of time in reboots and lockups. Note that she is using almost exclusively Microsoft software and certified drivers at work (in fact, all might be), but she still has lots of problems. The machine appears to be in fine condition.
But her machine crashes less than it did with Nt and certain combinations of service packs! Woohoo!
That last bit was sarcasm. And it might not be true.
-Paul Komarek
Why do people (including Microsoft) always feel the need to point out just how much better the current(-ish) version is than the old one? In one year, you'll still say "There's your problem right there, Windows 2003 is much more stable than Windows 2000".
If someone shot you last week, but only stabbed you this week, you don't have to thank them.
-Paul Komarek
gcc and Compaq's cc on Alpha automatically use 64 bit pointers and 32 bit ints. Perhaps this is because the Alpha has always been a 64 bit cpu?
-Paul Komarek
You may not believe me when I say this, but I'm really glad you showed me that constant. I try not to be outwardly pedantic, because it puts people off (or maybe it's my bad breath). But I shudder when I have to hardcode something like 8 bits per byte.
-Paul Komarek
[Apologies to Monty Python for stealing their punch line]
Any idea how DOC or other solid-state IDE interfaces compare?
Not particularly relevant (hence I unchecked the +1 bonus =-), but one of Cray's old vector machines had static ram for main memory with an incredible fabric between ram and cpus. If I remember correctly, it was a "butterfly" network (5 years after I learned about it, I don't much remember it anymore =-).
-Paul Komarek
Good to hear about the wall-brick PSUs for Mini-ITX systems; that TK fanless ATX psu is expensive. Also good about the net-booting.
I would definitely miss the serial bios console. I say this as an admin of 7 headless macines (4 Alpha boxes, 2 x86, one net4501). It is a real pain to drag monitors around to back of racks when you need to fiddle with them. Not that the Soekris has it, but remote management consoles rock; even if the kernel wedges, you can still power cycle remotely.
I'm really hoping you're going to reply with "such-and-such Mini-ITX board has a serial bios". If you tell me it has a remote management console, I'll know you're lying. =-)
Let's do a comparison, then:
Gateway, Router, Wirless Access Point, Firewall with DMZ iface (which is precisely why I bought a board with three ifaces):
$232+shipping: Soekris net4501 with case and psu
$60+shipping: D-Link DWL-520
-------
About $300.
I'm using nfs root right now, but might use an 8MB cf card soon. I have two that came free with digital cameras, but new ones seem to be about $4.
How much would a similar 3-iface Mini-ITX system cost? I'm not trying to prove anything here, just trying to find out if I should have waited before buying the Soekris setup for $300.
-Paul Komarek.
"No. MSFT isn't a monopoly".
Both previous courts disagree with you. In fact, the appealate court was unanimous that MSFT is a monoplist. You're probably not considering the exact same market as the courts were, which includes a certain time period.
-Paul Komarek
My "new technology" comment was definitely off-base. I really meant that the C3 was very new.
I'm not sure how to interpret the cpu power draw. I really want to know about system power, which is what I was measuring on the Soekris (with a regular VOM, not a power meter).
I agree that many applications want "PC" ports. The Soekris isn't being marketed or sold into those markets. By the same token, the Mini-ITX C3 stuff is tremendous overkill for VPN WAN routers.
The Mini-ITX stuff I've seen (not much, and only on the web) doesn't seem to have bios-over-serial and built-in net-booting, which are very nice (if not essential) for "embedded" operation.
I noted that VIA's suggested power supplies either had fans, except for that TK fanless ATX supply (which costs $150). I'm using a wall brick for my net4501.
I guess what I'm trying to say is that the price, power, and engineering considerations for a Mini-ITX system *must* include the net-booting eth card and other "accessories", the power supply that probably has moving parts, and the additional heat output.
I think I may use a C3 system when I get around to my fanless mailserver and Zope box, and I'll be willing to pay for all the extra equipment (including that TK supply). But the C3 isn't needed for serving nfs at home or acting as a DSL firewall. Since I want my gateway, nfs server, and web server to run on separate cpus (admin and security reasons), I'm trying to by the least expensive box for each task. The net4501 ended up an excellent choice for my network tasks, including gateway, router, wireless router, and firewall.
We're just thinking of different applications.
-Paul Komarek
I help run a small academic lab of Alphas (and one lonely x86 =-) for my research group. We see tremendously bursty usage patterns. I wouldn't be surprised if we were worse than 50% idle on average. These machines are also used by some folks off-campus and one other department on campus.
Few jobs take weeks, some take days, and most take a few hours. But when conference deadlines hit, we're slammed. Every cpu running full boar all the time. For these usage patterns renting time on a supercomputer might be very economical, even though our jobs are small-ish. Not only wouldn't we have to purchase expensive 64-bit big-memory machines that are doing something less than 50% of the time, but we wouldn't have to maintain them either.
-Paul Komarek
It seems unlikely that IBM, the one IT company whose stock held steady (actually grew slightly) after the .com bubble burst, is on its way out. IBM is a *very* smart company, with a lot of existing customers. It seems unlikely that they'd offer metered CPU hours now unless their current customers asked for it.
IBM may be big and slightly slow, but as a company it is smart and sees far. Their finger is on a lot of (customer) pulses.
-Paul Komarek