You must be either joking or trolling with your comment about RH.
Not at all. RedHat adds or modifies particular functions within the standard Linux kernel. It was several years ago when I researched this but there's no joking or trolling involved. My memory is hazy (and they may have stopped this practise now) but I believe they either modified the call to fsync() or added a variant of it. As I said, it was several years ago that I did this research.
How can tweaking a kernel lock you into a distro?
Because you can't upgrade the kernel from the standard kernel tree without patching the added/modified functionality in. IIRC there isn't another distro which does this.
Why don't you install apps from source if you don't like rpms? But I'm curious, how do you uninstall if the 'make install' drops files all over the filesystems? Lots of software doesn't include a 'make uninstall' afaik.
I do install from source. However if you use regular tarballs (i.e. there isn't an RPM or SRPM, you now have two package managers to worry about: RPM and the one in your head. Any distro based on packages has this fault. If you're using a package-based distro, chances are that you don't want to worry about this. You want the ease of installation and maintenance. Anything they don't provide, you become the maintainer for.
As far as 'make install' dropping files all over the network: It's called Checkinstall, and it is a godsend. It basically preloads a library to wrap normal fs functions which audit the install process so nothing is missed. Works for Slackware, RPM and DEB.
Ever heard of apt? Yes, it works with rpms (and damn fine too, I might add).
Yeah, I've started playing with Debian. It's a damned fine distro, although their exuberance about PAM and tattooing "GNU/" in front of Linux whenever possible doesn't jive with my personal preferences. Packages are great for checking out software, so long as there is a package available for it. (Binary) packages eliminate the need for a compiler suite on a system, which is a good thing. See my comment above about installing from source and its issues on package-laden distros. It takes some getting used to, but trusting dpkg works for the most part. Where it fails is when the package maintainer decides to bundle certain apps that really are not needed for their package (kde is a prime example of this).
Those examples I gave were from my own personal "been there, done that" file. The tips you give work as a general rule, but they don't ensure happiness and prosperity in the market, in my opinion.
Most customers are pretty straightforward and easy to work for. I would say above 80% of them fall in to this category. The last 20%, however, set out to make your life miserable. Warning them does nothing, because after the work is done and the bill is presented, they grumble about how it took so long and it shouldn't have, and no amount of warning them beforehand, or reminding them that they were warned does any good whatsoever. They will stand and argue for an hour or more. What to do? Write them off and blacklist them? Refuse to return their computer? Refuse to give the BIOS boot password until payment clears? I don't know. Customer service sucks in any sector.
In the specific case of the upgrade that took down NT4 (sorry it was NT4 not W2K) -- there was no warning in the driver documentation, nor in W2K's documentation -- I found it later buried in an obscure readme file that only Google found on Compaq's website. I remember it better now: it was upgrading from SP5 to SP6 -- it completely trashed the computer's ability to boot, and the 'uninstall' for the service pack didn't. Ghost would have helped, but let's be honest here -- do you ghost every system before you try to work on it, just in case? Especially when reading through all the warnings and seeing nothing?
I think that is why I steer clear of Win32 service work in general: Under Linux, I can get to the bottom of it and fix it; There is nothing stopping me or blocking me. With Windows, there's always a layer or three of secrecy and hidden gotchas, no matter how long you've been in the business. I don't like trying to fix things that are out to make sure I can't fix them without an MSDN subscription and a thousand man-hours of bookmarks and experiences about things that can go wrong to have a 95% success rate. I hate telling the customer that they have to reinstall "because that's just windows." I'm there to solve problems, not band-aid them.
Well, thank you very much.:-) Nice to be appreciated.
Believe me, I know how hard that particular industry works and what's stacked against them. I don't envy you guys one bit.
A simple, fast on-site computer help/repair hotline in a reasonably populated area would make you a millionaire, if set up properly.
Wrong.
You run in to the "$foo worked before you fixed $bar, so you should now fix $foo for free since it was your fault." You run in to the "You said it'd work if I did this, but that busted this that and the other" and you're now in for free work. You run in to the "I just need to update driver $foo. So you do and Win2k now bluescreens and you can't get it booted again, and realize that the guy's SCSI drives don't work no more, and you now have 8 hours in to a fix that you can only bill for 2" scenario. These are all real-world examples from working at a computer store.
Working for the general public is a whore's job, no matter what delusions of wealth you assign it. The people who work that sector make every penny.
Hrm - I've had my RH8 desktop for a few weeks now. I can't get NVidia drivers for it. They don't make RH8 drivers specifically, and the source requires that I have megs of kernel source on my drive. Yes, I could put it on, but haven't the time right now. Oh, wouldn't it be great for a binary version of a program IN ADDITION TO SOURCE CODE, so that those of us more interested in being productive than 'freedom' could just get work done?
Two things I can see right off the bat:
You're using RH -- custom RH kernels
You don't need the kernel source
I've always disliked RedHat for two things: RPM hell and custom kernels. Why the hell do distros think that they need to tweak the standard kernel in little ways to lock people in to their distro?
And you don't need the source -- you just need the kernel headers. The NV drivers really aren't all that tricky to compile and install. Hell I even did it under Debian, but then again my distro of choice is Slackware, as in "you learn everything you ever need to know about linux if you run Slackware," Slackware. LFS teaches you a little more, but Slackware is still king, IMO.
Your comment about Windows packages is woefully inaccurate -- try installing a windows program from SOURCE -- there is a big big difference and (IME) much more hassle involved in doing it under Linux. Win32 problems include, but are not limited to: Did you have the deps... were they installed in the same place... are you using the same development environment. The same version of the dev environment... the same OS version... Linux problems usually entail "do you have the deps, are you using a sane compiler (i.e. not what RH ships), can you type "./configure --help". That's really about it these days. The NV stuff is even simpler, since there are no deps outside of the kernel headers. make && make install. Done.
And there are thousands of uninstallable, ununinstallable and installable-but-don't-work-anyway shareware/freeware apps. Apps written for win95 that don't work in win98, 2k or XP. Apps that work in 2k but not 95. Apps that only work with win98osr2. Apps that only work if installed after some other app was installed. The list is endless. Compare that with a decent package-based distro like Debian. Yes, dselect and aptitude suck ass, but the raw system works great. apt-get install packagename. done. deps and all.
If your grandmother is anything like my grandmother she is hesitant about opening her email from the taskbar instead of through the start menu, let alone installing software. And in the case of linux, sudo and a shell script would work for you, too.
user: something about not enough memory. user: But I just want to save the file but it won't let me. I don't even have anything else open.
me: have you tried rebooting?
How, praytell, will you solve their problem? The system says it can't save due to memory, not "access denied." And secondly, how will rebooting help them save their file?:-)
Yes, I work through win32<-->linux problems, too. It's not magic. Hell, it's not even rocket science.
You don't have fancy uptime stats out of the box on Linux, no, but I've never found much use for them except for bragging rights. I get a fuck of a lot of work done under Linux, and I'm sure you get tons of it done under Win2k -- what's your point?
And for the record, I find that I am far more productive with Linux than I ever was with Win32. I am lucky enough to have most of the work I need to do able to be done under Linux, but then again, so does most of the world. The problem I see is that they don't realize that they can be just as productive without paying the Microsoft tax and existing within the upgrade cycle that Wintel places them in.
I'm running 2.4.19 compiled with gcc 3.1 and I haven't had a problem. What I do have is a 42 day uptime
Same kernel, same compiler (maybe a 3.1-pre though): 85 day uptime. And my notebook's been running 2.4.19 and 2.4.20-pre definately compiled with 3.0.4 with no troubles.
I am pretty sure that all the bad things related to gcc3 were in C++, not C.
The answer lies in the datasheets. Many RTC chips have more than 128 bytes of NVRAM, and use a bank swapping technique to select the alternate banks. Dallas Semiconductor is one such source, although most new systems have the RTC embedded in one of the main ICs of the chipset or on the super IO chip. Here is one example, the SMC FDC37N958FR, which is used in the Dauphin Orasis v1, an SBC I am experimenting with putting LinuxBIOS on to get around certain limitations. Page 215 is the start of the RTC/NVRAM access. This particular device has 256 bytes of NVRAM, several of which are reserved for the RTC and 8051 scratchpad. Since there is only 256 bytes, there is no bank switching. The DS1251 is an RTC/NVRAM chip with 512KB of static RAM. The little bitch is expensive, too.:-) It uses a banking method where one of the normally user-available registers is now a bank selection register.
I seriously doubt many systems have more than 256 bytes of NVRAM. That is a LOT of space for settings.
Those who abuse access are in part responsible for higher prices for all users, much like insurance fraud leads to unnecessarily high premiums for all policy holders.
Funny, I thought it was the insurance companies' notion that they have a right to a profit that continually drove up premiums.
Your notion of IT being cheap on Linux is very wrong. In fact if not properly implemented you will end up investing a lot on IT, just for the simple reason that you need linux admins who are good (considering that it is for govt). Even then administring linux is not as simple as windows.
Um, "not properly implemented," ANY IT infrastructure will cost a lot. That's not specific to Linux. Just because Win32 is simple get running quickly doesn't mean you don't end up paying a lot to keep it running smoothly. You need just as competent an admin for proper Win32 administration as you do for Linux administration. The difference, I think, is that Windows gives the illusion that it is easy to be a master admin, which it certainly is not. This, in my opinion, leads to many more cases of "well he looked like he knew what he was up to, now we're going to have to spend $x to fix it right" in the Win32 world (especially with Access or SQL Server), than you ever find in Unix.
Yes, but it's fundamentally different from XWT. I would not choose PicoGUI for an intranet. Hell, I don't think I'd choose PicoGUI for an embedded project. They decided to use their own protocol for server/client communications, which is something I don't understand/agree with. It's YAP - Yet Another Protocol. Yes, XMLRPC would have a larger overhead in terms of larger message size and a mini XML parser, but they could have used straight HTTP GET/POST, CORBA or something smaller and already out.
I also can't give PicoGUI to a web monkey and get them to design impressive widgets and therefore apps with it.
I'll stick to XWT, thanks.:-) And for embedded I would probably choose microwindows or maybe even fresco. The integer-only aspect of PicoGUI is nice but the reality is that anything that has a display requiring a more complex GUI will have enough horsepower to drive one of the bigger windowing toolkits.
This is a good idea, and maybe might be extended to the other TLDs: kids.com, kids.net, etc.
It's a stopgap measure at best; it's not a good idea, it's a good idea implemented very wrong.
All you're doing is balkanizing the internet -- I'd have MUCH rather given out.xxx or.adult and lock kids out, than give them their own section and lock them in. There is a huge difference there.
It's a money grab, it's a shmoo. If you want howthingswork.com to be accessible in.kids.us, you now pay another registration fee. If I want my domain in.kids.us, same thing. You're locking kids out of an enormous resource.
Then again, this is far easier to do than booting all the pr0n/goatse sites into.xxx or.adult. Maybe with ipv6 we can be a little smarter and dole out adult ips to 6969:6969:6969 or sommat.:-)
Why not just get a small pack of blank business card sized CDRs [cardiscs.com] and throw whatever bootable OS you want on it. Carry the OS and whatever other files you want in your wallet. Works for me, and most current BIOS's will at least boot via CD
Yeah, and get it lost in the tray-style CD-ROM drives because it happened to have stopped spinning in just the right direction.:-(
Using publicly available tools you can single-step execution of the Windows operating system. You can get the names of all the symbols in the kernel. You can set kernel mode breakpoints on any peice of the network (and rpc/app) stack you want. It's not the same as having the source, but you can isolate exactly what the OS is doing at any time if you want to. And you can just disassemble the body of any function you like, once you've isolated it (which is easy, since Microsoft publishes full symbol information).
Disclaimer: I am adept at reverse-engineering and debugging, at least on ix86 and microchip PIC.
Which would you prefer: single-stepping or getting an assembly dump of a kernel call, even with function names, or looking at the C that produced it, along with the author's comments and variable names?
Which would you prefer: a multi-megabyte assembly dump (with function names, of course), or the proper source code with variable names, thoughts and ideas in comment blocks and useful changelogs?
I don't give a rat's ass if I can get the assembly dump from kernel32 with debugging symbols: it's still a nightmare to figure out why it's doing something in a certain manner, or even figure out what it's doing at all (i.e. initializing drivers, different algorithms for speed optimizations, etc.) -- without the source. Sure it's possible but it's a hell of a lot harder.
So now, you present your contracted and capable security analysist with two things: the Linux source tree and a disassembled version of the NT/2k kernel. Which do you think is going to be a lot cheaper to have him go through?
How the hell do they aim a highly directional signal like that?
I mean you can get your location pretty easily with GPS -- do they just do some basic navigational math, use a compass and then fine-tune?
Incidentally, is that also how they aim a directional antenna on a 25' post on top of a tower? Figure out the elevation for each point, set the inclination on the antennas and then hoist them on top of the tower and rotate the post until they find the signal?
Why can't Busybox be used for regular, 24/7 server use?
While busybox is great, it does not use many GNU command line options, and many features are missing from the utilities in addition to that. awk, sed, grep -- none of these will work at all for autoconf-style development. tar has some options missing, I had some strange problems with vi (can't backspace?!) and find is missing almost every option -- my CF-based firewalls use GNU's findutils, grep and vim (tiny version) instead of the busybox builtins.
How much space saving? Well, at my work, we initially prototyped some programs that ended up at around 1 MByte, statically linked to glibc. The same program was 120K after statically linking to uclibc, and then 35K after dynamic linking to uclibc.
Yup, uClibc is pretty sweet. I use it and busybox for my CF firewall boxes. 2.4 kernel, ipsec, iptables, Perl, an XMLRPC daemon, IDS and some duct tape scripts in just over 8M. Perl is taking up most of that but if I need to bust out some complex scripting to get something done, it's there.
whoever said you need alot of computing power for a server is wrong
Not entirely true, but that's close to what I believe.
I had an old P5-233/64M/SCSI system for an office fileserver. Upgraded the SCSI system to UW2 hardware RAID5 and had all manner of problem. Turned out that the TX chipset couldn't handle the newer PCI bus master. There's now a P2-233 in there but the bandwidth utilization graphs seem to indicate that the CPU is still the bottleneck (the LAN isn't saturated, and the sustained disk I/O could bury the LAN several times over). I think samba is having trouble keeping numerous smaller feeds open than one or two big pipes. Oh well.:-)
Look at this chart [nasdaq.com] showing Boeing's stock price, and you tell me if their business has gotten bad.
See, that is what I don't understand about investors. What the fuck did Boeing do wrong to cause investors to dump it? Where they worried about a lawsuit naming Boeing as being responsible?
What aroused my curiosity is that I know the military switched from using infrared technology to infragreen for night vision because it worked much better.
You and I are running servers and gaming systems; we want pure performance, and don't mind rebooting a couple of times or recompiling a kernel to change hardware or upgrade drivers. In contrast, my mom has trouble right-clicking My Computer and choosing Properties to get a driver list. For her, a layered driver system that can dynamically load and unload drivers as needed and layer itself against instability is a MAJOR plus, even if it's not for me.
Linux can dynamically load and unload drivers with ease. It's been able to do this since the 2.0.x kernels, IIRC. Hell you can compile modules and install them without updating the kernel tree in many cases. And kernel OOPSes in the stable trees have been few and far between, to say the least.
Yes, microkernels are nifty but there isn't a whole hell of a lot one can do that a modular monolithic kernel can't. I hear there's even work to make the networking stack itself a module. Yer mom isn't going to notice any difference between a microkernel and the hybrid kernel that Linux offers.
Slashdot Mirror
You must be either joking or trolling with your comment about RH.
Not at all. RedHat adds or modifies particular functions within the standard Linux kernel. It was several years ago when I researched this but there's no joking or trolling involved. My memory is hazy (and they may have stopped this practise now) but I believe they either modified the call to fsync() or added a variant of it. As I said, it was several years ago that I did this research.
How can tweaking a kernel lock you into a distro?
Because you can't upgrade the kernel from the standard kernel tree without patching the added/modified functionality in. IIRC there isn't another distro which does this.
Why don't you install apps from source if you don't like rpms? But I'm curious, how do you uninstall if the 'make install' drops files all over the filesystems? Lots of software doesn't include a 'make uninstall' afaik.
I do install from source. However if you use regular tarballs (i.e. there isn't an RPM or SRPM, you now have two package managers to worry about: RPM and the one in your head. Any distro based on packages has this fault. If you're using a package-based distro, chances are that you don't want to worry about this. You want the ease of installation and maintenance. Anything they don't provide, you become the maintainer for.
As far as 'make install' dropping files all over the network: It's called Checkinstall, and it is a godsend. It basically preloads a library to wrap normal fs functions which audit the install process so nothing is missed. Works for Slackware, RPM and DEB.
Ever heard of apt? Yes, it works with rpms (and damn fine too, I might add).
Yeah, I've started playing with Debian. It's a damned fine distro, although their exuberance about PAM and tattooing "GNU/" in front of Linux whenever possible doesn't jive with my personal preferences. Packages are great for checking out software, so long as there is a package available for it. (Binary) packages eliminate the need for a compiler suite on a system, which is a good thing. See my comment above about installing from source and its issues on package-laden distros. It takes some getting used to, but trusting dpkg works for the most part. Where it fails is when the package maintainer decides to bundle certain apps that really are not needed for their package (kde is a prime example of this).
Secrets from the "been there, done that" file:
Those examples I gave were from my own personal "been there, done that" file. The tips you give work as a general rule, but they don't ensure happiness and prosperity in the market, in my opinion.
Most customers are pretty straightforward and easy to work for. I would say above 80% of them fall in to this category. The last 20%, however, set out to make your life miserable. Warning them does nothing, because after the work is done and the bill is presented, they grumble about how it took so long and it shouldn't have, and no amount of warning them beforehand, or reminding them that they were warned does any good whatsoever. They will stand and argue for an hour or more. What to do? Write them off and blacklist them? Refuse to return their computer? Refuse to give the BIOS boot password until payment clears? I don't know. Customer service sucks in any sector.
In the specific case of the upgrade that took down NT4 (sorry it was NT4 not W2K) -- there was no warning in the driver documentation, nor in W2K's documentation -- I found it later buried in an obscure readme file that only Google found on Compaq's website. I remember it better now: it was upgrading from SP5 to SP6 -- it completely trashed the computer's ability to boot, and the 'uninstall' for the service pack didn't. Ghost would have helped, but let's be honest here -- do you ghost every system before you try to work on it, just in case? Especially when reading through all the warnings and seeing nothing?
I think that is why I steer clear of Win32 service work in general: Under Linux, I can get to the bottom of it and fix it; There is nothing stopping me or blocking me. With Windows, there's always a layer or three of secrecy and hidden gotchas, no matter how long you've been in the business. I don't like trying to fix things that are out to make sure I can't fix them without an MSDN subscription and a thousand man-hours of bookmarks and experiences about things that can go wrong to have a 95% success rate. I hate telling the customer that they have to reinstall "because that's just windows." I'm there to solve problems, not band-aid them.
Well, thank you very much. :-) Nice to be appreciated.
Believe me, I know how hard that particular industry works and what's stacked against them. I don't envy you guys one bit.
Working for the general public is a whore's job, no matter what delusions of wealth you assign it. The people who work that sector make every penny.
Earn every penny. ugh.
Netmeeting? My god man, get yourself a real remote desktop utility. :-)
A simple, fast on-site computer help/repair hotline in a reasonably populated area would make you a millionaire, if set up properly.
Wrong.
You run in to the "$foo worked before you fixed $bar, so you should now fix $foo for free since it was your fault." You run in to the "You said it'd work if I did this, but that busted this that and the other" and you're now in for free work. You run in to the "I just need to update driver $foo. So you do and Win2k now bluescreens and you can't get it booted again, and realize that the guy's SCSI drives don't work no more, and you now have 8 hours in to a fix that you can only bill for 2" scenario. These are all real-world examples from working at a computer store.
Working for the general public is a whore's job, no matter what delusions of wealth you assign it. The people who work that sector make every penny.
Hrm - I've had my RH8 desktop for a few weeks now. I can't get NVidia drivers for it. They don't make RH8 drivers specifically, and the source requires that I have megs of kernel source on my drive. Yes, I could put it on, but haven't the time right now. Oh, wouldn't it be great for a binary version of a program IN ADDITION TO SOURCE CODE, so that those of us more interested in being productive than 'freedom' could just get work done?
Two things I can see right off the bat:
I've always disliked RedHat for two things: RPM hell and custom kernels. Why the hell do distros think that they need to tweak the standard kernel in little ways to lock people in to their distro?
And you don't need the source -- you just need the kernel headers. The NV drivers really aren't all that tricky to compile and install. Hell I even did it under Debian, but then again my distro of choice is Slackware, as in "you learn everything you ever need to know about linux if you run Slackware," Slackware. LFS teaches you a little more, but Slackware is still king, IMO.
Your comment about Windows packages is woefully inaccurate -- try installing a windows program from SOURCE -- there is a big big difference and (IME) much more hassle involved in doing it under Linux. Win32 problems include, but are not limited to: Did you have the deps... were they installed in the same place... are you using the same development environment. The same version of the dev environment... the same OS version... Linux problems usually entail "do you have the deps, are you using a sane compiler (i.e. not what RH ships), can you type "./configure --help". That's really about it these days. The NV stuff is even simpler, since there are no deps outside of the kernel headers. make && make install. Done.
And there are thousands of uninstallable, ununinstallable and installable-but-don't-work-anyway shareware/freeware apps. Apps written for win95 that don't work in win98, 2k or XP. Apps that work in 2k but not 95. Apps that only work with win98osr2. Apps that only work if installed after some other app was installed. The list is endless. Compare that with a decent package-based distro like Debian. Yes, dselect and aptitude suck ass, but the raw system works great. apt-get install packagename. done. deps and all.
If your grandmother is anything like my grandmother she is hesitant about opening her email from the taskbar instead of through the start menu, let alone installing software. And in the case of linux, sudo and a shell script would work for you, too.
- user: something about not enough memory.
me: have you tried rebooting?user: But I just want to save the file but it won't let me. I don't even have anything else open.
How, praytell, will you solve their problem? The system says it can't save due to memory, not "access denied." And secondly, how will rebooting help them save their file? :-)
Yes, I work through win32<-->linux problems, too. It's not magic. Hell, it's not even rocket science.
Yippee, so your systems stay up. I can do the same thing with my linux boxes:
mail/radius/postgres box (authenticates and accounts for all NAS ports across three cities (over 1000 ports)):
colocation box (~48 colocated web/mail hosts):
50-user fileserver (HW RAID5 UW2):
hmm, time's a bit off on that one.
You don't have fancy uptime stats out of the box on Linux, no, but I've never found much use for them except for bragging rights. I get a fuck of a lot of work done under Linux, and I'm sure you get tons of it done under Win2k -- what's your point?
And for the record, I find that I am far more productive with Linux than I ever was with Win32. I am lucky enough to have most of the work I need to do able to be done under Linux, but then again, so does most of the world. The problem I see is that they don't realize that they can be just as productive without paying the Microsoft tax and existing within the upgrade cycle that Wintel places them in.
I'm running 2.4.19 compiled with gcc 3.1 and I haven't had a problem. What I do have is a 42 day uptime
Same kernel, same compiler (maybe a 3.1-pre though): 85 day uptime. And my notebook's been running 2.4.19 and 2.4.20-pre definately compiled with 3.0.4 with no troubles.
I am pretty sure that all the bad things related to gcc3 were in C++, not C.
Sigh... and I was just thinking that I haven't seen a goatse.cx redirect in a long time. Ass.
The answer lies in the datasheets. Many RTC chips have more than 128 bytes of NVRAM, and use a bank swapping technique to select the alternate banks. Dallas Semiconductor is one such source, although most new systems have the RTC embedded in one of the main ICs of the chipset or on the super IO chip. Here is one example, the SMC FDC37N958FR, which is used in the Dauphin Orasis v1, an SBC I am experimenting with putting LinuxBIOS on to get around certain limitations. Page 215 is the start of the RTC/NVRAM access. This particular device has 256 bytes of NVRAM, several of which are reserved for the RTC and 8051 scratchpad. Since there is only 256 bytes, there is no bank switching. The DS1251 is an RTC/NVRAM chip with 512KB of static RAM. The little bitch is expensive, too. :-) It uses a banking method where one of the normally user-available registers is now a bank selection register.
I seriously doubt many systems have more than 256 bytes of NVRAM. That is a LOT of space for settings.
Those who abuse access are in part responsible for higher prices for all users, much like insurance fraud leads to unnecessarily high premiums for all policy holders.
Funny, I thought it was the insurance companies' notion that they have a right to a profit that continually drove up premiums.
Your notion of IT being cheap on Linux is very wrong. In fact if not properly implemented you will end up investing a lot on IT, just for the simple reason that you need linux admins who are good (considering that it is for govt). Even then administring linux is not as simple as windows.
Um, "not properly implemented," ANY IT infrastructure will cost a lot. That's not specific to Linux. Just because Win32 is simple get running quickly doesn't mean you don't end up paying a lot to keep it running smoothly. You need just as competent an admin for proper Win32 administration as you do for Linux administration. The difference, I think, is that Windows gives the illusion that it is easy to be a master admin, which it certainly is not. This, in my opinion, leads to many more cases of "well he looked like he knew what he was up to, now we're going to have to spend $x to fix it right" in the Win32 world (especially with Access or SQL Server), than you ever find in Unix.
Wow, PicoGUI is pretty impressive.
Yes, but it's fundamentally different from XWT. I would not choose PicoGUI for an intranet. Hell, I don't think I'd choose PicoGUI for an embedded project. They decided to use their own protocol for server/client communications, which is something I don't understand/agree with. It's YAP - Yet Another Protocol. Yes, XMLRPC would have a larger overhead in terms of larger message size and a mini XML parser, but they could have used straight HTTP GET/POST, CORBA or something smaller and already out.
I also can't give PicoGUI to a web monkey and get them to design impressive widgets and therefore apps with it.
I'll stick to XWT, thanks. :-) And for embedded I would probably choose microwindows or maybe even fresco. The integer-only aspect of PicoGUI is nice but the reality is that anything that has a display requiring a more complex GUI will have enough horsepower to drive one of the bigger windowing toolkits.
This is a good idea, and maybe might be extended to the other TLDs: kids.com, kids.net, etc.
It's a stopgap measure at best; it's not a good idea, it's a good idea implemented very wrong.
All you're doing is balkanizing the internet -- I'd have MUCH rather given out .xxx or .adult and lock kids out, than give them their own section and lock them in. There is a huge difference there.
It's a money grab, it's a shmoo. If you want howthingswork.com to be accessible in .kids.us, you now pay another registration fee. If I want my domain in .kids.us, same thing. You're locking kids out of an enormous resource.
Then again, this is far easier to do than booting all the pr0n/goatse sites into .xxx or .adult. Maybe with ipv6 we can be a little smarter and dole out adult ips to 6969:6969:6969 or sommat. :-)
Why not just get a small pack of blank business card sized CDRs [cardiscs.com] and throw whatever bootable OS you want on it. Carry the OS and whatever other files you want in your wallet. Works for me, and most current BIOS's will at least boot via CD
Yeah, and get it lost in the tray-style CD-ROM drives because it happened to have stopped spinning in just the right direction. :-(
Using publicly available tools you can single-step execution of the Windows operating system. You can get the names of all the symbols in the kernel. You can set kernel mode breakpoints on any peice of the network (and rpc/app) stack you want. It's not the same as having the source, but you can isolate exactly what the OS is doing at any time if you want to. And you can just disassemble the body of any function you like, once you've isolated it (which is easy, since Microsoft publishes full symbol information).
Disclaimer: I am adept at reverse-engineering and debugging, at least on ix86 and microchip PIC.
Which would you prefer: single-stepping or getting an assembly dump of a kernel call, even with function names, or looking at the C that produced it, along with the author's comments and variable names?
Which would you prefer: a multi-megabyte assembly dump (with function names, of course), or the proper source code with variable names, thoughts and ideas in comment blocks and useful changelogs?
I don't give a rat's ass if I can get the assembly dump from kernel32 with debugging symbols: it's still a nightmare to figure out why it's doing something in a certain manner, or even figure out what it's doing at all (i.e. initializing drivers, different algorithms for speed optimizations, etc.) -- without the source. Sure it's possible but it's a hell of a lot harder.
So now, you present your contracted and capable security analysist with two things: the Linux source tree and a disassembled version of the NT/2k kernel. Which do you think is going to be a lot cheaper to have him go through?
How the hell do they aim a highly directional signal like that?
I mean you can get your location pretty easily with GPS -- do they just do some basic navigational math, use a compass and then fine-tune?
Incidentally, is that also how they aim a directional antenna on a 25' post on top of a tower? Figure out the elevation for each point, set the inclination on the antennas and then hoist them on top of the tower and rotate the post until they find the signal?
Why can't Busybox be used for regular, 24/7 server use?
While busybox is great, it does not use many GNU command line options, and many features are missing from the utilities in addition to that. awk, sed, grep -- none of these will work at all for autoconf-style development. tar has some options missing, I had some strange problems with vi (can't backspace?!) and find is missing almost every option -- my CF-based firewalls use GNU's findutils, grep and vim (tiny version) instead of the busybox builtins.
How much space saving? Well, at my work, we initially prototyped some programs that ended up at around 1 MByte, statically linked to glibc. The same program was 120K after statically linking to uclibc, and then 35K after dynamic linking to uclibc.
Yup, uClibc is pretty sweet. I use it and busybox for my CF firewall boxes. 2.4 kernel, ipsec, iptables, Perl, an XMLRPC daemon, IDS and some duct tape scripts in just over 8M. Perl is taking up most of that but if I need to bust out some complex scripting to get something done, it's there.
whoever said you need alot of computing power for a server is wrong
Not entirely true, but that's close to what I believe.
I had an old P5-233/64M/SCSI system for an office fileserver. Upgraded the SCSI system to UW2 hardware RAID5 and had all manner of problem. Turned out that the TX chipset couldn't handle the newer PCI bus master. There's now a P2-233 in there but the bandwidth utilization graphs seem to indicate that the CPU is still the bottleneck (the LAN isn't saturated, and the sustained disk I/O could bury the LAN several times over). I think samba is having trouble keeping numerous smaller feeds open than one or two big pipes. Oh well. :-)
Romeo & Juliet for 1337 hax0rz! http://www.redcoat.net/pics/romjul.swf
That's one of the funniest swfs I've seen in a long time, thanks!
Look at this chart [nasdaq.com] showing Boeing's stock price, and you tell me if their business has gotten bad.
See, that is what I don't understand about investors. What the fuck did Boeing do wrong to cause investors to dump it? Where they worried about a lawsuit naming Boeing as being responsible?
What aroused my curiosity is that I know the military switched from using infrared technology to infragreen for night vision because it worked much better.
I'm sorry, but infragreen? WTF is that?
You and I are running servers and gaming systems; we want pure performance, and don't mind rebooting a couple of times or recompiling a kernel to change hardware or upgrade drivers. In contrast, my mom has trouble right-clicking My Computer and choosing Properties to get a driver list. For her, a layered driver system that can dynamically load and unload drivers as needed and layer itself against instability is a MAJOR plus, even if it's not for me.
Linux can dynamically load and unload drivers with ease. It's been able to do this since the 2.0.x kernels, IIRC. Hell you can compile modules and install them without updating the kernel tree in many cases. And kernel OOPSes in the stable trees have been few and far between, to say the least.
Yes, microkernels are nifty but there isn't a whole hell of a lot one can do that a modular monolithic kernel can't. I hear there's even work to make the networking stack itself a module. Yer mom isn't going to notice any difference between a microkernel and the hybrid kernel that Linux offers.