umm i went to cmu, it wasn't like a 24/7 party at, say, UCSB or univ florida.. but at the same time I had tons of fun and worked my ass off. Now I have a nice job, paying more than enough, doing what I want, I own the brand new Porsche _and_ I partied a lot/travelled in college. I realized that CMU's education was so much better than most other schools, I try my best to hire CMU grads, but it's too damn hard. You made the wrong decision. Live with it.
Yes they're idiots. They wouldn't know the difference bewtween GPRS vs WAP if it bit them in the ass. "Hey bubba this hear wap is crap". They should all be slaughtered for stupidity.
never gonna happen, it's too proprietery. The new verison of wap supports xhtml, and this provides your imode-like functionality. To get the imode-like speed we're going to have to do something more complex and scrap many of the cell towers we have now and put in more per area. Ie, not very soon.
So don't diss WAP.
This will be abused just like MIT's unix boxes were in the early 90's when they left the guest password open to everyone.
a rm -rf / by wing of MOD(I hate those bastards) ruined that for everyone - no more guest account afterthat.
Check out Intermail from Openwave. They have a corporate and a large-scale solution. Intermail includes an LDAP server, message store, mta, queue servers, etc.. it's a very nice setup, and is ported to most unix flavors.
Don't even think about iPlanet, that project is a complete disaster. Sun's crappy product plus Netscape's crappy message store = one shitty product. I don't know why intermail isn't mentioned on slashdot, it is the best unix mail software.
Never once has Solaris dbx dumped core on me and I have used it extremely heavily. I debug running processes almost daily, I look at cores almost daily. It stands up extremely well. It sounds like you used an old version of dbx.
Solaris has very evolved tools. My company does development on Solaris then ports to other platforms.
Solaris's dbx is _rock solid_ for mthread apps.
Also you have pstack and truss (ie strace for linux). Purify runs on Solaris too. Also you can get vsar, the most amazing tool ever invented. It enapsulates everything in sar but in a vt100 interface so you can monitor _everything_ realtime.
I don't know if Solaris x86 is as nice as Solaric Sparc.
I had the OS course with Satya. The best professor/class I've ever had - to this day I use concepts learned in that class.
I think this guy is just bitter since they upped the units from 12 to 18 after he took it.
I work for the company that makes verizon's email servers. I actually coded a few anti-spam techniques. Our servers can block mail user harvesters(try all email until they get good response) by banning IPs that unsuccesfully send too many email msgs. They could tighten relaying. if it's not done already. They could tie in third party spam servers... They could go nuts if they wanted to. We have a huge array of tools they could use which they don't seem to use. Adding IP's and domains is easy but more importantly it's fast. You can block >100,000 IP's and _still_ not drain CPU, I know cuz I coded the algorithms to store IPs;-). There is no excuse other than Verizon not admin'ing their servers correctly. But hopefully this should change.
Do a search on google - I just did a search on "cryptography society" and it came up with many usefull sites. Google is perfect for this type of search btw - if you wanna know why, learn about the algorithm. How do you find this? search google..
ignorance is bliss.. Microsoft has used their monopoly power to stifle innovation. Ie when they changed winsock so it wasn't fully unix socket-compatible. Put the Winsock company out of business. Microsoft was sued because of this, but they just paid Winsock off. This happened too many times to count - MSFT had too much money. They had too much power. At the IETF, MSFT employees usually refer to "standards" as Microsoft standards. This always generated a laugh at IEFT meetings. In fact, no one from MSFT showed up at the last IETF meeting, most likely because they realized IETF attendance is hopeless. In the IETF, MSFT has as much clout as any other Joe Schmoe. MSFT didn't like that.;-) RIP Microsoft, we have all had enough of you.
I remember the early trade wars games. I distinctly remember the first version I played - I changed my name to include ANSI codes so that when the score board was shown, the cursor would move up and put my "Izz be jammin" nick in the #1 spot, then move back down. Soon after that it filtered out ANSI codes.:(
Check out this palm doc to pdf converter. It's free but it is a plugin for a $30 shareware InstallBuddyForWindows pacakge. I don't have windows, but if someone else who has a Windows box could download InstallBuddy + this plugin as well as the pdf file, convert it and post the link to the doc file, I would be happy.
If I'm storing gigs of mp3's, I might want to play these mp3's at work as well as in my car. So what would be nice is a storage medium which I can put mp3s on and play them with a player which may be my desktop, small computer in my car,...
Sort of like a portable hard drive. If someone can put something like that out, they can leverage off of that and make many players... Is this done? If not someone needs to "make it so."
Yes it will indeed do wonders - optimally you should have N+1 threads where N = # of processors.
BTW - thanks to the Linux Scalability Project, the 2.3 Linux kernel will perform asynch I/O very efficiently. Netscape is 100% responsible for this - their imap server uses the same asynch. architecture so they patched the kernel for their imap server under the guise of this project.
This would be true if it weren't for the fact that most JVM's GCs are implemented poorly. You can have the fastest Java code in the world, but the JVM's GC bottleneck will almost always appear when you don't want it.
At my company, scalability is an issue since the largest ISP's you can think of use our product. Most of our servers are written in C++. We have been able to invest time into it over the last 5 years and make them very fast. Other servers are in Java and these servers have serious performance problems. The Java developpers, who are really really good, have spent >1 year trying to improve performance - minimizing allocations, increasing the frequency of the GC thread run, etc.. But still when there are huge spikes of usage, the GC thread can't keep up, the process exhausts memory and core dumps.
For a company like mine, Java currently doesn't cut it. C++ gives more bang for the buck. Sure it may take a _little_ longer to write fast, correct code with C++ but the added runtime performance is worth it in many cases.
There are many issues you should worry about when programming if you want fast code, issues which may be in the implementation details. That's real life.
Maybe a garbage collector could be efficient but current implementations are terriblly slow - a "smidgen slower" is an understatement. Specifically, look at jvm's. They use one GC thread and this is bad when for example there is a huge spike in memory allocation, this thread gets behind and memory is exhausted. This is primarily why Java just does not currently scale. C/C++ with no garbage collector will outperform java any day, esp. when there are spikes in allocations.
I claim that there is no need for a garbage collector. Finding memory leaks quickly is not a problem _if_ you have the right tools. Purify is good but it is too slow. There are tools I have developped which can show with pinpoint accuracy where the leaks are and have no performance penalty.
So if you are able to find leaks quickly, with no performance pentalty, then why would you want a garbage collector? I am still unconvinced.
Slashdot Mirror
all ikea's are like that throughout the world. Damn swedes.
umm i went to cmu, it wasn't like a 24/7 party at, say, UCSB or univ florida.. but at the same time I had tons of fun and worked my ass off. Now I have a nice job, paying more than enough, doing what I want, I own the brand new Porsche _and_ I partied a lot/travelled in college. I realized that CMU's education was so much better than most other schools, I try my best to hire CMU grads, but it's too damn hard. You made the wrong decision. Live with it.
Yes they're idiots. They wouldn't know the difference bewtween GPRS vs WAP if it bit them in the ass. "Hey bubba this hear wap is crap". They should all be slaughtered for stupidity.
Yeah uh except WAP 2.0 supports xHTML, a larger subset of HTML compared to cHTML. Plus it's standard whereas cHTML is no where near a standard.
never gonna happen, it's too proprietery. The new verison of wap supports xhtml, and this provides your imode-like functionality. To get the imode-like speed we're going to have to do something more complex and scrap many of the cell towers we have now and put in more per area. Ie, not very soon. So don't diss WAP.
You say aryan as if that's an insult. Is there any derogative word towards white people really, besides white trash.
This will be abused just like MIT's unix boxes were in the early 90's when they left the guest password open to everyone. a rm -rf / by wing of MOD(I hate those bastards) ruined that for everyone - no more guest account afterthat.
Check out Intermail from Openwave. They have a corporate and a large-scale solution. Intermail includes an LDAP server, message store, mta, queue servers, etc.. it's a very nice setup, and is ported to most unix flavors.
Here's a link.
Don't even think about iPlanet, that project is a complete disaster. Sun's crappy product plus Netscape's crappy message store = one shitty product. I don't know why intermail isn't mentioned on slashdot, it is the best unix mail software.
What about a bag of crap? How about some rotten meat? Cmon let's get some real tests done.
L'enfers c'est les autres!
If you intreview for a C position at my company I will laugh at you.
Never once has Solaris dbx dumped core on me and I have used it extremely heavily. I debug running processes almost daily, I look at cores almost daily. It stands up extremely well. It sounds like you used an old version of dbx.
Hrmm I seem to remember trying to use dbx to do this and it was even slower than Purify. I could be wrong though.
Solaris has very evolved tools. My company does development on Solaris then ports to other platforms. Solaris's dbx is _rock solid_ for mthread apps. Also you have pstack and truss (ie strace for linux). Purify runs on Solaris too. Also you can get vsar, the most amazing tool ever invented. It enapsulates everything in sar but in a vt100 interface so you can monitor _everything_ realtime. I don't know if Solaris x86 is as nice as Solaric Sparc.
I had the OS course with Satya. The best professor/class I've ever had - to this day I use concepts learned in that class. I think this guy is just bitter since they upped the units from 12 to 18 after he took it.
I work for the company that makes verizon's email servers. I actually coded a few anti-spam techniques. Our servers can block mail user harvesters(try all email until they get good response) by banning IPs that unsuccesfully send too many email msgs. They could tighten relaying. if it's not done already. They could tie in third party spam servers... They could go nuts if they wanted to. We have a huge array of tools they could use which they don't seem to use. Adding IP's and domains is easy but more importantly it's fast. You can block >100,000 IP's and _still_ not drain CPU, I know cuz I coded the algorithms to store IPs ;-). There is no excuse other than Verizon not admin'ing their servers correctly. But hopefully this should change.
Do a search on google - I just did a search on "cryptography society" and it came up with many usefull sites. Google is perfect for this type of search btw - if you wanna know why, learn about the algorithm. How do you find this? search google..
ignorance is bliss.. Microsoft has used their monopoly power to stifle innovation. Ie when they changed winsock so it wasn't fully unix socket-compatible. Put the Winsock company out of business. Microsoft was sued because of this, but they just paid Winsock off. This happened too many times to count - MSFT had too much money. They had too much power. At the IETF, MSFT employees usually refer to "standards" as Microsoft standards. This always generated a laugh at IEFT meetings. In fact, no one from MSFT showed up at the last IETF meeting, most likely because they realized IETF attendance is hopeless. In the IETF, MSFT has as much clout as any other Joe Schmoe. MSFT didn't like that. ;-) RIP Microsoft, we have all had enough of you.
I remember the early trade wars games. I distinctly remember the first version I played - I changed my name to include ANSI codes so that when the score board was shown, the cursor would move up and put my "Izz be jammin" nick in the #1 spot, then move back down. Soon after that it filtered out ANSI codes. :(
Yes the Cabal were bastards!
Check out this palm doc to pdf converter. It's free but it is a plugin for a $30 shareware InstallBuddyForWindows pacakge. I don't have windows, but if someone else who has a Windows box could download InstallBuddy + this plugin as well as the pdf file, convert it and post the link to the doc file, I would be happy.
"Do Bwah" is the French pronounciation
If I'm storing gigs of mp3's, I might want to play these mp3's at work as well as in my car. So what would be nice is a storage medium which I can put mp3s on and play them with a player which may be my desktop, small computer in my car,...
Sort of like a portable hard drive. If someone can put something like that out, they can leverage off of that and make many players... Is this done? If not someone needs to "make it so."
Yes it will indeed do wonders - optimally you should have N+1 threads where N = # of processors.
BTW - thanks to the Linux Scalability Project, the 2.3 Linux kernel will perform asynch I/O very efficiently. Netscape is 100% responsible for this - their imap server uses the same asynch. architecture so they patched the kernel for their imap server under the guise of this project.
This would be true if it weren't for the fact that most JVM's GCs are implemented poorly. You can have the fastest Java code in the world, but the JVM's GC bottleneck will almost always appear when you don't want it.
At my company, scalability is an issue since the largest ISP's you can think of use our product. Most of our servers are written in C++. We have been able to invest time into it over the last 5 years and make them very fast. Other servers are in Java and these servers have serious performance problems. The Java developpers, who are really really good, have spent >1 year trying to improve performance - minimizing allocations, increasing the frequency of the GC thread run, etc.. But still when there are huge spikes of usage, the GC thread can't keep up, the process exhausts memory and core dumps.
For a company like mine, Java currently doesn't cut it. C++ gives more bang for the buck. Sure it may take a _little_ longer to write fast, correct code with C++ but the added runtime performance is worth it in many cases.
There are many issues you should worry about when programming if you want fast code, issues which may be in the implementation details. That's real life.
Maybe a garbage collector could be efficient but current implementations are terriblly slow - a "smidgen slower" is an understatement. Specifically, look at jvm's. They use one GC thread and this is bad when for example there is a huge spike in memory allocation, this thread gets behind and memory is exhausted. This is primarily why Java just does not currently scale. C/C++ with no garbage collector will outperform java any day, esp. when there are spikes in allocations.
I claim that there is no need for a garbage collector. Finding memory leaks quickly is not a problem _if_ you have the right tools. Purify is good but it is too slow. There are tools I have developped which can show with pinpoint accuracy where the leaks are and have no performance penalty.
So if you are able to find leaks quickly, with no performance pentalty, then why would you want a garbage collector? I am still unconvinced.