This just reeks of the problem with Google. They are sooo into themselves and into thinking that everyone smart works there. The reality is that far less than "all great programming brains" actually work at Google. I'd bet that not even 5% of the best programming brains work there - in fact, I'd bet that less than 1% even work there.
Google has become so cocky as of late that they think all the good people want to or do work there. That's just not the case.
We've recently hired 4 or 5 guys away from Google, and they are so into themselves for having worked at Google that they are almost impossible to work with. They think that they have some special 'rights' just because their resume says "google" on it. They are far from the most talented engineers where I work. But don't go telling them that.
Not even sure where to begin. But I guess I should start my response by qualifying how I've used Perforce, so that we can at least agree that I'm qualified to comment on it (which you assert I am not qualified to do).
I use Perforce at a Fortune 500 company, with over 3000 developers on the system. My particular project spans Linux for the back end (approximately 5mm lines of code), and uses Windows on the front end (another 2mm lines of code). Our developers span the globe, from the West Coast of the USA, to New York, to London, to Hong Kong, Australia.
So I think we can agree that I'm qualified to comment on the platform.
I do not work at Perforce, never have, don't own shares, etc. etc.
Quite a few of the arguments I see you have above are the religious ones - you'd prefer that your files are read/write 100% of the time. Perforce's model is that if you don't have the file checked out, then you can't touch it or change it. The reason it does this is that it wants to know exactly when you _started_ editing a file, so that that it can keep track of any changes that might come after you. Timestamps in the filesystem are not able to do this, so you have to tell revision control when you start editing.
OK, so the religion of read-only takes some getting used to, and it's different than 100% writable files. But the idea is pretty simple - don't ever edit a read-only file. Check it out, then edit it.
100% of the IDE integrations just do this for you, so if you are on an IDE it's pretty seamless. If you are in VI or emacs, you have to check the file out before you touch it.
If you don't agree with that religion, p4 is not for you.
Some of the commands you say are impossible are actually super simple, so let me clarify for anyone reading who wants to know how to do these things.
- Host bound "client-workspace"
This isn't true - just set the host to an empty string, and it works on any host.
- All state on the server
You'll have to let me know why this is a bad thing. So if you lose your local machine, you have EVERYTHING still in revision control? That sounds great.
- Branching
You state that branching is impossible and takes many commands, but that just isn't true. o Create a branch
p4 integrate dev/... branchname/...
p4 submit branchname/...
o Merge between branches
p4 integrate dev/... branchname/...
p4 resolve branchname/... # resolve any conflicts
p4 submit branchname/...
- Every file that was opened by another developer while you were also working on it will be marked as a conflict
Umm.. if you didn't change the file, then you probably never checked it out. If you checked it out because the read-only thing was bothering you, then just revert any unchanged files via:
p4 revert -a...
- There is no easy way to see which files have changed from the CLI
p4 opened...
- Slow
I have to just plain disagree - given my project size, and the fact that we were on a WAN, it was very very fast. Our tree was enormous.
- Nullified changes
You don't really substantiate this with much information, so all I can say is that I haven't seen what you are referring to here.
- Actions take 5 to 10 seconds in CVS take 10 minutes in p4
Again, not much information here, but I've never seen p4 take time to do really much of anything. The only thing that can take a while is if you have to pull a 5 GB file over a slow network connection, then it's going to take a while in any revision control system. But that is just a network constraint, not a p4 constraint.
Seriously. I know it's $800 a user, or something like that, but it's worth every penny. (Yeah, yeah, it's closed source, and ships as a binary, but it _really_ works).
Note that Perforce is free for open source projects.
I have to say that I've had some people hired away from me to go to Google, and they have been hiring the people who can quote chapter and verse of some coding standards doc. But they haven't been my superstars. They have been "A" players. But not "superstars". I'll qualify that in one second.
The superstar is more than just somewhat hard to come by.
First, they are only going to be 1 out of every 100 programmers you work with. And that is only if you are lucky, and if you are good at hiring. If you hit job boards, you aren't good at hiring. (with apologies to the job board advertisement that is almost definitely above this post:)
Second, they can almost never identify themselves. Lots of people THINK they are the superstar. But then they get very little actually accomplished. These are the people I've lost to Google. But the superstar does much more than just know the tech details. They actually get stuff done. And their code really really works. And it is highly reusable. And they change others around them. The always make sure the best tools are in place, and they get others to use those tools, not just themselves. In this sense, they are also quite good leaders, although most do not want to manage large teams (and you'd be wise not to have them do so).
I've probably worked with 1000-2000 programmers in my lifetime, and I think I would give only about 10 of them the "superstar" status.
The superstars produce 2x to 10x what a very good programmers can produce in the same amount of time.
As far as finding and hiring them, the biggest problem is that they are very rarely on the market. So job boards are a bad place to start.
Just about all (maybe even 100%, actually) of the superprogrammers I've hired have come from friend referrals.
Go to your current employees, and give them very big checks if they can attract other programmers to your firm. Make sure this is worth their while (ie: $10,000 or more for bringing in someone). This will almost always be your best bet to find these guys.
Actually, I didn't hear the police asking him to leave. I just saw them forcing him to leave. I'll watch again, but I think they went straight to force.
But if you listen to what he is saying, he does say "I will get up and leave". In other words, there is a very large potential that no force at all was necessary.
Here in lies the problem. Once the police start using force, the natural reaction of any human being is to react back. Particularly if you are being arrested for really not doing any wrong. (Remember, it isn't a crime to be an a*hole. And it isn't a crime to talk loudly or obnoxiously.)
Once the police resorted to force, this outcome was inevitable.
So the question one has to ask is "did the police need to resort to force against this person?"
When you add up business class tickets from NYC ($2000), drinking the entire mini-bar at the hotel ($600), taxis to and from the airport ($100), a hooker ($500), and $50 to bribe the cop that turned out to be disguised as a hooker, you can get to $3000 in no time!
I think we are overstating Lawyer's ability to figure out what the next "big thing" will actually be. They are usually late to the game, and only in a position to post-sue, rather than preventitive sue.
I think (may be mis stating this) Napster was around for at least a year before the lawyers made their way into court. Of course, that just proves that "better late than never" is also on the lawyers play card.
Lets hope they don't shut down the current web as we know it!
OK, mod me flamebait, but I've read the same comments over and over again on Slashdot. To reiterate everything that is always said when this conversation comes up:
C++ is faster than java / C#
C++ has better templates
There is so much C++ out there that no one should ever use another language, or they will be forced into reinventing the wheel
Using a "simpler" language like java means that you will hire crappy programmers. Don't do that - use C++ and hire really good programmers.
Having worked on Wall St. as a programmer for 15 years or so, I'm sorry to say to all the people who constantly make that argument that they are officially behind the times.
The real reason to use java?
Compilation times
Compilation times
Compilation times
When programmers are a major cost (believe me, at Wall St. firms, they are a huge cost compared with the cost of machines), you should optimize the amount of time they spend waiting for compiles. We have a C++ library that is around 250,000 lines of code, that compiles in around 2 hours (and this is after a lot of work to try to make the compilation as fast as possible). We also have a 1,200,000 line java library that compiles in around 3 minutes. Any way you play it, programmers still spend a lot of time in the write code / compile / test cycle. In fact, this is where they spend most of their time.
In terms of performance, that is largely a myth that is no longer true. We run pricing simulations in java (which is about as computationally intensive as it gets) and we are able to do so at the same speed is the equivalent C++ code. In fact, the java code is a bit faster now, as the tools for finding the bottlenecks are much faster.
The hiring market? This is largely moving towards java. Many schools are not teaching C++ any more.
The work problem? In java, you spend much more of your time working on the business problem, and much less time trying to figure out why GCC arcane error message on file X means that your template is out of whack in file Y. Or in C++, trying to locate the source of a core dump (getting java.lang.NullPointerException is SOO much better).
Yeah, great, then we can all get in arguments about whether or not C++ laws run faster than java laws, and how C# laws are biased towards large corporations. And, of course, about how the VB laws aren't really laws after all.
I just got it from CompUSA. (After being turned down at EB Games for not having "reserved" a copy - this game was harder to get than Britney tickets - although frankly, D3 warrents it and Britney doesn't:)
After the 1+ hour install, I launched the game, chose my strategy and POOF!! bye-bye doom3. Looks like I'll need a patch of some sort before I can play. Rats.
Has anyone else had stability issues with the game? Any pointers?
I think that there are places out there you can go to find good music. However, you have to like one type of band - jambands.
Etree.org is a large reference site for bands that allow taping. I know that, as a fan of Phish and Dave Matthews, I was able to find another 15 bands that I really liked thanks to the bands taping policies, and high bandwidth becoming the norm.
It is odd to me that Mr. Katz would publish something so filled with anger, and so lacking in evidence. All he seems to talk about here is how we should all hate Microsoft.
If I'm not mistaken, Mr. Katz is saying that we should hate Microsoft because they are releasing new products and actively marketing them.
On Slashdot, I would think that we can be open minded about everything, even Micro$oft. Lets look at the history:
o Solaris 2.8 and Linux both ship with Browsers. So surely it has to be OK for Microsoft to also ship an operating system with a browser.
o Linux has no marketing department (some comapnies have taken on the burden of marketing Linux, but not to the degree that Microsoft markets its products). This has to be considered one of the many burdens currently holding Linux back from mass popularity like Microsoft has. Linux can't strike the level of deals that Microsoft has with vendors such as Dell, Compaq, Gateway and HP.
o Linux is free. This is certainly a burdon on Microsoft.
This is going to be one of the great battles of all time, and we are only getting started. Lets not muddy the waters with slander. It's a pretty even fight:
A product that costs $$$, and where you are paying for (mostly) marketing of that product
vs.
A product that is free, technically superior, but is not marketed.
Slashdot Mirror
This just reeks of the problem with Google. They are sooo into themselves and into thinking that everyone smart works there. The reality is that far less than "all great programming brains" actually work at Google. I'd bet that not even 5% of the best programming brains work there - in fact, I'd bet that less than 1% even work there.
Google has become so cocky as of late that they think all the good people want to or do work there. That's just not the case.
We've recently hired 4 or 5 guys away from Google, and they are so into themselves for having worked at Google that they are almost impossible to work with. They think that they have some special 'rights' just because their resume says "google" on it. They are far from the most talented engineers where I work. But don't go telling them that.
Technically, it's MSFT. MS is Morgan Stanley.
Not even sure where to begin. But I guess I should start my response by qualifying how I've used Perforce, so that we can at least agree that I'm qualified to comment on it (which you assert I am not qualified to do).
I use Perforce at a Fortune 500 company, with over 3000 developers on the system. My particular project spans Linux for the back end (approximately 5mm lines of code), and uses Windows on the front end (another 2mm lines of code). Our developers span the globe, from the West Coast of the USA, to New York, to London, to Hong Kong, Australia.
So I think we can agree that I'm qualified to comment on the platform.
I do not work at Perforce, never have, don't own shares, etc. etc.
Quite a few of the arguments I see you have above are the religious ones - you'd prefer that your files are read/write 100% of the time. Perforce's model is that if you don't have the file checked out, then you can't touch it or change it. The reason it does this is that it wants to know exactly when you _started_ editing a file, so that that it can keep track of any changes that might come after you. Timestamps in the filesystem are not able to do this, so you have to tell revision control when you start editing.
OK, so the religion of read-only takes some getting used to, and it's different than 100% writable files. But the idea is pretty simple - don't ever edit a read-only file. Check it out, then edit it.
100% of the IDE integrations just do this for you, so if you are on an IDE it's pretty seamless. If you are in VI or emacs, you have to check the file out before you touch it.
If you don't agree with that religion, p4 is not for you.
Some of the commands you say are impossible are actually super simple, so let me clarify for anyone reading who wants to know how to do these things.
- Host bound "client-workspace"
This isn't true - just set the host to an empty string, and it works on any host.
- All state on the server
You'll have to let me know why this is a bad thing. So if you lose your local machine, you have EVERYTHING still in revision control? That sounds great.
- Branching
You state that branching is impossible and takes many commands, but that just isn't true.
o Create a branch
p4 integrate dev/... branchname/...
p4 submit branchname/...
o Merge between branches
p4 integrate dev/... branchname/...
p4 resolve branchname/... # resolve any conflicts
p4 submit branchname/...
- Every file that was opened by another developer while you were also working on it will be marked as a conflict
Umm .. if you didn't change the file, then you probably never checked it out. If you checked it out because the read-only thing was bothering you, then just revert any unchanged files via:
p4 revert -a ...
- There is no easy way to see which files have changed from the CLI
p4 opened ...
- Slow
I have to just plain disagree - given my project size, and the fact that we were on a WAN, it was very very fast. Our tree was enormous.
- Nullified changes
You don't really substantiate this with much information, so all I can say is that I haven't seen what you are referring to here.
- Actions take 5 to 10 seconds in CVS take 10 minutes in p4
Again, not much information here, but I've never seen p4 take time to do really much of anything. The only thing that can take a while is if you have to pull a 5 GB file over a slow network connection, then it's going to take a while in any revision control system. But that is just a network constraint, not a p4 constraint.
- You have to shell out good money
Seriously. I know it's $800 a user, or something like that, but it's worth every penny. (Yeah, yeah, it's closed source, and ships as a binary, but it _really_ works).
Note that Perforce is free for open source projects.
I have to say that I've had some people hired away from me to go to Google, and they have been hiring the people who can quote chapter and verse of some coding standards doc. But they haven't been my superstars. They have been "A" players. But not "superstars". I'll qualify that in one second.
:)
The superstar is more than just somewhat hard to come by.
First, they are only going to be 1 out of every 100 programmers you work with. And that is only if you are lucky, and if you are good at hiring. If you hit job boards, you aren't good at hiring. (with apologies to the job board advertisement that is almost definitely above this post
Second, they can almost never identify themselves. Lots of people THINK they are the superstar. But then they get very little actually accomplished. These are the people I've lost to Google. But the superstar does much more than just know the tech details. They actually get stuff done. And their code really really works. And it is highly reusable. And they change others around them. The always make sure the best tools are in place, and they get others to use those tools, not just themselves. In this sense, they are also quite good leaders, although most do not want to manage large teams (and you'd be wise not to have them do so).
I've probably worked with 1000-2000 programmers in my lifetime, and I think I would give only about 10 of them the "superstar" status.
The superstars produce 2x to 10x what a very good programmers can produce in the same amount of time.
As far as finding and hiring them, the biggest problem is that they are very rarely on the market. So job boards are a bad place to start.
Just about all (maybe even 100%, actually) of the superprogrammers I've hired have come from friend referrals.
Go to your current employees, and give them very big checks if they can attract other programmers to your firm. Make sure this is worth their while (ie: $10,000 or more for bringing in someone). This will almost always be your best bet to find these guys.
Actually, I didn't hear the police asking him to leave. I just saw them forcing him to leave. I'll watch again, but I think they went straight to force.
But if you listen to what he is saying, he does say "I will get up and leave". In other words, there is a very large potential that no force at all was necessary.
...
Here in lies the problem. Once the police start using force, the natural reaction of any human being is to react back. Particularly if you are being arrested for really not doing any wrong. (Remember, it isn't a crime to be an a*hole. And it isn't a crime to talk loudly or obnoxiously.)
Once the police resorted to force, this outcome was inevitable.
So the question one has to ask is "did the police need to resort to force against this person?"
Probably not
When you add up business class tickets from NYC ($2000), drinking the entire mini-bar at the hotel ($600), taxis to and from the airport ($100), a hooker ($500), and $50 to bribe the cop that turned out to be disguised as a hooker, you can get to $3000 in no time!
I think we are overstating Lawyer's ability to figure out what the next "big thing" will actually be. They are usually late to the game, and only in a position to post-sue, rather than preventitive sue.
I think (may be mis stating this) Napster was around for at least a year before the lawyers made their way into court. Of course, that just proves that "better late than never" is also on the lawyers play card.
Lets hope they don't shut down the current web as we know it!
C++ is faster than java / C#
C++ has better templates
There is so much C++ out there that no one should ever use another language, or they will be forced into reinventing the wheel
Using a "simpler" language like java means that you will hire crappy programmers. Don't do that - use C++ and hire really good programmers.
Having worked on Wall St. as a programmer for 15 years or so, I'm sorry to say to all the people who constantly make that argument that they are officially behind the times.
The real reason to use java?
Compilation times
Compilation times
Compilation times
When programmers are a major cost (believe me, at Wall St. firms, they are a huge cost compared with the cost of machines), you should optimize the amount of time they spend waiting for compiles. We have a C++ library that is around 250,000 lines of code, that compiles in around 2 hours (and this is after a lot of work to try to make the compilation as fast as possible). We also have a 1,200,000 line java library that compiles in around 3 minutes. Any way you play it, programmers still spend a lot of time in the write code / compile / test cycle. In fact, this is where they spend most of their time.
In terms of performance, that is largely a myth that is no longer true. We run pricing simulations in java (which is about as computationally intensive as it gets) and we are able to do so at the same speed is the equivalent C++ code. In fact, the java code is a bit faster now, as the tools for finding the bottlenecks are much faster.
The hiring market? This is largely moving towards java. Many schools are not teaching C++ any more.
The work problem? In java, you spend much more of your time working on the business problem, and much less time trying to figure out why GCC arcane error message on file X means that your template is out of whack in file Y. Or in C++, trying to locate the source of a core dump (getting java.lang.NullPointerException is SOO much better).
Now, mod me flame.
Yeah, great, then we can all get in arguments about whether or not C++ laws run faster than java laws, and how C# laws are biased towards large corporations. And, of course, about how the VB laws aren't really laws after all.
That was exactly the problem. Thanks! Newest catalyst and DX 9.0c did the trick.
I just got it from CompUSA. (After being turned down at EB Games for not having "reserved" a copy - this game was harder to get than Britney tickets - although frankly, D3 warrents it and Britney doesn't :)
After the 1+ hour install, I launched the game, chose my strategy and POOF!! bye-bye doom3. Looks like I'll need a patch of some sort before I can play. Rats.
Has anyone else had stability issues with the game? Any pointers?
-mb
Actually, C# has this feature - you can use triple quotes. As in:
String foo = """quote your \n stuff here""";
Etree.org is a large reference site for bands that allow taping. I know that, as a fan of Phish and Dave Matthews, I was able to find another 15 bands that I really liked thanks to the bands taping policies, and high bandwidth becoming the norm.
-mb
If I'm not mistaken, Mr. Katz is saying that we should hate Microsoft because they are releasing new products and actively marketing them.
On Slashdot, I would think that we can be open minded about everything, even Micro$oft. Lets look at the history:
o Solaris 2.8 and Linux both ship with Browsers. So surely it has to be OK for Microsoft to also ship an operating system with a browser.
o Linux has no marketing department (some comapnies have taken on the burden of marketing Linux, but not to the degree that Microsoft markets its products). This has to be considered one of the many burdens currently holding Linux back from mass popularity like Microsoft has. Linux can't strike the level of deals that Microsoft has with vendors such as Dell, Compaq, Gateway and HP.
o Linux is free. This is certainly a burdon on Microsoft.
This is going to be one of the great battles of all time, and we are only getting started. Lets not muddy the waters with slander. It's a pretty even fight:
A product that costs $$$, and where you are paying for (mostly) marketing of that product
vs.
A product that is free, technically superior, but is not marketed.
I wish both sides well.