The application I work on is a 2-node cluster, but it can have dozens of clients.
Debugging the distributed nature of the problem usually comes down to two things - content of the message (we use our own protocol on top of TCP) and the timing. Our debug logging can log both on the server and client sides, so usualy I just end up looking at the logs (just a visual examination) to see where a problem lies. I can see what messages were sent by whom and when, and what was recieved.
If the problem gets more complex, I have used perl to parse the logs, especially if we have a problem in the field and we get a lot of data back. As I mentioned before, I have also used Ethereal to examine what is being sent across the wire when some even more bizarre problems surfaced. I have also been using a Linux router running NIST Net to help simulate various network conditions when trying to debug problems.
One place I did find logging to be better than the debugger though was in debugging distributed applications.
Actually, I code distributed applications for a living:-) We do some logging that can be turned on and off on the fly so that we can help debug distributed problems both in house and in the field.
I guess I didn't mean to imply printf() is useless, just that it is just one small tool in debugging a problem. If printf() is the only thing you use, then you either have a simple problem or you are going to have to spend a lot of time debugging. I also get some use of of packet sniffers like Ethereal, and as I said you can do a lot with dbx. This all goes back to the original point - use all of the tools you have available, and learn how to use them well!
Indeed. I don't use VS.NET, but I can tell you that dbx on Solaris can do the same things. It can also check for buffer overruns, memory leaks, and do performance profiling. However, many people seem not to know about these features.
Somehow I think the "use printf!" crowd here are debugging their university CS assignments, and not any type of complex real-world application.
The lady or gentleman who finds it more entertaining to tweak the kernel than play Quake is that much geekier and worthy of respect in my book.
And when you make it out of University you'll realize that you have to get some work done too. I use Linux on my desktop at work - do you think I have the time to be compiling software from sratch every time I want to install or upgrade something? I use RedHat, because I know everything is tested and will work, and I can get it installed/upgraded quickly, and get on with work that needs to get done.
I realize that people like source-based distros, and I understand the appeal. However, I have to imagine that a lot of the people who run these distros are university students on the other end of a fat network pipe with a lot of spare time.
Re:Sgi gets employee to market product, film at 11
on
New SGI Altix 3000
·
· Score: 2
Who cares? Would you have felt better if I had submitted it (I would have if I had noticed the press release sooner) or if some AC had done so?
This is News for Nerds. Groundbreaking stuff for Linux in terms of perfomance and scalability. Article already written at Newsforge about it. As long as the editors post interesting stories, who cares who submitted it? Isn't this better than a story about how Microsoft did stupid thing X today?
If a bunch of people are out to get the U.S., then why are they doing it?
The U.S. is the world's superpower. The only country with economic strength and the ability to project military power. People are going to hate you because:
The U.S took the other side in some international dispute.
They resent their culture being pushed aside by U.S. pop culture.
They are jealous of the standard of living in the U.S.
I think if you look you'll find that dislike of America boils down into one of those three categories. I am Canadian, and despite the fact that we have a theoretically higher standard of living, you'll find reason #2 is most likely why someone from Canada dislikes the U.S. - we know everything about American culture, they know nothing about ours.
The problem is, when you are the world's superpower, it is hard to hide from these problems. Isolationisim was tried in the 1930s, but that didn't work out too well.
Sure, the U.S. has made foreign policy mistakes - maybe even lots of them - but there aren't any magic solutions that will make this all go away.
Funny how you make fun of the original poster, but the only piece of information in your post is complete bullshit...
Oh no! Windows Media Player sends a GUID that uniquely identifies the bundle of requests, but is in no way able to link back to your machine.
Wrong. Why not check out Microsoft's own website on this and see what is has to say: "there are occasions when unique machine-identifying information is transmitted across the Internet". That sure sounds like they can link back to your machine to me!
I'm not saying that what Media Player trasnmits is neccissarily a big deal (depends on your views about privacy, I suppose...) but lets be truthful about what they are doing.
If you had a higher UID, I'd accuse you of being a Microsoft astroturfer...
all products from MS that include.NET runtime must supply Sun Java.......this is ass-backwards..NET is a runtime enviroment (as is java of course).....if an application uses.NET at its core, for example Visual Studio.NET, they need to include the runtime
Are you being deliberately obtuse? Nobody is talking about applications - if MS wants to ship the.NET runtime with their O/S (future versions of XP, etc.) then they also have to ship the Java runtime.
MS adds extensions for Windows only development, which are optional to developers depending on their target market (HINT: Apple has Cocoa extensions in Java......samething......they are optional)
Another poster has pointed out how misleading this is. Java has a method that you can use to add extensions to the language - MS deliberaely chose not to use it, thus creating an incompatible Java implementation. This is not the "same thing" at all.
Microsoft should counter Sun needs to carry.NET with Solaris and StarOffice, as they both include Java. (And MS has a BSD runtime now...for developers, not fully completed libraries yet)
First, I don't see Microsoft shipping a.NET implementation for Solaris now or any time in the future. Second, even if they did, Sun is not a convicted monopolist, which means that they don't have to play by the same rules as MS. When will you MS apologists get this through your thick heads?
Every interesting game for the XBox (besides Halo) is available on my PC. Mechwarrior! Splinter Cell! Yay! - I can play the same or practically the same game for my PC already. I'm not saying that these games aren't fun, but they leave me zero motivation to buy an XBox since I already own a platform I can play them on.
Xbox Live is the coolest internet technology since Napster
Sorry, I've been playing online multiplayer games for many years on my PC. XBox live brings nothing new to the table here.
XBox will no doubt succeed because Microsoft is behind it with their deep pockets, and because they are pushing rehashed first-person-shooters to the teenage male demographic that eats that kind of thing up. But lets not pretend that the XBox doing anything more than that.
My friend that works for a bank here in Canada has OS/2 running on his desktop. I would imagine that banks, being an IBM stringhold, still have quite a bit of OS/2 in place.
This book reminds me of my university education. When I was done, we had done transistors, digital logic, PC architecture, assembly language, C and C++ and network protocols. I have always felt that having an understanding of how everything works has made me a better programmer.
I can't see that 'nice to have' features like specifying restrictions on textual element content outweigh the huge extra complexity compared to its predecessor
The ability to specify these restrictions are a huge advantage. For example...
<ip-addres>foobar!</ip-address>
... is almost certainly not valid input. You can easily use schema to validate that this should appear in IPv4 dotted decimal form, for example. Take any moderate size document, and suddenly schema validation saves you hundreds of lines of validation logic inside your own code. This kind of validation is also a huge improvement for XML remote procedure calls (SOAP, etc.)
I'll agree that the W3C schema language is complex, and I prefer something like Relax NG. However, I think both are a huge improvement over DTD for many applications, and if you don't need the extra features, you can stick with DTD.
For example, IIS6 is configured like Apache's httpd.conf (but true XML)
Really? Is there any public documentation for that? I'd be interested to see what they are doing.
It always seemed to be that the people who designed IIS never set up web applications for a living. For example, I have an IIS test server configured, and I want my production server to have the exact same configuration. How do you do it? Go through all of the IIS configuration windows and tabs and make sure all of the same checkboxes are checked? Talk about prone to error.
It is sad that it has taken them this long to come out with a text config file. A step in the right direction anyway.
Funny, I was never told to RTFM when I asked for commercial support.
I doubt anyone offering commercial support of closed or open source software would tell you to RTFM.
With the free support you get online, you are just as likely to get a "RTFM" from a Microsoft newsgroup/IRC channel as you are from a Linux one. Such is life with unmoderated public forums.
The open source community direly needs to lose the punks for it to be reliable for education and commercial support.
Yeah, because those "punks" are the same ones running companies who charge for commercial support of open source. Get real. Have you ever paid for commercial support of an open source product? It works just like closed-source support - a professional providing a support service for a fee, not some jackoff in an IRC channel.
I still use pine when I am dialed in, but when I am at work I have been running Evolution for about 8 months.
I guess I never got attached enough to the pine shortcuts to miss them, and some things (e.g. looking up several recipients from an LDAP address book) are much faster from Evo than from pine.
I have to say that Evolution is one fantastic application. I have been keeping up-to-date using Ximian Red Carpet, and I haven't had a problem the entire time. Kudos to the Ximian Evo hackers.
they also get stable, mostly competent majority governments most of the time
No, we get stuck with a middle-of-the-road Liberal party in power becuase the other interests in the country can't get their act together and get them out. And thanks to the lack of term limits and anything resembling the power of the US house/senate, we get essentially a dictatorship that has been in power so long that they are corrupt.
At least you have change in government every now and again, and some way to oppose a decision taken by the president.
Canadians are friendly, most countries like us, so we don't really need such protection as we don't piss people off on a regular basis.
As a Candadian, people who have this kind of myopic viewpoint really piss me off. The only reason we can spend so little on defense is that the US is our neighbor.
The US can project military power, and we simply can't. Hell, we can't even deploy troops to Afghanistan without US help.
If the US wasn't there, we'd be spending a whole lot more on defense, or we'd be run over by the first crackpot who wanted to try.
That assumes no exploitable bugs. Not convincing. The 'something they make up for tracking' could also hide a host of evils.
This is just bunk. What "exploitable bug" is going to give some website infomation about me that I never gave them? What could someone put in a session id that I would care about, if I never gave them any information? Come up with one concrete example here please, instead of making up conspiracy theories.
Logins are better for that, and my browser is happy to remember this when I go back. If you want data to persist, store it on your own server, not my client.
ROTFLMAO. You mean to tell me that you don't trust me putting a user identifier in your cookie, but you trust your browser to store your password locally? Time for a sanity check.
It never takes too long for the cookie conspirators to come out of the woodwork, does it?
It's just about impossible to know what information is being gathered through the cookie mechanism
Wrong. The only thing a website can put in a cookie is what information you give it, or something they make up for tracking a session. And better than that, you can examine your cookie file and see what is there. If you don't like cookies that are attached to ad images, get yourself a browser that blocks cookies that don't originate from the site you are visiting.
What's wrong with encoding a session identifier in the URL?
Persistence beyond the surrent session? Easy and ubiquitous support in all web development environments?
Cookies are evil and software architects need to get that through their heads.
Riiiight, because you say so. You leave your tinfoil hat on, and 99.9% of the rest of the world will go on using cookies, especially software developers who can deeply analyze that you are full of it.
So you're telling me that Sun's current attitude towards GNU is right? You don't deal with perl much, do you?
Well, I do, but I have the Sun compiler, so I am not so concerned. As long as Sun ships their own compiler, it makes sense to ship Perl compiled against it, no?
If they start shipping gcc as part of the core Solaris install then maybe, but otherwise I don't think it makes any sense.
Slashdot Mirror
The application I work on is a 2-node cluster, but it can have dozens of clients.
Debugging the distributed nature of the problem usually comes down to two things - content of the message (we use our own protocol on top of TCP) and the timing. Our debug logging can log both on the server and client sides, so usualy I just end up looking at the logs (just a visual examination) to see where a problem lies. I can see what messages were sent by whom and when, and what was recieved.
If the problem gets more complex, I have used perl to parse the logs, especially if we have a problem in the field and we get a lot of data back. As I mentioned before, I have also used Ethereal to examine what is being sent across the wire when some even more bizarre problems surfaced. I have also been using a Linux router running NIST Net to help simulate various network conditions when trying to debug problems.
One place I did find logging to be better than the debugger though was in debugging distributed applications.
Actually, I code distributed applications for a living :-) We do some logging that can be turned on and off on the fly so that we can help debug distributed problems both in house and in the field.
I guess I didn't mean to imply printf() is useless, just that it is just one small tool in debugging a problem. If printf() is the only thing you use, then you either have a simple problem or you are going to have to spend a lot of time debugging. I also get some use of of packet sniffers like Ethereal, and as I said you can do a lot with dbx. This all goes back to the original point - use all of the tools you have available, and learn how to use them well!
But how many people know about these facilities?
Indeed. I don't use VS.NET, but I can tell you that dbx on Solaris can do the same things. It can also check for buffer overruns, memory leaks, and do performance profiling. However, many people seem not to know about these features.
Somehow I think the "use printf!" crowd here are debugging their university CS assignments, and not any type of complex real-world application.
The lady or gentleman who finds it more entertaining to tweak the kernel than play Quake is that much geekier and worthy of respect in my book.
And when you make it out of University you'll realize that you have to get some work done too. I use Linux on my desktop at work - do you think I have the time to be compiling software from sratch every time I want to install or upgrade something? I use RedHat, because I know everything is tested and will work, and I can get it installed/upgraded quickly, and get on with work that needs to get done.
I realize that people like source-based distros, and I understand the appeal. However, I have to imagine that a lot of the people who run these distros are university students on the other end of a fat network pipe with a lot of spare time.
Who cares? Would you have felt better if I had submitted it (I would have if I had noticed the press release sooner) or if some AC had done so?
This is News for Nerds. Groundbreaking stuff for Linux in terms of perfomance and scalability. Article already written at Newsforge about it. As long as the editors post interesting stories, who cares who submitted it? Isn't this better than a story about how Microsoft did stupid thing X today?
You can use all the gdb commands in the command window, plus when you execute a gui action it will show you the command in the command window.
The debugger with Sun Workshop does the same thing. Sometimes I use the GUI as a "help" feature when I can't remember a command ;-)
If a bunch of people are out to get the U.S., then why are they doing it?
The U.S. is the world's superpower. The only country with economic strength and the ability to project military power. People are going to hate you because:
I think if you look you'll find that dislike of America boils down into one of those three categories. I am Canadian, and despite the fact that we have a theoretically higher standard of living, you'll find reason #2 is most likely why someone from Canada dislikes the U.S. - we know everything about American culture, they know nothing about ours.
The problem is, when you are the world's superpower, it is hard to hide from these problems. Isolationisim was tried in the 1930s, but that didn't work out too well.
Sure, the U.S. has made foreign policy mistakes - maybe even lots of them - but there aren't any magic solutions that will make this all go away.
Funny how you make fun of the original poster, but the only piece of information in your post is complete bullshit...
Oh no! Windows Media Player sends a GUID that uniquely identifies the bundle of requests, but is in no way able to link back to your machine.
Wrong. Why not check out Microsoft's own website on this and see what is has to say: "there are occasions when unique machine-identifying information is transmitted across the Internet". That sure sounds like they can link back to your machine to me!
I'm not saying that what Media Player trasnmits is neccissarily a big deal (depends on your views about privacy, I suppose...) but lets be truthful about what they are doing.
If you had a higher UID, I'd accuse you of being a Microsoft astroturfer...
all products from MS that include .NET runtime must supply Sun Java.......this is ass-backwards. .NET is a runtime enviroment (as is java of course).....if an application uses .NET at its core, for example Visual Studio .NET, they need to include the runtime
Are you being deliberately obtuse? Nobody is talking about applications - if MS wants to ship the .NET runtime with their O/S (future versions of XP, etc.) then they also have to ship the Java runtime.
MS adds extensions for Windows only development, which are optional to developers depending on their target market (HINT: Apple has Cocoa extensions in Java......samething......they are optional)
Another poster has pointed out how misleading this is. Java has a method that you can use to add extensions to the language - MS deliberaely chose not to use it, thus creating an incompatible Java implementation. This is not the "same thing" at all.
Microsoft should counter Sun needs to carry .NET with Solaris and StarOffice, as they both include Java. (And MS has a BSD runtime now...for developers, not fully completed libraries yet)
First, I don't see Microsoft shipping a .NET implementation for Solaris now or any time in the future. Second, even if they did, Sun is not a convicted monopolist, which means that they don't have to play by the same rules as MS. When will you MS apologists get this through your thick heads?
Man, I can still remember how much I hated E.T. on the 2600 as a little kid. Some Atari 2600 E.T. karma whoring:
More about the game.
The author discusses coding it up in 5 weeks.
the Xbox is outselling the GameCube everywhere
Everywhere in North America you mean.
Don't own a Xbox? Too bad for you. It rocks.
Every interesting game for the XBox (besides Halo) is available on my PC. Mechwarrior! Splinter Cell! Yay! - I can play the same or practically the same game for my PC already. I'm not saying that these games aren't fun, but they leave me zero motivation to buy an XBox since I already own a platform I can play them on.
Xbox Live is the coolest internet technology since Napster
Sorry, I've been playing online multiplayer games for many years on my PC. XBox live brings nothing new to the table here.
XBox will no doubt succeed because Microsoft is behind it with their deep pockets, and because they are pushing rehashed first-person-shooters to the teenage male demographic that eats that kind of thing up. But lets not pretend that the XBox doing anything more than that.
Where have you been?
My friend that works for a bank here in Canada has OS/2 running on his desktop. I would imagine that banks, being an IBM stringhold, still have quite a bit of OS/2 in place.
This book reminds me of my university education. When I was done, we had done transistors, digital logic, PC architecture, assembly language, C and C++ and network protocols. I have always felt that having an understanding of how everything works has made me a better programmer.
I can't see that 'nice to have' features like specifying restrictions on textual element content outweigh the huge extra complexity compared to its predecessor
The ability to specify these restrictions are a huge advantage. For example...
<ip-addres>foobar!</ip-address>
... is almost certainly not valid input. You can easily use schema to validate that this should appear in IPv4 dotted decimal form, for example. Take any moderate size document, and suddenly schema validation saves you hundreds of lines of validation logic inside your own code. This kind of validation is also a huge improvement for XML remote procedure calls (SOAP, etc.)
I'll agree that the W3C schema language is complex, and I prefer something like Relax NG. However, I think both are a huge improvement over DTD for many applications, and if you don't need the extra features, you can stick with DTD.
For example, IIS6 is configured like Apache's httpd.conf (but true XML)
Really? Is there any public documentation for that? I'd be interested to see what they are doing.
It always seemed to be that the people who designed IIS never set up web applications for a living. For example, I have an IIS test server configured, and I want my production server to have the exact same configuration. How do you do it? Go through all of the IIS configuration windows and tabs and make sure all of the same checkboxes are checked? Talk about prone to error.
It is sad that it has taken them this long to come out with a text config file. A step in the right direction anyway.
Funny, I was never told to RTFM when I asked for commercial support.
I doubt anyone offering commercial support of closed or open source software would tell you to RTFM.
With the free support you get online, you are just as likely to get a "RTFM" from a Microsoft newsgroup/IRC channel as you are from a Linux one. Such is life with unmoderated public forums.
The open source community direly needs to lose the punks for it to be reliable for education and commercial support.
Yeah, because those "punks" are the same ones running companies who charge for commercial support of open source. Get real. Have you ever paid for commercial support of an open source product? It works just like closed-source support - a professional providing a support service for a fee, not some jackoff in an IRC channel.
I still use pine when I am dialed in, but when I am at work I have been running Evolution for about 8 months.
I guess I never got attached enough to the pine shortcuts to miss them, and some things (e.g. looking up several recipients from an LDAP address book) are much faster from Evo than from pine.
I have to say that Evolution is one fantastic application. I have been keeping up-to-date using Ximian Red Carpet, and I haven't had a problem the entire time. Kudos to the Ximian Evo hackers.
Exactly; everybody uses RADIUS. Why the hell would you use DHCP for auth?
Check out how this was done at years past at the Ottawa Linux Symposium.
No, we get stuck with a middle-of-the-road Liberal party in power becuase the other interests in the country can't get their act together and get them out. And thanks to the lack of term limits and anything resembling the power of the US house/senate, we get essentially a dictatorship that has been in power so long that they are corrupt.
At least you have change in government every now and again, and some way to oppose a decision taken by the president.
Canadians are friendly, most countries like us, so we don't really need such protection as we don't piss people off on a regular basis.
As a Candadian, people who have this kind of myopic viewpoint really piss me off. The only reason we can spend so little on defense is that the US is our neighbor.
The US can project military power, and we simply can't. Hell, we can't even deploy troops to Afghanistan without US help.
If the US wasn't there, we'd be spending a whole lot more on defense, or we'd be run over by the first crackpot who wanted to try.
This is just bunk. What "exploitable bug" is going to give some website infomation about me that I never gave them? What could someone put in a session id that I would care about, if I never gave them any information? Come up with one concrete example here please, instead of making up conspiracy theories.
Logins are better for that, and my browser is happy to remember this when I go back. If you want data to persist, store it on your own server, not my client.
ROTFLMAO. You mean to tell me that you don't trust me putting a user identifier in your cookie, but you trust your browser to store your password locally? Time for a sanity check.
It never takes too long for the cookie conspirators to come out of the woodwork, does it?
It's just about impossible to know what information is being gathered through the cookie mechanism
Wrong. The only thing a website can put in a cookie is what information you give it, or something they make up for tracking a session. And better than that, you can examine your cookie file and see what is there. If you don't like cookies that are attached to ad images, get yourself a browser that blocks cookies that don't originate from the site you are visiting.
What's wrong with encoding a session identifier in the URL?
Persistence beyond the surrent session? Easy and ubiquitous support in all web development environments?
Cookies are evil and software architects need to get that through their heads.
Riiiight, because you say so. You leave your tinfoil hat on, and 99.9% of the rest of the world will go on using cookies, especially software developers who can deeply analyze that you are full of it.
Well, I do, but I have the Sun compiler, so I am not so concerned. As long as Sun ships their own compiler, it makes sense to ship Perl compiled against it, no?
If they start shipping gcc as part of the core Solaris install then maybe, but otherwise I don't think it makes any sense.