XP was more stable than ME (the previous incarnation of Windows that I had installed on my computer). I could see the difference straight away, because it didn't feel as flaky. I didn't have any fancy hardware to support, so XP was pretty much what I needed (more stability).
My complaint here is not about Microsoft releasing operating systems before the computers are ready for them (it's a chicken and egg situation and I totally understand that), it's about retailers not giving an option. At the moment 2GB systems are not that common (at least around here), yet vista is forced on you. It would make sense to offer an alternative at least until 4GB machines were pretty much the standard (meaning their price was not in the "Advanced Gamers" range, but in the "Home User" range). I also understand that it's Microsoft not allowing retailers to do that, rather than retailers trying to annoy users. That's why I only complain on slashdot instead of going to the shop and demanding a refund...
How will that become less important over time? If I go to PC World and buy a computer, and it has vista (well, it does), then I'm pretty much stuck with it. It's not going to become less important to me - I'm stuck with a computer that could have had better performance but doesn't, and it wasn't through a choice of mine.
Of course, if I plan to upgrade my computer every year, then it will eventually become less important to me. But it is so difficult to get rid of computers these days that I won't be upgrading any time soon. Until then, I've got Yahoo games not working with vista ("known problem"), a slow shutdown, and a 2GB RAM half-decent system that would have carried a lot more on XP.
It's not even "early adoption" - it's been out for a while, and it doesn't look like it's getting any better.
But did you ask what to do on the forums? There's a very friendly community out there, and someone must have experienced exactly the same problem. And you've got the code, so you can see what's causing it.
Better yet, I'll sell you a "boot subscription" from my off-shore company. Whenever you boot, you have to request your encryption keys from my server. The thing is that you can cancel your subscription with a simple phone call ("Please enter your account number..."). Once you do that your computer is a brick.
Being off-shore, I don't have to reveal anything to anyone.
I filled the forms online, but wasn't sure about the year the house was built (it was home insurance).
Later on I found out, and phoned them to update the details. They then informed me that actually they don't insure tenants, and that the insurance wasn't actually valid (it wasn't mentioned on the website - the form actually had a "tenant" option).
I had no choice but to say thank you and cancel it, but I still got charged for the two days I was "insured".
To be honest, I couldn't be bothered with chasing them (the phone calls would have cost me more), but of course I cancelled my other policies with them and will never be their customer (nor recommend them).
At the time I thought it was extremely cheeky - had I not phoned them I could have experienced a disaster only to find that I wasn't insured!
By the way, an insurance company rarely has a problem with giving you back the premiums when they don't want to pay. If I wanted to pay for insurance only to get the premiums back, I would have opened a bank account...
- We only cover your car if you drive according to the law. Three years ago you were going 2mph above the speed limit, hence you invalidated your policy and we are not obliged to pay. - Why didn't you notify me then? - According to the policy, we're not obliged to do that either. - Are you obliged to do anything? - Maybe, but we're not obliged to answer that question.
Yes, a nation of managers who couldn't care less whether they're working for Nike, the car industry, the tobacco industry, or the weapon industry.
As long as the money keeps rolling in, they're doing their job and can sleep at night.
When you don't care about the product you're selling, how exactly do you motivate your staff? I hope you have a very long whip and a very strong hand...
So how many lives has vaccination actually saved? I want a pie chart by 3:00pm today. I don't buy this "spend now to save later" crap. We're a business, and if it's not profit, it must be an expense.
By the way, we have the same business-focus attitude towards software design and QA. Prove to me that QA actually saves money when all I see is a payroll and a bunch of people clicking random buttons and running the same tests over and over again. As for software design, I don't even have to open my mouth - time away from the keyboard is simply wasted (again, where's the revenue from the "process"?)
You don't even need strncpy to write unsafe code. Look:
int *p;
*p = 5;
Amazing, isn't it?
Yes, it does make sense to learn how to use a programming language before using it. It's possible to use most <string.h> functions in an unsafe way - so what? The point is that some functions are inherently unsafe (strlen, strcpy) whereas some can actually be safe, if one knows how to use them, of course.
What do you mean "safety pin"? I just pricked myself!
I totally agree - the one example I always give when I meet a "self-documenting code" programmer is the case of implementing a standard.
You accept one form of floating point, but reject another - the code is crystal clear, but the customer wants to be able to write 3.8e+8 and you don't allow it. Another programmer goes into the code and quickly changes it to allow that 3.8e+8 format. Next thing a test fails (the test that makes sure that 3.8e+8 will not be accepted). The programmer thinks "ah, that's a bad test, it doesn't allow all possible floating points formats, probably someone was day-dreaming" and quickly changes the test not to fail.
So far, so good - the code was clear, the test was clear, but someone, somewhere will have a very big problem with the "numbers separated by the character 'e'" case as described in the standard.
A comment/* Not allowing other formats, as per section 3.27.1.55.3 of the standard */ is all it takes.
I am about to write a tiny XML parser for an embedded system (why XML? I didn't write the spec...). I know all available element and attribute names, and I just need to find them very quickly.
gperf was the first thing I was thinking of, but looks like Cuckoo hashes is going to be my choice now.
No idea why a score of 0 with such a contribution, but thanks anyway.
One thing still missing
on
Project Arcade
·
· Score: 2, Interesting
Having to pay for each game, and taking pride in spending 15 minutes (or 45 minutes) on a single quarter (or any other coin).
I think it was the fact that I had to pay for each attempt that made me memorise most of Slapfight and R-Type. I've played both on MAME (R-Type on the Wii as well), and I don't have the same motivation. If I die, I die. Press 5 a few more times and have another go. It's also the amount invested so far that can make a difference (when you realise that you've already spent around $100 or $500 on this game, you're not going to leave it now, not without your name on the high score list.
This will never be replicated at home, even if you add the coin mechanism, because you know you still have the money there.
I'm not suggesting that money is the only incentive, and that you can't enjoy arcade games without it, but I do believe it's a big part of it.
Choose a place for the King: 8 choose 1 = 8 Choose a place for the Queen: 7 choose 1 = 7 Choose two places for the Rooks: 6 choose 2 = 15 Choose two places for the Bishops: 4 choose 2 = 6 The other two places are for the Knights.
All together: 8 * 7 * 15 * 6 = 7!
A different way: there are 8! possible orders, but then we have two Rooks, two Bishops and two Knights, so all together 8!/2^3 = 7!
This is true for each player, so we have (7!)^2 = 25401600 starting positions
16! = 20922789888000 which is slightly bigger.
It would be interesting to play with two same-coloured bishops... very limiting indeed.
But the ZX Spectrum had many 'B' games, simply because it was so easy to write and publish at the time.
"Gulpman", anyone? How about "Pedro"?
I once bought an adventure game only to find that it was text only and quite boring. I didn't laugh. I can't even remember the name (but side B wouldn't load... disgraceful).
So you're not using Links at work? Actually, you can also read slashdot with emacs (but it does look suspicious if you don't normally use it... "Yes, I really like this editor. Here, have a look... wait... how do I open a file here... just a sec... I'm sure it was ctrl-something or another, or maybe alt-shift something...")
I'm always amazed at those messages: "We recently hired xyz and it turns out he's totally useless."
Not blaming any one, but are job interviews out of fashion? What sort of questions do people ask in interviews? Or, in other words, can't you tell within 20 minutes that someone simply doesn't have a clue?
It doesn't explain *why* you do something, it only explains (in a sort of indirect way) *what* you're telling a machine to do.
Reading the highly optimised multidimensional probability calculations involved in speech recognition doesn't give you a clue about *why* you see what you see. Unfortunately, not even memorising the code back to front does that. You'll have to sit and read the theory, and the math, and the articles, and learn the expertise from the people you work with - the expertise that isn't even written yet, because it's a trade secret.
Nobody would bother writing an entire ISO/IEC standard into the C comments. But you still have to know it to understand why something is done one way and not another. Or even worse - how to fix it to comply with the standard. That's not something you will find in the code. You'll have to ask someone or just read the 576 pages of densely written Jargonese to understand what you're trying to do.
Code is only code. It's like reading without nouns, adjectives and vowels. Adding documentation adds the vowels to the verbs, but you still only have verbs. Reading a poem made of verbs is not funny. Amending that poem in a meaningful way is even worse, until you can actually hear the poem in your mind.
Slashdot Mirror
XP was more stable than ME (the previous incarnation of Windows that I had installed on my computer). I could see the difference straight away, because it didn't feel as flaky. I didn't have any fancy hardware to support, so XP was pretty much what I needed (more stability).
My complaint here is not about Microsoft releasing operating systems before the computers are ready for them (it's a chicken and egg situation and I totally understand that), it's about retailers not giving an option. At the moment 2GB systems are not that common (at least around here), yet vista is forced on you. It would make sense to offer an alternative at least until 4GB machines were pretty much the standard (meaning their price was not in the "Advanced Gamers" range, but in the "Home User" range). I also understand that it's Microsoft not allowing retailers to do that, rather than retailers trying to annoy users. That's why I only complain on slashdot instead of going to the shop and demanding a refund...
How will that become less important over time? If I go to PC World and buy a computer, and it has vista (well, it does), then I'm pretty much stuck with it. It's not going to become less important to me - I'm stuck with a computer that could have had better performance but doesn't, and it wasn't through a choice of mine.
Of course, if I plan to upgrade my computer every year, then it will eventually become less important to me. But it is so difficult to get rid of computers these days that I won't be upgrading any time soon. Until then, I've got Yahoo games not working with vista ("known problem"), a slow shutdown, and a 2GB RAM half-decent system that would have carried a lot more on XP.
It's not even "early adoption" - it's been out for a while, and it doesn't look like it's getting any better.
But did you ask what to do on the forums? There's a very friendly community out there, and someone must have experienced exactly the same problem. And you've got the code, so you can see what's causing it.
Better yet, I'll sell you a "boot subscription" from my off-shore company. Whenever you boot, you have to request your encryption keys from my server. The thing is that you can cancel your subscription with a simple phone call ("Please enter your account number..."). Once you do that your computer is a brick.
Being off-shore, I don't have to reveal anything to anyone.
... and sell them for safe-keeping.
But is it better than hotbot?
"Excellent question, but this is not the issue. The issue is really ... <political drivel>"
In other words, the advertisers are stealing from me: time, bandwidth (=money) and attention.
According to Microsoft, this is obviously the other way around: websites should change themselves to support the new Operating System.
Because we don't like this "OS independency" that websites seem to enjoy at the moment.
This actually happened to me.
I filled the forms online, but wasn't sure about the year the house was built (it was home insurance).
Later on I found out, and phoned them to update the details. They then informed me that actually they don't insure tenants, and that the insurance wasn't actually valid (it wasn't mentioned on the website - the form actually had a "tenant" option).
I had no choice but to say thank you and cancel it, but I still got charged for the two days I was "insured".
To be honest, I couldn't be bothered with chasing them (the phone calls would have cost me more), but of course I cancelled my other policies with them and will never be their customer (nor recommend them).
At the time I thought it was extremely cheeky - had I not phoned them I could have experienced a disaster only to find that I wasn't insured!
By the way, an insurance company rarely has a problem with giving you back the premiums when they don't want to pay. If I wanted to pay for insurance only to get the premiums back, I would have opened a bank account...
The insurance industry would love that as well.
- We only cover your car if you drive according to the law. Three years ago you were going 2mph above the speed limit, hence you invalidated your policy and we are not obliged to pay.
- Why didn't you notify me then?
- According to the policy, we're not obliged to do that either.
- Are you obliged to do anything?
- Maybe, but we're not obliged to answer that question.
I've come up with a name for this invention.
How about "telescreen"?
Yes, a nation of managers who couldn't care less whether they're working for Nike, the car industry, the tobacco industry, or the weapon industry.
As long as the money keeps rolling in, they're doing their job and can sleep at night.
When you don't care about the product you're selling, how exactly do you motivate your staff? I hope you have a very long whip and a very strong hand...
Or vaccination for that matter.
So how many lives has vaccination actually saved? I want a pie chart by 3:00pm today. I don't buy this "spend now to save later" crap. We're a business, and if it's not profit, it must be an expense.
By the way, we have the same business-focus attitude towards software design and QA. Prove to me that QA actually saves money when all I see is a payroll and a bunch of people clicking random buttons and running the same tests over and over again. As for software design, I don't even have to open my mouth - time away from the keyboard is simply wasted (again, where's the revenue from the "process"?)
You don't even need strncpy to write unsafe code. Look:
int *p;
*p = 5;
Amazing, isn't it?
Yes, it does make sense to learn how to use a programming language before using it. It's possible to use most <string.h> functions in an unsafe way - so what? The point is that some functions are inherently unsafe (strlen, strcpy) whereas some can actually be safe, if one knows how to use them, of course.
What do you mean "safety pin"? I just pricked myself!
I totally agree - the one example I always give when I meet a "self-documenting code" programmer is the case of implementing a standard.
/* Not allowing other formats, as per section 3.27.1.55.3 of the standard */ is all it takes.
You accept one form of floating point, but reject another - the code is crystal clear, but the customer wants to be able to write 3.8e+8 and you don't allow it. Another programmer goes into the code and quickly changes it to allow that 3.8e+8 format. Next thing a test fails (the test that makes sure that 3.8e+8 will not be accepted). The programmer thinks "ah, that's a bad test, it doesn't allow all possible floating points formats, probably someone was day-dreaming" and quickly changes the test not to fail.
So far, so good - the code was clear, the test was clear, but someone, somewhere will have a very big problem with the "numbers separated by the character 'e'" case as described in the standard.
A comment
Thank you!
I am about to write a tiny XML parser for an embedded system (why XML? I didn't write the spec...). I know all available element and attribute names, and I just need to find them very quickly.
gperf was the first thing I was thinking of, but looks like Cuckoo hashes is going to be my choice now.
No idea why a score of 0 with such a contribution, but thanks anyway.
Having to pay for each game, and taking pride in spending 15 minutes (or 45 minutes) on a single quarter (or any other coin).
I think it was the fact that I had to pay for each attempt that made me memorise most of Slapfight and R-Type. I've played both on MAME (R-Type on the Wii as well), and I don't have the same motivation. If I die, I die. Press 5 a few more times and have another go. It's also the amount invested so far that can make a difference (when you realise that you've already spent around $100 or $500 on this game, you're not going to leave it now, not without your name on the high score list.
This will never be replicated at home, even if you add the coin mechanism, because you know you still have the money there.
I'm not suggesting that money is the only incentive, and that you can't enjoy arcade games without it, but I do believe it's a big part of it.
I thought it goes like this:
Choose a place for the King: 8 choose 1 = 8
Choose a place for the Queen: 7 choose 1 = 7
Choose two places for the Rooks: 6 choose 2 = 15
Choose two places for the Bishops: 4 choose 2 = 6
The other two places are for the Knights.
All together: 8 * 7 * 15 * 6 = 7!
A different way: there are 8! possible orders, but then we have two Rooks, two Bishops and two Knights, so all together 8!/2^3 = 7!
This is true for each player, so we have (7!)^2 = 25401600 starting positions
16! = 20922789888000 which is slightly bigger.
It would be interesting to play with two same-coloured bishops... very limiting indeed.
But the ZX Spectrum had many 'B' games, simply because it was so easy to write and publish at the time.
"Gulpman", anyone?
How about "Pedro"?
I once bought an adventure game only to find that it was text only and quite boring. I didn't laugh. I can't even remember the name (but side B wouldn't load... disgraceful).
I have upgraded to Feisty, also with zero problems, so now we're winning...
So you're not using Links at work? Actually, you can also read slashdot with emacs (but it does look suspicious if you don't normally use it... "Yes, I really like this editor. Here, have a look... wait... how do I open a file here... just a sec... I'm sure it was ctrl-something or another, or maybe alt-shift something...")
I'm always amazed at those messages: "We recently hired xyz and it turns out he's totally useless."
Not blaming any one, but are job interviews out of fashion? What sort of questions do people ask in interviews? Or, in other words, can't you tell within 20 minutes that someone simply doesn't have a clue?
"I'm sorry, we don't provide that service any more, because there is no suitable category in our invoicing software."
I wonder why they call code "code"...
It doesn't explain *why* you do something, it only explains (in a sort of indirect way) *what* you're telling a machine to do.
Reading the highly optimised multidimensional probability calculations involved in speech recognition doesn't give you a clue about *why* you see what you see. Unfortunately, not even memorising the code back to front does that. You'll have to sit and read the theory, and the math, and the articles, and learn the expertise from the people you work with - the expertise that isn't even written yet, because it's a trade secret.
Nobody would bother writing an entire ISO/IEC standard into the C comments. But you still have to know it to understand why something is done one way and not another. Or even worse - how to fix it to comply with the standard. That's not something you will find in the code. You'll have to ask someone or just read the 576 pages of densely written Jargonese to understand what you're trying to do.
Code is only code. It's like reading without nouns, adjectives and vowels. Adding documentation adds the vowels to the verbs, but you still only have verbs. Reading a poem made of verbs is not funny. Amending that poem in a meaningful way is even worse, until you can actually hear the poem in your mind.