It's not just the people, it's the way they are treated.
Most people with no worries are nice pleasant folks. People who have to regard life as a constant battle are likely to be hostile.
Are you more likely to be of pleasant demeanour, if you know that after a work day in one hour you are
a) Going home to a nicely appointed white-picket-fence home where your wife, who can afford not to work full time, has a good home-cooked meal waiting for you
OR
b) Going to your next job as fry cook in the McDonalds. You'll probably eat the free food there again, because you're trying to save money because you're behind on the rent. The bus journey there will eat about a fifth of your wage packet for the day.
Steve Jobs, and many others, could probably live like Gregory Peck in The Million Pound Note ; people are willing to provide you goods and services if they think you're good for it, and will bring them more business later...
HTTP requests are raw user input. You don't want raw user input anywhere near a kernel module.
Kernel-mode caching. Requests for cached responses are served without switching to user mode.
If you hadn't put an HTTP handler in the kernel, you wouldn't need a switch of context.
Kernel-mode request queuing. Requests cause less overhead in context switching, because the kernel forwards requests directly to the correct worker process. If no worker process is available to accept a request, the kernel-mode request queue holds the request until a worker process picks it up.
You could do that in a user process.
When a worker process fails, service is not interrupted; the failure is undetectable by the user because the kernel queues the requests while the WWW service starts a new worker process for that application pool.
You could do that in a user process too.
Requests are processed faster because they are routed directly from the kernel to the appropriate user-mode worker process instead of being routed between two user-mode processes.
And there's the real reason it's done - it should say "IPC sucks real bad in Windows, so we made this stupid, stupid, idiotic hack to try and compete with that other OS we're not mentioning."
> "When an electronic medical record is printed out, the amount of repetitive data in it is ridiculous,"
Some of that is because the standards mandate it. In order that you don't lose context for a stray page, the patient demographics have top be on each sheet, etc, etc.
It's actually Intersystems Caché - it's a trade name.
I had a look at it about 10-12 years ago when I was actively developing on EMRs. My overwhelming reaction to it was "ugh".
It's one of those lock-in products that's incredibly expensive because people have built vast sprawling systems on it and they don't dare migrate to something else.
I was always fascinated by them, but they were a rich kid's toy back then - they cost about the same in numbers, but the forces of inflation have made them so much cheaper now...
The team decides what the indentation style should be
Not on any project I run. "Discussions" about indentation style lead to some of the most unpleasant, passive-aggressive, time-wasting and downright unproductive meetings I've ever attended. There's always at least one or two people who are unwilling to compromise on what they consider to be the "best" style (as you can see by scrolling around this discussion).
The reason I tend to prefer the IDE default (for languages where it's a no-brainer to use the IDE, like Java) is because it requires the least effort to get right. There's always some bonehead who joins the project later, doesn't read the style guide, and starts committing using the default settings. If your project standard is the default settings, this is fine. On projects where you wasted around 8 hours determining that the indent should be four spaces and there should be a rigid right-column margin of 80 chars, this means that someone will check out the source, open it in Eclipse, and start committing changes indented with tabs. The IDE guys already had all those arguments about style, and they almost certainly have more collective experience than you. Use that experience, don't waste your life reinventing the wheel.
Have the team lead pick a simple, easily followed style, and allow a certain amount of leeway - as the saying goes foolish consistency is the hobgoblin of little minds. The point is to make the code easier to understand, and if there are occasions when deviating from the style guide improves this, do it.
Spend the time you would have wasted arguing about tabs or spaces creating a culture of discussion about what makes for clearer, more understandable code. Rules of thumb like "if you find yourself with the urge to comment a particular stretch of code, consider that you might just write the code more clearly".
There's the much more readily verifiable fact that red-green colour blindness allows you to see right through various forms of camouflage, and this has been exploited by the military in various settings.
Lining up comments after the code is a stupid bonehead thing to do, because per-line comments are a stupid bonehead thing to do.
If you have to comment every line to explain it's purpose, spend that effort making the code clearer instead.
Comments should be reserved for areas of special magic which are not immediately obvious and can't be refactored for whatever reason, and expanding on the contracts of public methods.
I had project leads do that, use code beautifiers that re-indented, split, and re-ordered methods in files that were widely forked...
I'm a tabs person, because tabs give you more choice about how far a single level of indent is.
Spaces are more the totalitarian "I think 4 spaces is the right indent" person's choice.
But yes, suck it up and follow the convention on the project.
If the project uses a particular IDE, use the default for that IDE (which for Java / Eclipse - is tabs. Clearly you should also munge all the line-endings to plain LF with your version control system....)
While it's no Harry Potter and the Order of the Phoenix (which apparently lost $167M after grossing $938M worldwide, on a $150M budget - go figure!), you'd think that was actually a respectable profit.
The reason they sign them is because the labels can offer them promotion and distribution. Faced with the choice of building their name through hard graft, and having a pre-made audience delivered to them on a plate, many choose to sign for the label. Faced with a flash, successful, A&R man who promises the earth and knows some really cool people, many choose to sign for the label.
You have deals where the artist sells $188M of CDs, and declares bankruptcy.
You can't continue to have that patriotism in view of the observable facts - your government is by the few, for the few, and the people be damned. I agree that it's great that a government was founded with these ideals - but as laid out in the constitution, the time would seem to have come to throw what you have away and replace it with a government for the people again.
The buck stops with management. They get the pay, they get the responsibility.
Of course, they're the ones who assess performance as well. No way are they actually going to take the heat for that.
So the story is : bad management. They're not putting in the appropriate checks and balances, probably because they cost money. They're not interested in making a good product, they want to pad their pay packets. So the buck goes all the way to the top, to the people who decide remuneration policies.
If the software developers don't give a damn, they're not being selected or motivated appropriately by management.
And this is one of the myriad reasons why bonus culture sucks.
Excel also succeeded because it had no format lock-in. Because it could WRITE Lotus 1-2-3 just as well as it read it, there was no risk to using Excel and finding that it didn't perform as well as Lotus.
Lotus was the incumbent at the time. 1-2-3 was the killer app that drove adoption of the PC. Yes, Excel worked in pretty graphics mode. Yes, Excel was better than 1-2-3. But you've seen management clinging like limpets to older solutions to things just because of their elevated perception of risk. If Excel hadn't been able to read 1-2-3 files perfectly, it would never have happened.
It's exactly the same reason why people won't migrate from MS Office to LibreOffice - because it's not entirely compatible, and everyone else uses it. It's all but impossible to make an entirely compatible program though - because even the MOO-XML formats are just a serialization of binary structs and even *puke* Windows API calls. Office isn't a standalone program - it only works on Windows.
Yeah. Most people use software like it was voodoo. This specific ritual gets you these results.
Any change to their rituals and they end up like a rabbit in the headlights. These rituals have been built up over a lifetime of use, without any real understanding of the primitives they are composed of.
Indeed. Being a programmer who's moved further into the management zone, I'm having to cope with Office a lot more as my colleagues have a seriously restricted comfort zone when parted from the Microsoft teat for more than a few minutes.
The behaviour of both MS Office and LibreOffice infuriate me. As a relative novice (the last word processor I used seriously was Wordperfect 5.1 when I was at university), even the most basic, simple tasks seem so kludgy and clumsy.
It has got to the point where I'm considering writing a Markdown parser in VBA so I can continue to write my documents in wikitext and read them into the corporate template by running a macro. At this point, people will suggest using something like Pandoc, but it sucks horribly no matter what you do - I've tried direct conversion to DOCX, I've tried converting to ODT and then saving that as DOC(X) from LibreOffice - compatibility with Office is still horrible, and not surprisingly, as another poster points out, the internal formats are horrific and the 3,600 page MOO-XML documentation really reveals that.
The only thing that writes Office documents properly is Office - so automate Office to load documents from other formats, don't try and duplicate all that horror in other programs to convert to it.
It might be an open-source mod for Apache, but it's backed by an Oracle database. It only issues PL/SQL calls, so you can't just slap in another RDBMS.
It's such a common case that Java will usually optimize looped string concat into a StringBuilder now. I imagine Python does the same thing.
The performance probably still sucks because the buffer needs re-allocating periodically (in StringBuilder, it doubles each time), but not as much as it would in VB6 which has no such cleverness in it's immutable string type. There's also all the garbage collection overhead for all those discarded byte arrays.
Would be more interested to see a benchmark if they declared the StringBuilder with a buffer of the size they expected to use.
They're only examining the performance of concatenating immutable strings, versus the performance of writing to a (buffered) stream.
This is a problem that's been known about for donkey's ages. It's just that computers are so stupidly powerful it's no longer an issue that many programmers ever have to confront.
In VB6 you had to jump through hoops to do it properly, but it's such a common case in Java that the compiler will optimize repeated concatenations in a loop into using a StringBuilder instead. I presume Python has similar optimizations.
I used to just stick all the strings in an array, allocate a new string of the appropriate length, and copy them into it.
News at 11 : many less experienced programmers are ignorant of the internal workings of their chosen frameworks, because they never had to write their own implementation at a lower level.
You're young, and this change ought to scare you. Us oldies have an outside chance of being able to afford some of these fine robots to replace your ass.
Slashdot Mirror
It's not just the people, it's the way they are treated.
Most people with no worries are nice pleasant folks. People who have to regard life as a constant battle are likely to be hostile.
Are you more likely to be of pleasant demeanour, if you know that after a work day in one hour you are
a) Going home to a nicely appointed white-picket-fence home where your wife, who can afford not to work full time, has a good home-cooked meal waiting for you
OR
b) Going to your next job as fry cook in the McDonalds. You'll probably eat the free food there again, because you're trying to save money because you're behind on the rent. The bus journey there will eat about a fifth of your wage packet for the day.
Steve Jobs, and many others, could probably live like Gregory Peck in The Million Pound Note ; people are willing to provide you goods and services if they think you're good for it, and will bring them more business later...
Decent *and* sensible.
He might draw a much smaller salary - but as he notes, a $1M salary had low marginal value for him.
What he just did was remove all money worries from his staff. Now all their focus can go on increasing the value of his business.
Good for them, good for him. Good morals, and good sense.
And they're fucking stupid reasons.
HTTP requests are raw user input. You don't want raw user input anywhere near a kernel module.
Kernel-mode caching. Requests for cached responses are served without switching to user mode.
If you hadn't put an HTTP handler in the kernel, you wouldn't need a switch of context.
Kernel-mode request queuing. Requests cause less overhead in context switching, because the kernel forwards requests directly to the correct worker process. If no worker process is available to accept a request, the kernel-mode request queue holds the request until a worker process picks it up.
You could do that in a user process.
When a worker process fails, service is not interrupted; the failure is undetectable by the user because the kernel queues the requests while the WWW service starts a new worker process for that application pool.
You could do that in a user process too.
Requests are processed faster because they are routed directly from the kernel to the appropriate user-mode worker process instead of being routed between two user-mode processes.
And there's the real reason it's done - it should say "IPC sucks real bad in Windows, so we made this stupid, stupid, idiotic hack to try and compete with that other OS we're not mentioning."
> "When an electronic medical record is printed out, the amount of repetitive data in it is ridiculous,"
Some of that is because the standards mandate it. In order that you don't lose context for a stray page, the patient demographics have top be on each sheet, etc, etc.
It's actually Intersystems Caché - it's a trade name.
I had a look at it about 10-12 years ago when I was actively developing on EMRs. My overwhelming reaction to it was "ugh".
It's one of those lock-in products that's incredibly expensive because people have built vast sprawling systems on it and they don't dare migrate to something else.
Bigtrak!
I was always fascinated by them, but they were a rich kid's toy back then - they cost about the same in numbers, but the forces of inflation have made them so much cheaper now...
Now they have one that has an app!
The team decides what the indentation style should be
Not on any project I run. "Discussions" about indentation style lead to some of the most unpleasant, passive-aggressive, time-wasting and downright unproductive meetings I've ever attended. There's always at least one or two people who are unwilling to compromise on what they consider to be the "best" style (as you can see by scrolling around this discussion).
The reason I tend to prefer the IDE default (for languages where it's a no-brainer to use the IDE, like Java) is because it requires the least effort to get right. There's always some bonehead who joins the project later, doesn't read the style guide, and starts committing using the default settings. If your project standard is the default settings, this is fine. On projects where you wasted around 8 hours determining that the indent should be four spaces and there should be a rigid right-column margin of 80 chars, this means that someone will check out the source, open it in Eclipse, and start committing changes indented with tabs. The IDE guys already had all those arguments about style, and they almost certainly have more collective experience than you. Use that experience, don't waste your life reinventing the wheel.
Have the team lead pick a simple, easily followed style, and allow a certain amount of leeway - as the saying goes foolish consistency is the hobgoblin of little minds. The point is to make the code easier to understand, and if there are occasions when deviating from the style guide improves this, do it.
Spend the time you would have wasted arguing about tabs or spaces creating a culture of discussion about what makes for clearer, more understandable code. Rules of thumb like "if you find yourself with the urge to comment a particular stretch of code, consider that you might just write the code more clearly".
This was the most reliably citable thing I could find : a medical paper discussing side effects of cataract lens replacements.
http://crstodayeurope.com/2011...
There's the much more readily verifiable fact that red-green colour blindness allows you to see right through various forms of camouflage, and this has been exploited by the military in various settings.
Lots of anecdotes here : http://www.reddit.com/r/todayi...
Lining up comments after the code is a stupid bonehead thing to do, because per-line comments are a stupid bonehead thing to do.
If you have to comment every line to explain it's purpose, spend that effort making the code clearer instead.
Comments should be reserved for areas of special magic which are not immediately obvious and can't be refactored for whatever reason, and expanding on the contracts of public methods.
I had project leads do that, use code beautifiers that re-indented, split, and re-ordered methods in files that were widely forked...
I'm a tabs person, because tabs give you more choice about how far a single level of indent is.
Spaces are more the totalitarian "I think 4 spaces is the right indent" person's choice.
But yes, suck it up and follow the convention on the project.
If the project uses a particular IDE, use the default for that IDE (which for Java / Eclipse - is tabs. Clearly you should also munge all the line-endings to plain LF with your version control system....)
made almost no money
Budget : $5M
Takings : $59M
While it's no Harry Potter and the Order of the Phoenix (which apparently lost $167M after grossing $938M worldwide, on a $150M budget - go figure!), you'd think that was actually a respectable profit.
The whole of Hollywood is a scam.
The contracts are only lucrative for the labels.
The reason they sign them is because the labels can offer them promotion and distribution. Faced with the choice of building their name through hard graft, and having a pre-made audience delivered to them on a plate, many choose to sign for the label. Faced with a flash, successful, A&R man who promises the earth and knows some really cool people, many choose to sign for the label.
You have deals where the artist sells $188M of CDs, and declares bankruptcy.
You can't continue to have that patriotism in view of the observable facts - your government is by the few, for the few, and the people be damned. I agree that it's great that a government was founded with these ideals - but as laid out in the constitution, the time would seem to have come to throw what you have away and replace it with a government for the people again.
True patriotism would be revolution.
The buck stops with management. They get the pay, they get the responsibility.
Of course, they're the ones who assess performance as well. No way are they actually going to take the heat for that.
So the story is : bad management. They're not putting in the appropriate checks and balances, probably because they cost money. They're not interested in making a good product, they want to pad their pay packets. So the buck goes all the way to the top, to the people who decide remuneration policies.
If the software developers don't give a damn, they're not being selected or motivated appropriately by management.
And this is one of the myriad reasons why bonus culture sucks.
Medical bills in the US are high for a variety of reasons, more expertly laid out than I can here by one of the vlogbrothers
Most of the reason is your costs.
e.g.
Lipitor, 30 days dosage ; US $124 ; New Zealand $7
Excel also succeeded because it had no format lock-in. Because it could WRITE Lotus 1-2-3 just as well as it read it, there was no risk to using Excel and finding that it didn't perform as well as Lotus.
Lotus was the incumbent at the time. 1-2-3 was the killer app that drove adoption of the PC. Yes, Excel worked in pretty graphics mode. Yes, Excel was better than 1-2-3. But you've seen management clinging like limpets to older solutions to things just because of their elevated perception of risk. If Excel hadn't been able to read 1-2-3 files perfectly, it would never have happened.
It's exactly the same reason why people won't migrate from MS Office to LibreOffice - because it's not entirely compatible, and everyone else uses it. It's all but impossible to make an entirely compatible program though - because even the MOO-XML formats are just a serialization of binary structs and even *puke* Windows API calls. Office isn't a standalone program - it only works on Windows.
Yeah. Most people use software like it was voodoo. This specific ritual gets you these results.
Any change to their rituals and they end up like a rabbit in the headlights. These rituals have been built up over a lifetime of use, without any real understanding of the primitives they are composed of.
Indeed. Being a programmer who's moved further into the management zone, I'm having to cope with Office a lot more as my colleagues have a seriously restricted comfort zone when parted from the Microsoft teat for more than a few minutes.
The behaviour of both MS Office and LibreOffice infuriate me. As a relative novice (the last word processor I used seriously was Wordperfect 5.1 when I was at university), even the most basic, simple tasks seem so kludgy and clumsy.
It has got to the point where I'm considering writing a Markdown parser in VBA so I can continue to write my documents in wikitext and read them into the corporate template by running a macro. At this point, people will suggest using something like Pandoc, but it sucks horribly no matter what you do - I've tried direct conversion to DOCX, I've tried converting to ODT and then saving that as DOC(X) from LibreOffice - compatibility with Office is still horrible, and not surprisingly, as another poster points out, the internal formats are horrific and the 3,600 page MOO-XML documentation really reveals that.
The only thing that writes Office documents properly is Office - so automate Office to load documents from other formats, don't try and duplicate all that horror in other programs to convert to it.
It might be an open-source mod for Apache, but it's backed by an Oracle database. It only issues PL/SQL calls, so you can't just slap in another RDBMS.
Oracle costs some serious money.
Don't forget the economics of decomissioning the thing. You're writing up a huge debt for your country in 60 years.
It's such a common case that Java will usually optimize looped string concat into a StringBuilder now. I imagine Python does the same thing.
The performance probably still sucks because the buffer needs re-allocating periodically (in StringBuilder, it doubles each time), but not as much as it would in VB6 which has no such cleverness in it's immutable string type. There's also all the garbage collection overhead for all those discarded byte arrays.
Would be more interested to see a benchmark if they declared the StringBuilder with a buffer of the size they expected to use.
They're only examining the performance of concatenating immutable strings, versus the performance of writing to a (buffered) stream.
This is a problem that's been known about for donkey's ages. It's just that computers are so stupidly powerful it's no longer an issue that many programmers ever have to confront.
In VB6 you had to jump through hoops to do it properly, but it's such a common case in Java that the compiler will optimize repeated concatenations in a loop into using a StringBuilder instead. I presume Python has similar optimizations.
I used to just stick all the strings in an array, allocate a new string of the appropriate length, and copy them into it.
News at 11 : many less experienced programmers are ignorant of the internal workings of their chosen frameworks, because they never had to write their own implementation at a lower level.
The FAQ posted by Pixar explictly allows this.
They only prohibit direct revenue.
Indirect revenue (e.g. YouTube ad fees) are permitted, but you must credit RenderMan.
You're young, and this change ought to scare you. Us oldies have an outside chance of being able to afford some of these fine robots to replace your ass.