The issue with "breaking the sound barrier" was largely an engineering problem, not a theoretical one.
Bullets, for example, had been supersonic for quite some time.
The trick was learning to design an _airfoil_ that could provide sufficiant lift at speed ranges that would allow subsonic takeoff/landing but yet still allow controlled supersonic flight.
It's all about the behaviour of air at high speeds - its material properties, if you will. But there was very little physics going on.
This is vastly different from the FTLiaV problem, where it appears the very nature of the universe conspires against you.
DG
Re:So what kind of stickers...
on
dB Drag Racing
·
· Score: 2
Damn, forgot this in the parent post. SlashDot needs an "edit" feature...
The Linux game client binary on the SoU expansion CD is V1.30, where the game client on their website is V1.29. So the SoU version is NEWER than this one.
I can confirm that the "sticky mouse" problem that showed up on some resolutions is fixed with V1.30. I actually finished the game at 800X600 with 1.29B5. I tried running at 1024X768 with B5, and got the "sticky mouse". Tried it again with 1.30 after I installed SoU, and it worked - although I seem to have hit the bandwidth limits of a PCI-based GeForce MX400, 'cause it was a little slow.
At 800X600, Athlon 2100+, RH8, latest NVidia drivers, 32Mb textures, game was nice and snappy.
I expect that with a more modern, AGP-based card, the game would scale better to the higher resolutions.
I didn't get very deep into the SoU single-player campaign last night, but the little bit I did shows that it is MUCH improved from the original. The BioWare module people are getting very good. Lots of nice little touches. This is a game worth picking up.
The expansion pack Shadows of Urdentide even ships with a Linux installer on the CD
Almost...
It seems that the CD mastering tool BioWare/Atari used converted all the text files - that includes shell scripts, mind - to Windows-style text, and when you try to run the installer/bin/sh chokes on all the ^M characters.
Note to all single-player-campaign people - pick up the SoU expansion pack and install it concurrently with NWN. The expansion pack adds many more spells, classes, feats etc and they work with the original game, plus some minor bugs are fixed in the process (the SoU expansion patches the original game content too)
Huh, I submitted this as a story this AM, and it was rejected in favour of this. Go figure.
That would be fine and dandy if a Microsoft product was like a physical widget, but it's not.
Can you get support from Microsoft for MS-DOS? Windows 3.1? Win 95? If you want to buy 300 more licences for Windows NT, or for Word 97, can you get it? If you want Win98 but with that pesky security hole patched, can it be had for ANY price?
The Microsoft forced upgrade cycle means that the stuff "self-breaks" every few years ANYWAY. So which is worse - converting entire agencies to Windows Whatever from Win95, or converting the same agency from Win 95 to Linux?
You're going to have to do *A* conversion eventually; you might as well get onto the one that doesn't require you to pay licences and which can be upgraded or bugfixed piecemeal.
Well, not *right this second* but it's been eating up all of my spare time the last couple of weeks.
7th/7th sorcerer/ranger, currently on the second level of the "Ruins of Illusk" under Luskan, chapter 2. Whoohoo!:)
And as for installing it... I just bought the game, DLed the latest NWN client from Bioware, DLed that 3rd party installer, ran the installer, unpacked the client, and started playing - just that easy.
I guaren-damn-tee you that had there been any reasonable alternative to human-wave attacks at the time, that the Chinese generals would have been all over that.
But war is an ultimate expression of policy. You do whatever you have to do to win. You make use of whatever resources you have, no matter how unpleasent.
A Chinese-style human wave is not indicitive of a lack of regard for human life - it is instead an indicator of how badly they wanted to win.
My guess is that all that bravada about being able to take on IBM's legal team in an extended fight, plus Novell as well, while making all your money selling alternatives to a FREE OS (that by your own admission is a drop-in replacement for your product) doesn't spell investor confidence.
They intend to increase these licences as they "agressively protect our intellectual property rights"
1/3rd of their projected earnings is from licencing, the rest is from OS products.
Lots of rah-rah "our best quarter ever" "we have eliminated all long-term debt"
The core business is still selling SCO UNIX
(which as an aside, means that you really cannot afford a wave of consumer hostility, can you?)
Looking to provide "subscription-based" services.
They have cash in the bank.
Their prime focus is defending themselves against Linux.
A nod towards the Novell thing - "SCO owns the UNIX contract rights" not the UNIX trademarks or IP.
A list of customers is presented... British Army just bought an asset/lifecycle management system for helicopters based on SCO UNIX
"SCO Source" is their new shakedown unit. 2 big licencees - one is secret, the other is Microsoft. Microsoft bought a licence for "Microsoft services for UNIX" putting UNIX source into Windows (?)
IBM is NOT the secret licencee.
They claim 3 code teams found violations where their code was in Linux, independantly.
No mention of what code, where.
Ahhhh... something makes sense now. Their Linux sales were tanking hard - they were making more money off SCO UNIX. So they're out to go toe-to-toe with Linux (a big part of which is anti-Linux FUD)
Q&A now starting:
First question mentions Novell release and asks for explicit listing with the code violations. They claim IBM has them (?)
Claim Novell blew them off; that there was a meeting yesterday that Novell skipped.
They refuse to publish the violations publically.
Next question is about how long the SCO Source revenue stream is likely to be. So far, they claim two contracts. No way to quantify how many contracts are in the pipeline.
What is happening Jun 14th re IBM/AIX? They will revoke the AIX licence. Details will follow.
Next question is from a capital company, asking about aquisitions. They want to get into web services, interact with Java and.net.
They claim 2.2 million SCO servers - original business plan was to migrate those to Linux, and they lost money doing that. So now they want to keep those people on SCOX (at $1500 a server)
Nothing about why those people should stay on SCO when they have Linux readily availible....
Mr Wall Street thinks SCO is going to gain some short-term cash.
They think that their stock price spike reflects market confidence in SCO vs Linux
Another venture capitalist.... wants to know if licence deals will drive SCOX product sales.
They used SCO Source to raise capital instead of going to veture capital.
Next question asks for an estimate on legal costs. So far they are below their legal budget, but they are in for the long haul and are ready for countersuits.
Next caller wants to know why SCO was asking for transfer of UNIX copyrights from Novell and why they were denied. SCO doesn't seem to know what they own. They claim they own the UNIX copyrights, based on a contract review of the Novell contracts. They feel they can go toe-to-toe with Novell and win that fight too.
Next question wants to know if SCO has abandoned Linux sales. They claim lots of people are putting Linux implementation on hold, and that Linux "leadership" is dismissive to SCO IP rights. SCO is probably done with Linux. Caldara upgrade path is SCO UNIX.
Not exactly a job interview, but a similar situation in a military context.
I was on my Armoured Recce Troop Leader's Course, and I was being tested on Patrol Commanding in the Advance.
A Patrol consists of two light armoured vehicles. A student is in charge of each vehicle, and one of them is the Patrol Commander who is in charge. The driver and observer in both vehicles are instructors.
In the advance, you're out looking for the bad guys, so you take turns leapfrogging each other. One vehicle watches while the other vehicle moves forward to a new position of observation. If you encounter obstacles, the bad guys, etc there are a series of drills to carry out.
If you fail the course, you lose your job, no possibility for a do-over. The course was broken down into sections, and each section had a practical exam. You could fail it once. Fail your second crack at it, and you were gone.
So anyway, we're on my exam for Patrol Commanding in the Advance, and the guy assigned as my junior is a complete fuckwit. Couldn't find his own ass in the dark with both hands and a flashlight.
He takes the first bound while I observe, encounters a blind corner, and fucks up the drill. The I take the next bound, leapfrogging him, and that goes OK. He leapfrogs me, encouters some other obstacle, and fucks it up again.
In order to pass the exam, you had to do four bounds without error within a time limit, and by this time, we're starting to get close to the limit and there's only one good bound in the bag.
So I'm looking at him floundering around through my binos, and I realize that I've already failed the exam... but By God I'm not going to let this whole experience go to waste. So I hop out of the callsign, storm forward to his position, drag him out of the vehicle, and tear him a new asshole.
Normally, this Just Isn't Done - students don't yell at other students so that they don't look bad in front of the instructors. But as far as I was concerned, the damage was already done, and if I didn't do something about this numbnuts then the next poor bastard who he is assigned to as junior is going to get screwed too, so I have to sort him out right now.
Once I've finished expressing my displeasure and explaining how he SHOULD have been carrying out his job, I tell him we're going to carry on with the exam until time runs out - but with one major difference. Instead of leapfrogging, we're going to catapillar, and I'm going to take the lead.
This means that he drives forward to my position, and I move forward to find the next one. Lather rince repeat.
It's slower than moving leapfrog, and it exposes me to all the risk because I'm always up front, but it also prevents him from screwing anything up because all he has to do is take up position in the spot I just vacated.
We get two more bounds in and then time runs out.
So I'm being debriefed, and the first thing the instructor asks me is how I think I did. The answer is obvious - not enough bounds done correctly, chewed out another student in the middle of the exam... it's pretty clear to me I've failed.
But he passed me.
Up until the point when I went foreward to have words with junior, I had been failing miserably. Chewing junior out (when it was clear he needed it) got his attention, but didn't necessarily _mean_ anything - anybody can get angry.
Nope, what passed me was taking effective steps to solve the problem, by taking the lead and moving to catapillar movement to ensure I kept it. As soon as I did that, he passed me.
I was told that a leader who can carry out a plan effectively is good to have around, but a leader who can take a plan that is all FUBAR and turn it around is something else again.
There is, however, a rather unfortunate epilogue.
The message that junior got out of this was rather different. The way he understood what had happened was "yelling at subordinates will get you passed" so he spent the rest of the course screaming his head off any time he had command of anything. Icing on the top of a perfectly enjoyable experience.:|
Assume that I produce widgets, and that my profits are tie to widget sales.
So I can:
1) Sell more widgets - this is fine, but bounded. Eventually, EVERYBODY has a widget, and, assuming that widgets have an infinate life, that means my market shrinks with time.
(Of course, one can sidestep this by building a fixed life into widgets - perhaps an ever-decreasing life - but there are problems with this approach too)
2) As you mentioned, get more efficient at widget production, so that the overhead-per-widget falls with time, leaving a greater difference between price of sale and cost of production (ie, increasing profits)
But this is bounded too! Assuming perfect efficiency, eventually you get down to cost of raw material + cost of energy (at 100% efficiency) and can go no further.
So while for any given company, there may well be room within the theroretical bounds of market size and efficiency of production for some pretty sizeable growth, eventually you MUST hit a limiting factor. Then what?
You have set up here a classic Straw Man argument, in which you pose a question, provide a flawed response, and then proceed to tear down the flawed premise as "proof" of your assertation.
A debate tactic as old as the hills, but no more effective today than it was in ancient Greece.
The crux of your straw man is this - the assertation that any attempt at presenting a perpetual motion machine is de facto rejected by "physicists" by virtue of the nomenclature "perpetual motion".
This is in error.
The reason why every single claim to date of the creation of a "perpetual motion" machine has been refuted, is because every single one of them has been deconstructed, analyzed, and the flaw found and demonstrated.
If you (or any other) would-be natural-law revolutionary can ACTUALLY DEMONSTRATE a machine that taps into some yet-undiscovered law of nature, you will be lauded and celebrated, not ridiculed.
The problem is that no machine so presented has ever survived the analysis - and many of them have been outright attempts at FRAUD.
Your argument is invalid at its heart, I'm afraid.
Isn't the stock market, at least the way it is operated these days, more of a scam than an investment?
Perhaps some business major can explain this to me.
As I see it, here's the problem:
1) I buy x shares of stock in SomeCompany, at some price n.
2) I want to sell this stock at some point in the future for some price y > n.
3) In order for the price of the stock to increase, it must be "worth more".
4) Assuming no further stock is issued and that I purchase no additional stock, that means my stock constitutes a fixed percentage "ownership" in the company.
5) Given that my percentage ownership is fixed, in order for the stock to be worth more at some time in the future, the value (read: size) of the company must grow.
Well, that's fine and dandy for a small company, but that seems to assume that any given company will grow without bounds pretty much forever (especially when we start talking about options)
But the "economy" is a bounded system. Those bounds may be high, but they DO exist. So the fundamental premise of "stock price increases with time" seems fundamentally flawed to me. What do you do with a successful company that produces a flat amount of profit each year, and does not seek to grow?
Now DIVIDENDS, THAT I understand. I own some share q of the company, the company produces w amount of profit, so I'm entitled to a dividend w/q - that makes sense.
But this whole speculation thing I just cannot wrap my head around. It looks like a pyramid scheme (with a gentle slope) to me.
(Obligatory Austin Powers "nutshell" joke deleted to save space)
1) A research company did a survey of commercial Linux developers with the aim of learning more about them.
2) They gave the data they collected to Nick, to see if he could glean anything interesting from the data.
3) Nick discovered that, of the commercial Linux developers surveyed:
a) 40% considered Linux to be their "primary focus"
b) Of these 40%, more of them switched their "primary focus" from Windows than from a flavour of UNIX
3) From this, Nick concludes that Linux is stealing more development share from Windows than from UNIX derivatives - which makes inuative sense, because there are a lot more Windows developers out there to steal share from.
4) Nick is satisfied that the research company was doing good science in their survey, rather than trying to spin him.
My own analysis agrees with his conclusions.
It is an interesting conclusion, because, given the API simularity between UNIXen and Linux, one might expect that the majority of commercial Linux developers would have been UNIX developers who ported their product over to Linux, but this seems to not be the case. Developers are in fact "defecting" from Windows to Linux.
What isn't addressed is the reason for the "defection". I for one would not be suprised that the prime motivator was Microsoft eating their lunch - not much point producing a product if M$ baked it into Windows....
A quick aside: I HATE the term "scripting", as if it were some degenerate form of "real programming" - especially with feature-rich languages like perl that never have to call other applications.
Anyway, first-hand experience: thanks to the concept of perl modules and the incredible CPAN archive, writing applications that have to go to the network for things like HTTP or (especially) LDAP are trivial in perl but seriously heavy lifting in C.
You also get string parsing, regular expressions, and garbage collection built right in. Not to mention the incredibly powerful (from a code legibility standpoint) associative array or "hash" data structure.
Believe it or not, correctly written perl is orders of magnitude more legible than C or Java, because it works at a higher level of abstraction.
I wrote an LDAP->LDAP replication program, with schema and data format translation, in a couple of hours using perl.
Doing stuff like comparing the contents of a database dump (provided as a CSV) against an LDAP directory is trivial in perl.
C is best used when you won't have a perl environment availible and need the binary to stand alone. For pretty much every other task I've encountered in the last 6 years, perl got the job done faster and with much better maintainability.
They've got 2 engine programs now - hell, more than 2. The restrictor plate engines have nothing in common with a short track engine save the basic parts.
Running a completely different, smaller, unrestricted engine at the superspeedways would be far cheaper than the restrictor engines.
Before you go getting all bent out of shape over restrictor plates and safety, consider the following:
1) Daytona and Talledega are LONG
2) Daytona and Talledega are HIGHLY BANKED
That combination of the length of the track and especially the high banking (which provides gobs of extra cornering force) means that the cars can sustain astronomical top speeds without needing major revolutions in tire technology or wing-and-undertray levels of downforce. It's the banking that lets 'em run flat out.
NASCAR was running over 200 MPH at Daytona in the 60's, back when the cars really were production based and had stones for tires. With modern (for NASCAR) tires and suspensions, that banking could probably support speeds in excess of 260 MPH before the cars got cornering-force limited and had to slow down on corner entry.
Now with the frontal area that they have, no NASCAR car is going to be turning 260 with even unrestricted engines. The power consumed by aero drag is a function of the square of the speed, so it takes more power for the same delta v the faster you go. There's a limit to how much power you can squeeze out of even an unrestricted motor, so the real top speed would probably be somewhere in the 235 area.
But note that the guy who makes 5 HP more than his neighbor is only going to make a small fraction of a MPH more in terminal velocity.
So guess what pulling the restrictor plates off did? You get the EXACT SAME scenario as you had with the plates on, except now the speeds are 30-50 MPH faster. And kinetic energy (that must be dissipated in a crash) is a function of the square of velocity squared as well....
As bad as a Big Wreck at a buck ninety is, that pales in comparision to the same wreck at 230. And these aren't 1500lb Champ cars, these are 3600lb locomotives.
The problem with restrictor plates isn't that they cause the tight grouping of cars and the inability to pass unassisted - that's the fault of the banking. The big issue with the restrictor plate is that it takes a tremendous amount of engineering to try and coax extra air through that plate, and to get the engine to run in the odd environment the plate creates in the intake manifold. R&D costs for a 'plate engine run easily 10 times higher than a short track motor.
What NASCAR should do is make the actual engine displacement for the superspeedways smaller. Make 'em run a 3 litre V6. That'd bring costs way down while still preserving the safety.
http://www.aceshardware.com/read.jsp?id=50000363...and know not only what track you're on, but that the car is in a position (in the braking area leading up to the chicane, on the wrong side of the track, with a couple of degrees of steering lock in the wrong direction) to crash in the next heartbeat.
I've been coding in an enterprise environment for quite some time now, and I have one rule that is cast in gold:
Always optimise source code for legibility above all else. Never trade legibility for performance unless you have no other choice, and then document your cleverness in the code so that those who follow behind you can keep up.
Here's why:
When you first write a system, it will spend its first few months of life in a very intensive quality control feedback loop. Bugs are found and very quickly exterminated. The code is still fresh in your mind and you're "in the zone".
But as the system stabilises, there is less and less reason to go back to the code, so that freshness wears off. After a little while, other priorities will take over and the internal model of the code will fade away.
But there's still bugs in there - there always is. But any bug that makes it past the first few months is non-obvious, intermittant, rare, and so on (thus, harder to find)
When one finally surfaces, _somebody_ is going to have to fix it. Sometimes it will be you, and you will appreciate code legibilty when you have to dust off source that has laid untouched for years. Not only does it increase the probability that you'll be able to actually find the bug, it cuts down on the time needed to fix it.
There's nothing like being the guy who finds and fixes bugs within seconds of them being pointed out to enhance your reputation.
But more often than not, it will be some other poor sap who gets saddled with your code and a deadline to get it fixed - and the guy who draws the short straw is normally not the biggest brain in the shop. There is no gratitude like the gratitude from someone forced to dive into somebody else's code, and who subsequently discovers that you have gone out of your way to make it easier for them to understand.
This is _also_ a reputation enhancer. "That code was so well written that not only did it take no time at all to track down the bug, but I also learned a couple of new techniques in the process!"
The true guru is a TEACHER.
Oh, and ALWAYS check the return code from every system call and provide appropriate error trapping. That's good too.
Slashdot Mirror
Ahhh, no.
The issue with "breaking the sound barrier" was largely an engineering problem, not a theoretical one.
Bullets, for example, had been supersonic for quite some time.
The trick was learning to design an _airfoil_ that could provide sufficiant lift at speed ranges that would allow subsonic takeoff/landing but yet still allow controlled supersonic flight.
It's all about the behaviour of air at high speeds - its material properties, if you will. But there was very little physics going on.
This is vastly different from the FTLiaV problem, where it appears the very nature of the universe conspires against you.
DG
If you can't be fast, be pretty.
If you can't be pretty, be loud.
DG
Damn, forgot this in the parent post. SlashDot needs an "edit" feature...
The Linux game client binary on the SoU expansion CD is V1.30, where the game client on their website is V1.29. So the SoU version is NEWER than this one.
I can confirm that the "sticky mouse" problem that showed up on some resolutions is fixed with V1.30. I actually finished the game at 800X600 with 1.29B5. I tried running at 1024X768 with B5, and got the "sticky mouse". Tried it again with 1.30 after I installed SoU, and it worked - although I seem to have hit the bandwidth limits of a PCI-based GeForce MX400, 'cause it was a little slow.
At 800X600, Athlon 2100+, RH8, latest NVidia drivers, 32Mb textures, game was nice and snappy.
I expect that with a more modern, AGP-based card, the game would scale better to the higher resolutions.
I didn't get very deep into the SoU single-player campaign last night, but the little bit I did shows that it is MUCH improved from the original. The BioWare module people are getting very good. Lots of nice little touches. This is a game worth picking up.
DG
The expansion pack Shadows of Urdentide even ships with a Linux installer on the CD
Almost...
It seems that the CD mastering tool BioWare/Atari used converted all the text files - that includes shell scripts, mind - to Windows-style text, and when you try to run the installer /bin/sh chokes on all the ^M characters.
Happily, there is a workaround. See HERE
Note to all single-player-campaign people - pick up the SoU expansion pack and install it concurrently with NWN. The expansion pack adds many more spells, classes, feats etc and they work with the original game, plus some minor bugs are fixed in the process (the SoU expansion patches the original game content too)
Huh, I submitted this as a story this AM, and it was rejected in favour of this. Go figure.
DG
"If it ain't broke, don't fix it"
That would be fine and dandy if a Microsoft product was like a physical widget, but it's not.
Can you get support from Microsoft for MS-DOS? Windows 3.1? Win 95? If you want to buy 300 more licences for Windows NT, or for Word 97, can you get it? If you want Win98 but with that pesky security hole patched, can it be had for ANY price?
The Microsoft forced upgrade cycle means that the stuff "self-breaks" every few years ANYWAY. So which is worse - converting entire agencies to Windows Whatever from Win95, or converting the same agency from Win 95 to Linux?
You're going to have to do *A* conversion eventually; you might as well get onto the one that doesn't require you to pay licences and which can be upgraded or bugfixed piecemeal.
DG
Well, not *right this second* but it's been eating up all of my spare time the last couple of weeks.
:)
7th/7th sorcerer/ranger, currently on the second level of the "Ruins of Illusk" under Luskan, chapter 2. Whoohoo!
And as for installing it... I just bought the game, DLed the latest NWN client from Bioware, DLed that 3rd party installer, ran the installer, unpacked the client, and started playing - just that easy.
So I don't see a "fiasco" here.
DG
I guaren-damn-tee you that had there been any reasonable alternative to human-wave attacks at the time, that the Chinese generals would have been all over that.
But war is an ultimate expression of policy. You do whatever you have to do to win. You make use of whatever resources you have, no matter how unpleasent.
A Chinese-style human wave is not indicitive of a lack of regard for human life - it is instead an indicator of how badly they wanted to win.
DG
Well, it seems that a few people ain't buying it.
Today's trading has their stock down almost 30%.
My guess is that all that bravada about being able to take on IBM's legal team in an extended fight, plus Novell as well, while making all your money selling alternatives to a FREE OS (that by your own admission is a drop-in replacement for your product) doesn't spell investor confidence.
Bye-bye SCO.
DG
I'm on now - and it has started.
.net.
8.3 million in revenue from SCO Source licencing.
6000 licences.
They intend to increase these licences as they "agressively protect our intellectual property rights"
1/3rd of their projected earnings is from licencing, the rest is from OS products.
Lots of rah-rah "our best quarter ever" "we have eliminated all long-term debt"
The core business is still selling SCO UNIX
(which as an aside, means that you really cannot afford a wave of consumer hostility, can you?)
Looking to provide "subscription-based" services.
They have cash in the bank.
Their prime focus is defending themselves against Linux.
A nod towards the Novell thing - "SCO owns the UNIX contract rights" not the UNIX trademarks or IP.
A list of customers is presented... British Army just bought an asset/lifecycle management system for helicopters based on SCO UNIX
"SCO Source" is their new shakedown unit. 2 big licencees - one is secret, the other is Microsoft. Microsoft bought a licence for "Microsoft services for UNIX" putting UNIX source into Windows (?)
IBM is NOT the secret licencee.
They claim 3 code teams found violations where their code was in Linux, independantly.
No mention of what code, where.
Ahhhh... something makes sense now. Their Linux sales were tanking hard - they were making more money off SCO UNIX. So they're out to go toe-to-toe with Linux (a big part of which is anti-Linux FUD)
Q&A now starting:
First question mentions Novell release and asks for explicit listing with the code violations. They claim IBM has them (?)
Claim Novell blew them off; that there was a meeting yesterday that Novell skipped.
They refuse to publish the violations publically.
Next question is about how long the SCO Source revenue stream is likely to be. So far, they claim two contracts. No way to quantify how many contracts are in the pipeline.
What is happening Jun 14th re IBM/AIX? They will revoke the AIX licence. Details will follow.
Next question is from a capital company, asking about aquisitions. They want to get into web services, interact with Java and
They claim 2.2 million SCO servers - original business plan was to migrate those to Linux, and they lost money doing that. So now they want to keep those people on SCOX (at $1500 a server)
Nothing about why those people should stay on SCO when they have Linux readily availible....
Mr Wall Street thinks SCO is going to gain some short-term cash.
They think that their stock price spike reflects market confidence in SCO vs Linux
Another venture capitalist.... wants to know if licence deals will drive SCOX product sales.
They used SCO Source to raise capital instead of going to veture capital.
Next question asks for an estimate on legal costs. So far they are below their legal budget, but they are in for the long haul and are ready for countersuits.
Next caller wants to know why SCO was asking for transfer of UNIX copyrights from Novell and why they were denied. SCO doesn't seem to know what they own. They claim they own the UNIX copyrights, based on a contract review of the Novell contracts. They feel they can go toe-to-toe with Novell and win that fight too.
Next question wants to know if SCO has abandoned Linux sales. They claim lots of people are putting Linux implementation on hold, and that Linux "leadership" is dismissive to SCO IP rights. SCO is probably done with Linux. Caldara upgrade path is SCO UNIX.
That's it.
DG
Not exactly a job interview, but a similar situation in a military context.
:|
I was on my Armoured Recce Troop Leader's Course, and I was being tested on Patrol Commanding in the Advance.
A Patrol consists of two light armoured vehicles. A student is in charge of each vehicle, and one of them is the Patrol Commander who is in charge. The driver and observer in both vehicles are instructors.
In the advance, you're out looking for the bad guys, so you take turns leapfrogging each other. One vehicle watches while the other vehicle moves forward to a new position of observation. If you encounter obstacles, the bad guys, etc there are a series of drills to carry out.
If you fail the course, you lose your job, no possibility for a do-over. The course was broken down into sections, and each section had a practical exam. You could fail it once. Fail your second crack at it, and you were gone.
So anyway, we're on my exam for Patrol Commanding in the Advance, and the guy assigned as my junior is a complete fuckwit. Couldn't find his own ass in the dark with both hands and a flashlight.
He takes the first bound while I observe, encounters a blind corner, and fucks up the drill. The I take the next bound, leapfrogging him, and that goes OK. He leapfrogs me, encouters some other obstacle, and fucks it up again.
In order to pass the exam, you had to do four bounds without error within a time limit, and by this time, we're starting to get close to the limit and there's only one good bound in the bag.
So I'm looking at him floundering around through my binos, and I realize that I've already failed the exam... but By God I'm not going to let this whole experience go to waste. So I hop out of the callsign, storm forward to his position, drag him out of the vehicle, and tear him a new asshole.
Normally, this Just Isn't Done - students don't yell at other students so that they don't look bad in front of the instructors. But as far as I was concerned, the damage was already done, and if I didn't do something about this numbnuts then the next poor bastard who he is assigned to as junior is going to get screwed too, so I have to sort him out right now.
Once I've finished expressing my displeasure and explaining how he SHOULD have been carrying out his job, I tell him we're going to carry on with the exam until time runs out - but with one major difference. Instead of leapfrogging, we're going to catapillar, and I'm going to take the lead.
This means that he drives forward to my position, and I move forward to find the next one. Lather rince repeat.
It's slower than moving leapfrog, and it exposes me to all the risk because I'm always up front, but it also prevents him from screwing anything up because all he has to do is take up position in the spot I just vacated.
We get two more bounds in and then time runs out.
So I'm being debriefed, and the first thing the instructor asks me is how I think I did. The answer is obvious - not enough bounds done correctly, chewed out another student in the middle of the exam... it's pretty clear to me I've failed.
But he passed me.
Up until the point when I went foreward to have words with junior, I had been failing miserably. Chewing junior out (when it was clear he needed it) got his attention, but didn't necessarily _mean_ anything - anybody can get angry.
Nope, what passed me was taking effective steps to solve the problem, by taking the lead and moving to catapillar movement to ensure I kept it. As soon as I did that, he passed me.
I was told that a leader who can carry out a plan effectively is good to have around, but a leader who can take a plan that is all FUBAR and turn it around is something else again.
There is, however, a rather unfortunate epilogue.
The message that junior got out of this was rather different. The way he understood what had happened was "yelling at subordinates will get you passed" so he spent the rest of the course screaming his head off any time he had command of anything. Icing on the top of a perfectly enjoyable experience.
DG
Something has happened to ol' Ebert in the last year or so... I can't put my finger on exactly what it is, but there's a definate change in him.
He's telling it like it is, big time. He pulls no punches, and isn't afraid to venture into some deep and muddy waters.
As a consequence, I've found myself paying much more attention to him lately, and mostly agreeing with him after the fact too.
Ebert rocks.
BTW, hi from Windsor.
DG
But then I got a better job, and the hosting site went down... and I just let it drop.
I wonder if anybody even remembers the first Netrek Home Page anymore....
http://www.cycor.ca/TCave/netrek.html
DG
But there are fixed bounds to this process.
Assume that I produce widgets, and that my profits are tie to widget sales.
So I can:
1) Sell more widgets - this is fine, but bounded. Eventually, EVERYBODY has a widget, and, assuming that widgets have an infinate life, that means my market shrinks with time.
(Of course, one can sidestep this by building a fixed life into widgets - perhaps an ever-decreasing life - but there are problems with this approach too)
2) As you mentioned, get more efficient at widget production, so that the overhead-per-widget falls with time, leaving a greater difference between price of sale and cost of production (ie, increasing profits)
But this is bounded too! Assuming perfect efficiency, eventually you get down to cost of raw material + cost of energy (at 100% efficiency) and can go no further.
So while for any given company, there may well be room within the theroretical bounds of market size and efficiency of production for some pretty sizeable growth, eventually you MUST hit a limiting factor. Then what?
DG
You have set up here a classic Straw Man argument, in which you pose a question, provide a flawed response, and then proceed to tear down the flawed premise as "proof" of your assertation.
A debate tactic as old as the hills, but no more effective today than it was in ancient Greece.
The crux of your straw man is this - the assertation that any attempt at presenting a perpetual motion machine is de facto rejected by "physicists" by virtue of the nomenclature "perpetual motion".
This is in error.
The reason why every single claim to date of the creation of a "perpetual motion" machine has been refuted, is because every single one of them has been deconstructed, analyzed, and the flaw found and demonstrated.
If you (or any other) would-be natural-law revolutionary can ACTUALLY DEMONSTRATE a machine that taps into some yet-undiscovered law of nature, you will be lauded and celebrated, not ridiculed.
The problem is that no machine so presented has ever survived the analysis - and many of them have been outright attempts at FRAUD.
Your argument is invalid at its heart, I'm afraid.
DG
(slightly off topic)
Isn't the stock market, at least the way it is operated these days, more of a scam than an investment?
Perhaps some business major can explain this to me.
As I see it, here's the problem:
1) I buy x shares of stock in SomeCompany, at some price n.
2) I want to sell this stock at some point in the future for some price y > n.
3) In order for the price of the stock to increase, it must be "worth more".
4) Assuming no further stock is issued and that I purchase no additional stock, that means my stock constitutes a fixed percentage "ownership" in the company.
5) Given that my percentage ownership is fixed, in order for the stock to be worth more at some time in the future, the value (read: size) of the company must grow.
Well, that's fine and dandy for a small company, but that seems to assume that any given company will grow without bounds pretty much forever (especially when we start talking about options)
But the "economy" is a bounded system. Those bounds may be high, but they DO exist. So the fundamental premise of "stock price increases with time" seems fundamentally flawed to me. What do you do with a successful company that produces a flat amount of profit each year, and does not seek to grow?
Now DIVIDENDS, THAT I understand. I own some share q of the company, the company produces w amount of profit, so I'm entitled to a dividend w/q - that makes sense.
But this whole speculation thing I just cannot wrap my head around. It looks like a pyramid scheme (with a gentle slope) to me.
Somebody want to explain?
DG
(Obligatory Austin Powers "nutshell" joke deleted to save space)
1) A research company did a survey of commercial Linux developers with the aim of learning more about them.
2) They gave the data they collected to Nick, to see if he could glean anything interesting from the data.
3) Nick discovered that, of the commercial Linux developers surveyed:
a) 40% considered Linux to be their "primary focus"
b) Of these 40%, more of them switched their "primary focus" from Windows than from a flavour of UNIX
3) From this, Nick concludes that Linux is stealing more development share from Windows than from UNIX derivatives - which makes inuative sense, because there are a lot more Windows developers out there to steal share from.
4) Nick is satisfied that the research company was doing good science in their survey, rather than trying to spin him.
My own analysis agrees with his conclusions.
It is an interesting conclusion, because, given the API simularity between UNIXen and Linux, one might expect that the majority of commercial Linux developers would have been UNIX developers who ported their product over to Linux, but this seems to not be the case. Developers are in fact "defecting" from Windows to Linux.
What isn't addressed is the reason for the "defection". I for one would not be suprised that the prime motivator was Microsoft eating their lunch - not much point producing a product if M$ baked it into Windows....
DG
If any of the people involved in producing these updates hang out here on Slashdot, a Linux version of these updates would be verra nice!
DG
You know, there are places where you can borrow movies for a limited time for a small fee.
DG
A quick aside: I HATE the term "scripting", as if it were some degenerate form of "real programming" - especially with feature-rich languages like perl that never have to call other applications.
Anyway, first-hand experience: thanks to the concept of perl modules and the incredible CPAN archive, writing applications that have to go to the network for things like HTTP or (especially) LDAP are trivial in perl but seriously heavy lifting in C.
You also get string parsing, regular expressions, and garbage collection built right in. Not to mention the incredibly powerful (from a code legibility standpoint) associative array or "hash" data structure.
Believe it or not, correctly written perl is orders of magnitude more legible than C or Java, because it works at a higher level of abstraction.
I wrote an LDAP->LDAP replication program, with schema and data format translation, in a couple of hours using perl.
Doing stuff like comparing the contents of a database dump (provided as a CSV) against an LDAP directory is trivial in perl.
C is best used when you won't have a perl environment availible and need the binary to stand alone. For pretty much every other task I've encountered in the last 6 years, perl got the job done faster and with much better maintainability.
DG
They've got 2 engine programs now - hell, more than 2. The restrictor plate engines have nothing in common with a short track engine save the basic parts.
Running a completely different, smaller, unrestricted engine at the superspeedways would be far cheaper than the restrictor engines.
DG
Before you go getting all bent out of shape over restrictor plates and safety, consider the following:
1) Daytona and Talledega are LONG
2) Daytona and Talledega are HIGHLY BANKED
That combination of the length of the track and especially the high banking (which provides gobs of extra cornering force) means that the cars can sustain astronomical top speeds without needing major revolutions in tire technology or wing-and-undertray levels of downforce. It's the banking that lets 'em run flat out.
NASCAR was running over 200 MPH at Daytona in the 60's, back when the cars really were production based and had stones for tires. With modern (for NASCAR) tires and suspensions, that banking could probably support speeds in excess of 260 MPH before the cars got cornering-force limited and had to slow down on corner entry.
Now with the frontal area that they have, no NASCAR car is going to be turning 260 with even unrestricted engines. The power consumed by aero drag is a function of the square of the speed, so it takes more power for the same delta v the faster you go. There's a limit to how much power you can squeeze out of even an unrestricted motor, so the real top speed would probably be somewhere in the 235 area.
But note that the guy who makes 5 HP more than his neighbor is only going to make a small fraction of a MPH more in terminal velocity.
So guess what pulling the restrictor plates off did? You get the EXACT SAME scenario as you had with the plates on, except now the speeds are 30-50 MPH faster. And kinetic energy (that must be dissipated in a crash) is a function of the square of velocity squared as well....
As bad as a Big Wreck at a buck ninety is, that pales in comparision to the same wreck at 230. And these aren't 1500lb Champ cars, these are 3600lb locomotives.
The problem with restrictor plates isn't that they cause the tight grouping of cars and the inability to pass unassisted - that's the fault of the banking. The big issue with the restrictor plate is that it takes a tremendous amount of engineering to try and coax extra air through that plate, and to get the engine to run in the odd environment the plate creates in the intake manifold. R&D costs for a 'plate engine run easily 10 times higher than a short track motor.
What NASCAR should do is make the actual engine displacement for the superspeedways smaller. Make 'em run a 3 litre V6. That'd bring costs way down while still preserving the safety.
DG
...you see this screenshot:
3 ...and know not only what track you're on, but that the car is in a position (in the braking area leading up to the chicane, on the wrong side of the track, with a couple of degrees of steering lock in the wrong direction) to crash in the next heartbeat.
http://www.aceshardware.com/read.jsp?id=5000036
DG
90% of our codebase is perl.
Yes Virginia, perl source CAN be written to optimise for legibility.
DG
I've been coding in an enterprise environment for quite some time now, and I have one rule that is cast in gold:
Always optimise source code for legibility above all else. Never trade legibility for performance unless you have no other choice, and then document your cleverness in the code so that those who follow behind you can keep up.
Here's why:
When you first write a system, it will spend its first few months of life in a very intensive quality control feedback loop. Bugs are found and very quickly exterminated. The code is still fresh in your mind and you're "in the zone".
But as the system stabilises, there is less and less reason to go back to the code, so that freshness wears off. After a little while, other priorities will take over and the internal model of the code will fade away.
But there's still bugs in there - there always is. But any bug that makes it past the first few months is non-obvious, intermittant, rare, and so on (thus, harder to find)
When one finally surfaces, _somebody_ is going to have to fix it. Sometimes it will be you, and you will appreciate code legibilty when you have to dust off source that has laid untouched for years. Not only does it increase the probability that you'll be able to actually find the bug, it cuts down on the time needed to fix it.
There's nothing like being the guy who finds and fixes bugs within seconds of them being pointed out to enhance your reputation.
But more often than not, it will be some other poor sap who gets saddled with your code and a deadline to get it fixed - and the guy who draws the short straw is normally not the biggest brain in the shop. There is no gratitude like the gratitude from someone forced to dive into somebody else's code, and who subsequently discovers that you have gone out of your way to make it easier for them to understand.
This is _also_ a reputation enhancer. "That code was so well written that not only did it take no time at all to track down the bug, but I also learned a couple of new techniques in the process!"
The true guru is a TEACHER.
Oh, and ALWAYS check the return code from every system call and provide appropriate error trapping. That's good too.
DG
OK, so the stuff is lighthearted and not particularly deep, it's still a fun read and the stories truck right along.
Eric seems to have a ken interest in history, and his work is well researched.
It's nae "Dune", but it's fun. And a few of his books are readable for free at the Baen Free Library
Check him out.
DG