For all the talk of Linux vs Win2k, when I look at the benchmarks, what I notice immediately is that IBM's DB2 is at the top of most of those lists, and Oracle is not. I know that Oracle is best known for handling extrememly large databases, but isn't it supposed to perform well too? It only shows up once in the list, where the Sun/Oracle box beaten *badly* by a less expensive Intel box!!
Anyway, the reason that struck me is that I'm very interested in DB2. I've heard good things about it, but not nearly as often as I hear about MS-SQL, Oracle, PostgreSQL, and MySQL, all of which are covered in O'Reily's "SQL in a Nutshell". I was only disappointed that this book didn't cover any others. If I wanted to get similar information for DB2, where would I look? It'd be cool to see some DB2-lovin hippies to publish information in the form of an addendum to that book for comparison's (and reference's) sake.
Actually, I meant "any NT admins" when I said "admins". I live up here in the land of Microsoft (Seattle area), and all of the NT admins that I know are NT geeks. They're mostly elitists who only run Win2k at home.
I don't personally run any Windows OS on my home computer, but after using Windows 2000 for about a month, my girlfriend decided that Win98 crashed WAY less, and I reinstalled it for her.
And, BTW, your guess is quite off the mark. I probably touched a nerve, but I'm not calling you an idiot. Relax, guy.:-)
Uh, yeah, sorta. About as amusing as breakins get.:-)
People in general need to stick to what they know best. Bosses generally don't have the required paranoia level required to be an admin. Were I Scott, I don't think I would have even asked that question of my boss... If his boss had the background to make that sort of decision, he'd be an admin.
The previous rant is still wearing off... No one should be offended by this post.
There have been few times when I wished I had moderator points more. Moderators: A post should only be judged insightful if you KNOW that it is. YHBT YHL HAND.
But so have I.
This is not a rant against the parent post, it's a troll. This is a rant against stupid moderators.
setup scripts, and do some last minute checking on up2date daemons OK, besides the fact that up2date is a Red Hat Linux specific feature, connecting any system to the internet without looking at the services that run at startup time and applying errata/service packs is extrememly irresponsible and arrogant. Windows NT is worse than good UNIX systems in this respect, because just about everything is *on* by default, and many of those services are exploitable, or need good hard configuring not to be.
admins all use windows95 on their system at home I don't know ANY admins that use Win95 on their home systems (or 98, or ME). At least none who have one clue what they're doing. The resemblance between Win95 and WinNT is totally superficial. There is nothing even related to administration of an NT server that you can practice on a Win95 box. Win95 doesn't have features as basic as file permissions or services. The only thing you can practice on Win95 is point and click.
I still do not know how to lock the ports below 1024 like redhat linux does Every UNIX I know of does this by default. It's a feature of the kernel, and not something that you have to DO.
NT is quite secure but not really stable And in the breath before, suggesting to visit Bugtraq to check the bugs in each. Look at the number of hacks per OS (I beleive that attrition.org keeps track of that sort of thing), and you will see that NT has a disproportionally large number of breakins. Although they are not the most commonly used servers, they are the most frequently hacked. Repeat after me: security is not a feature, it's a process. Your security lies not in your OS, it's in your admin.
The differences between FAT32 and XFS may be interesting, but keep perspective. What you're seeing is the difference between the Linux driver for XFS and the Linux driver for FAT32, and not necessarily the inherent properties of either filesystem.
Don't get me wrong, I'm not comparing FAT32 to XFS by a long shot! But FAT32 is a fs that not a lot of hackers care about enough to improve the performance under Linux. Personally, I've always found that FAT32 access under Linux has been abysmal compared to access to the same filesystem under Windows.
(according to top, which is known to lie without shame, particularly with memory)
I _do_ wish people would stop saying that. 'top' does NOT lie. Users just don't understand what it's telling them. Your misunderstanding doesn't constitute a lie. The "SIZE" column displays the total size of _everything_ the process is using, including memory on the card that the X server has mmap()'d. If you want to see how much of your system memory is used, use 'ps' and look at the RSS column.
It should be noted that Linux takes longer to switch tasks on the PIII if it was compiled to support SSE2 instructions. Five dollars says that FreeBSD doesn't support them. Perhaps those benchmarks would have showed different results if the kernel had been built without SSE2 support?
Since your follow-ups seem to indicate that you don't think people are getting your message quite right, I'm going to try to be particular about this reply...
The internet is cool stuff. but really, how useful or widespread would it be if not for MS? No less than it is now, and possilbly more. Seriously. I recall using the net *Way* before MS got interested in supporting it, and in those days it was significantly easier to find what you wanted. A higher signal-to-noise ratio, if you will. I believe that qualifies as being more useful than it is now.
As for widespread, well I don't believe the utter bullshit myth that MS brought computing to the masses. MS didn't bring us innovations or ease of use, they stopped anyone else from doing it. PC users from back in the day will tell you about MS superior competitors: Geoworks, DR-DOS, OS/2. Guess what? MS's monopoly is universally attributed with the death of hundreds or thousands of good products. Now, you may notice that over the years every part of the PC has decreased in cost except for the OS/software, which has increased, largly because it's MS and no alternatives are left. In most cases, the software bundle included in a PC is the most expensive component: more expensive than the CPU, more expensive than even the monitor. Expensive enough that a lot of people who can't afford a computer now could afford one if not for the cost of software. I suggest that you consider that MS has actually hindered the spread of home computing, not helped it, and that widespread computer access and use is a reality despite MS, not because of them.
The ability for average non-technical persons to use a personal computer... With only the flavors of Unix, early Linux distro's, and other non-friendly OS's, Internet access would have been rare Now, the post to which you were replying was counting the *contributions* of MS vs. Open IP, so you *are* suggesting that MS is responsible for this here. I don't believe that MS really is responsible for or contributed to the ease of use of computer systems. As a matter of opinion, MS has always been behind its competition in terms of ease of use. Always. It still is. Maybe it's true that with only UNIX, Linux or unfriendly OS's internet access wouldn't be as widespread, but that wasn't all that there was. As mentioned before, MS out-marketed most of those competitors that stood to really bring ease of use and widespread computing to everyone.
The inclusion of web-browsers I hate to repeat arguments, but... again, what leads you to believe that this was a contribution? It was being done before MS did it. Netscape was available with new computers before IE was, but MS killed that because it was allowing customers to grow at a pace and along a course that MS could neither keep up with nor control.
Like it or not, MS filled a serious demand in the computer world. They broke lots of laws, squashed lots people, and did it the closed way, but... I'm not going to argue that MS filled that demand, but the point that has to be made is that the demand would have been filled by someone else if not for MS. Squashing competitors, killing products, and stiffling the growth of an industry isn't something that MS should be lauded for just because they ended up with a lot of cash and customers. And, they may have made more money than any other company, but if their prospertity came at the cost of the jobs and businesses of thousands (easily tens of thousands) of competing developers, so they've cost us as Americans more than they've made us.
Writing MS off because we don't like they way they do business isn't "a serious mistake", it's the responsible course of action.
volume at the expense of clarity, he claims that this volume suffers neither fault.
It's early, and when I read this, I thought he meant that the book had neither volume nor clarity.
Re:Can we please give them the benefit of the doub
on
Sony Violating GPL?
·
· Score: 2
Looking at the new emulator, it was just posted 2 days ago (5/1/01). My guess is that they'll release the source real soon now, but it probably wasn't quite ready for prime time just yet.
The GPL doesn't say anything about giving them a few days. If the product was ready for release, then it should *all* be ready, and if it's not, then nothing should have been released. Even if they release the source soon, they require users to agree to a EULA which is not compatible with the GPL, and is not the license for that software, which is illegal. I do believe that Corel did that with their Linux distro not so very long ago. The community reminded them of their error, and they fixed it.
The problem is that several companies don't seem to understand that the GPL is a license, and that its terms are every bit as binding as any license that they pay for. If they need reminders, we've got plenty of them.
I think your math is off, and doubt that you've dealt with MS servers recently. Let me give you my take on the situation:
- Lets say your "entrepeneur" wants support for his Linux, so he buys Red Hat Linux Professional Server. For $180 bucks he can provide file sharing, printer sharing, domain logons for Windows, mail service, etc for all of the users in his office.
- Windows 2000 professional might make a fine desktop (still crashes too often for my taste), but it will in no way provide you with a server. For that you will have to buy Windows 2000 Server with client licenses for each user. If this is a *really* small office, you might get away with 10 CAL's, and pay around $1000. You're a lot more likely to need 25 for even a small office and pay closer to $3000. On top of that, you still have to buy Exchange for email, and probably MS SBS, too. Throw in a few thousand more dollars. The cost difference for the OS is not insignificant. It could easily pay for several visits from a consultant.
- IBM's SBS isn't free, but neither is Microsoft's.
- A linux box can be set up in short order by almost any technically apt person. You probably have one on staff if you work with any number of computers, but you may not if that's not your feild. So you hire a consultant to set it up. Once done, Linux installations rarely need baby sitting. A configured box can usually be left on its own to do its job.
- Windows is not easy to "use and administer", and if you think it is, you haven't used MS SBS. For instance, adding users in SBS will create users with a default login script. However, that login script will be overwritten every time you add a user. Managers want to use MS SBS because (this is a quote from a customer) "I don't have to do things the NT way, this is simpler". What's that I hear? The NT way is too complicated? However, when working for this particular customer, we had to do most things the "NT way" because the SBS way screwed them up. Overwriting the default login script meant loosing any local changes they'd made. The only option was to add users with the standard NT user manager. I have other examples, but listing them would take so long, I'd never finish this post. And as an aside... I've *never* seen an RPM install hose a system, so I don't know what you're referring to in your "user error" comment.
- Whether Windows or Linux, your need for a dedicated admin will depend on the number of users and tasks handled by the system. It's about the same for both, and most small businesses can get by with outsourced support. The difference is that Windows will need it more often. Having serviced customers with Windows and Linux boxen, I am qualified to make that statement, and I have the experience (as opposed to your speculation) to back it up.
- It's easier and faster for me to install PostgreSQL or MySQL on Linux than MSSQL on Windows. And don't forget that MS's default installation was responsible for *many* hacks in the last several months, including Wells Fargo.
In summary, people don't buy and use MS because it's better. They buy and use MS because of FUD like yours. Them that have consultants that they *trust* frequently choose Linux.
My guess is that with your change things appear faster, but in the end you've actually hurt your overall throughput (the amount of computational work you actually get done in a given time).
You almost got that one right. IIRC, increasing the time slice increases throughput, the price you pay is increased latency.
You also have failed to realize that perception is reality. If it seems faster, then it IS faster.
First, I feel obligated to state that I think Python is totally tits. I love it. Thank you.
Reading "What's New in Python 2.1", I'm curious about nested scopes. Given that name lookup has always been the cause of poor performance in Python (or so I've heard), it would initially seem that introducing nested scopes would further reduce the speed at which Python scripts run if you use this feature.
Is there any information available about Python's performance with this addition?
If you guys are open to including an open source ksh, I think it'd make a lot of people very happy.
David Korn recently did a slashdot interview, and his answer to the first question should best express why an authentic ksh should replace the pdksh included now. See this document.
But, calling it an industry standard is probably taking that a bit too far.
Says who? Linux is shipping on more servers than any other UNIX variant. More software is being developed on Linux than on any other UNIX variant.
No, it's not the standard PC platform, but the previous comment doesn't indicate that it was... He just said that Linux is the industry standard UNIX. It may not be the most powerful or mature, but from a lot of people's point of view, it's the standard.
When Red Hat 7.0 was released, there was NO known exploit for that LPD. It was discovered weeks after the release of 7.0. As always, errata was released in a timely fashion.
Red Hat's gcc 2.96 was hardly useless. I've used it to build everything on my system, and it's been totally stable. The only thing that 2.96 wouldn't build correctly was the 2.2 kernel. Guess what, that's not a new problem! IIRC, egcs also wouldn't compile kernels correctly. Assembly code isn't defined as well as, say, C. egcs broke the previous assembler semantics, and the kernel had to be fixed. This is pretty much the same issue with 2.96. The 2.4 kernel has semantics that the 2.96 compiler will accept, and the new compiler is officially used to build kernels now.
C++ support for GCC has never been that good for exactly those reasons. If Red Hat had used any version of GCC other than egcs 1.1.2 (from 6.2) their C++ binaries would not have been compatible with the older, 6.x versions or the 8.x series which will most likey include gcc 3.0.
Based on that unavoidable problem, and their need to support the Alpha platform, Red Hat's engineers decided to use a version of gcc from CVS, and have done a lot of work to make sure that it's stable. Red Hat 7.0 has been rock steady for me. None of the components have suffered because of gcc 2.96rh. It produces stable binaries. Reports of Red Hat's demise are greatly exaggerated.
But, it appears (see other post) the consensus is to downgrade GCC
This consensus is generally reached by the uninformed. You will only cause yourself headache by doing this.
Know of any place I could look for more information? Like how the server decides what programs are given privileges and how those programs request privileges from the server?
The limitation that you're running into is this: you can't SUID a library in UNIX. I understand that you could do this in Multics, but I believe that it was never a feature of UNIX since it's too dangerous. Especially in a threaded application, one could do all kinds of nasty things while the library's running to get unauthorized access. (It might actually work if the library ran as a separate process with its own resources, but I'm going way off topic. Bring this up on a kernel development list and see what kind of responce you get.)
In order for PAM to access any password database, including the shadow file, the application running must be able to access that database. If you don't care about non root users accessing your password database, then you shouldn't be using shadow files. There are other things to consider as well.
Authentication isn't the only reason that a server needs to run as root. A typical login sequence for an interactive daemon works like:
* daemon starts as root
* bind to port (may require root, depending on service)
* accept connection
* read username and password
* compare username and password to authentication db (probably requires root)
* drop root privileges, switching to the authenticated user (definitely requires root)
The last step is an important one. If you want your daemons to be able to switch UID, they have to be run as root. The alternative is to run the imap server as a non-root user with access to the password database, or proxy the username and password to an external authentication daemon, but then all of the mailboxes must be readable and writable by the imap user. It can be done, but then each daemon needs its own hack to get everything working.
Since you can't switch UID's without root, and you can't drop root until after authentication, you're stuck doing authentication as root. You would *think* that this would be a small enough section of code that it wouldn't be commonly exploitable, but it is all the same. </me sighs>
We fully support data, directory and system interop with UNIX, Linux, Novell, Mac, IBM mainframes through our base OS protocol support as well as through products like Services for UNIX, Interix, Services for NetWare, MetaDirectory and Host Integration Server.
Well, I'm glad this was first so we can get this out of the way. Doug defends interoperability by stating that MS offers products that will provide it. However, their OS's alone provide interoperability only with other, similar (read: MS) products. If you want interoperability with, say, a UNIX system, you have to pay for additional software licenses. Their customers are criticizing them not because interoperability isn't available, but because they feel like they have to pay for MS licenses for every machine they run, whether it actually uses MS software or not.
Further, Dave goes on to defend MS tactics of 'Embrace and (Extend|Extinguish)' as protecting their IP. If we consider MS's undocumented (for our intents and purposes) extentions to Kerberos, we can clearly see why customers complain about 'vendor lock-in'. In order for a Kerberos ticket to be valid in a Windows domain network, it must contain the additional information that a Windows server will place there. That's valid. That's why there's a chunk of a Kerberos ticket left undefined by the standard. However, because that information isn't documented, only a Windows Kerberos server can be used for authentication. That is not interoperability. That does not give your customers the ability to choose the best solution for the job.
If your ISP used NT servers as terminal servers (as opposed to Portmasters, MAX's, Cisco, etc.), then they *would* have to pay for a client license for every line that would be in use at max capacity. If they're just running NT as a Radius server, then they only have to pay for client licenses for each of the terminal servers that connect to them. An ISP using NT would normally have to pay fairly hefty client license fees. Figure out how many users will be checking their mail, how many web pages are being viewed and how many radius clients are connecting simultaniously at maximum capacity, and pay client licenses for all of them.
In the Alaska Airlines situation you describe, the clients in question are connecting directly to the NT servers and using their resources. According to MS, that means they have to pay for client licenses.
Slashdot Mirror
For all the talk of Linux vs Win2k, when I look at the benchmarks, what I notice immediately is that IBM's DB2 is at the top of most of those lists, and Oracle is not. I know that Oracle is best known for handling extrememly large databases, but isn't it supposed to perform well too? It only shows up once in the list, where the Sun/Oracle box beaten *badly* by a less expensive Intel box!!
Anyway, the reason that struck me is that I'm very interested in DB2. I've heard good things about it, but not nearly as often as I hear about MS-SQL, Oracle, PostgreSQL, and MySQL, all of which are covered in O'Reily's "SQL in a Nutshell". I was only disappointed that this book didn't cover any others. If I wanted to get similar information for DB2, where would I look? It'd be cool to see some DB2-lovin hippies to publish information in the form of an addendum to that book for comparison's (and reference's) sake.
You Have Been Tolled. You Have Lost. Have A Nice Day.
Actually, I meant "any NT admins" when I said "admins". I live up here in the land of Microsoft (Seattle area), and all of the NT admins that I know are NT geeks. They're mostly elitists who only run Win2k at home.
:-)
I don't personally run any Windows OS on my home computer, but after using Windows 2000 for about a month, my girlfriend decided that Win98 crashed WAY less, and I reinstalled it for her.
And, BTW, your guess is quite off the mark. I probably touched a nerve, but I'm not calling you an idiot. Relax, guy.
Uh, yeah, sorta. About as amusing as breakins get. :-)
People in general need to stick to what they know best. Bosses generally don't have the required paranoia level required to be an admin. Were I Scott, I don't think I would have even asked that question of my boss... If his boss had the background to make that sort of decision, he'd be an admin.
The previous rant is still wearing off... No one should be offended by this post.
There have been few times when I wished I had moderator points more. Moderators: A post should only be judged insightful if you KNOW that it is. YHBT YHL HAND.
But so have I.
This is not a rant against the parent post, it's a troll. This is a rant against stupid moderators.
setup scripts, and do some last minute checking on up2date daemons
OK, besides the fact that up2date is a Red Hat Linux specific feature, connecting any system to the internet without looking at the services that run at startup time and applying errata/service packs is extrememly irresponsible and arrogant. Windows NT is worse than good UNIX systems in this respect, because just about everything is *on* by default, and many of those services are exploitable, or need good hard configuring not to be.
admins all use windows95 on their system at home
I don't know ANY admins that use Win95 on their home systems (or 98, or ME). At least none who have one clue what they're doing. The resemblance between Win95 and WinNT is totally superficial. There is nothing even related to administration of an NT server that you can practice on a Win95 box. Win95 doesn't have features as basic as file permissions or services. The only thing you can practice on Win95 is point and click.
I still do not know how to lock the ports below 1024 like redhat linux does
Every UNIX I know of does this by default. It's a feature of the kernel, and not something that you have to DO.
NT is quite secure but not really stable
And in the breath before, suggesting to visit Bugtraq to check the bugs in each. Look at the number of hacks per OS (I beleive that attrition.org keeps track of that sort of thing), and you will see that NT has a disproportionally large number of breakins. Although they are not the most commonly used servers, they are the most frequently hacked. Repeat after me: security is not a feature, it's a process. Your security lies not in your OS, it's in your admin.
The differences between FAT32 and XFS may be interesting, but keep perspective. What you're seeing is the difference between the Linux driver for XFS and the Linux driver for FAT32, and not necessarily the inherent properties of either filesystem.
Don't get me wrong, I'm not comparing FAT32 to XFS by a long shot! But FAT32 is a fs that not a lot of hackers care about enough to improve the performance under Linux. Personally, I've always found that FAT32 access under Linux has been abysmal compared to access to the same filesystem under Windows.
Rule #4:
One can have a graph of any shape that he wants by carefully choosing the axis'.
(according to top, which is known to lie without shame, particularly with memory)
I _do_ wish people would stop saying that. 'top' does NOT lie. Users just don't understand what it's telling them. Your misunderstanding doesn't constitute a lie. The "SIZE" column displays the total size of _everything_ the process is using, including memory on the card that the X server has mmap()'d. If you want to see how much of your system memory is used, use 'ps' and look at the RSS column.
It should be noted that Linux takes longer to switch tasks on the PIII if it was compiled to support SSE2 instructions. Five dollars says that FreeBSD doesn't support them. Perhaps those benchmarks would have showed different results if the kernel had been built without SSE2 support?
- 05 -03-007-20-NW-KN
http://linuxtoday.com/news_story.php3?ltsn=2001
Since your follow-ups seem to indicate that you don't think people are getting your message quite right, I'm going to try to be particular about this reply...
The internet is cool stuff. but really, how useful or widespread would it be if not for MS?
No less than it is now, and possilbly more. Seriously. I recall using the net *Way* before MS got interested in supporting it, and in those days it was significantly easier to find what you wanted. A higher signal-to-noise ratio, if you will. I believe that qualifies as being more useful than it is now.
As for widespread, well I don't believe the utter bullshit myth that MS brought computing to the masses. MS didn't bring us innovations or ease of use, they stopped anyone else from doing it. PC users from back in the day will tell you about MS superior competitors: Geoworks, DR-DOS, OS/2. Guess what? MS's monopoly is universally attributed with the death of hundreds or thousands of good products. Now, you may notice that over the years every part of the PC has decreased in cost except for the OS/software, which has increased, largly because it's MS and no alternatives are left. In most cases, the software bundle included in a PC is the most expensive component: more expensive than the CPU, more expensive than even the monitor. Expensive enough that a lot of people who can't afford a computer now could afford one if not for the cost of software. I suggest that you consider that MS has actually hindered the spread of home computing, not helped it, and that widespread computer access and use is a reality despite MS, not because of them.
The ability for average non-technical persons to use a personal computer... With only the flavors of Unix, early Linux distro's, and other non-friendly OS's, Internet access would have been rare
Now, the post to which you were replying was counting the *contributions* of MS vs. Open IP, so you *are* suggesting that MS is responsible for this here. I don't believe that MS really is responsible for or contributed to the ease of use of computer systems. As a matter of opinion, MS has always been behind its competition in terms of ease of use. Always. It still is. Maybe it's true that with only UNIX, Linux or unfriendly OS's internet access wouldn't be as widespread, but that wasn't all that there was. As mentioned before, MS out-marketed most of those competitors that stood to really bring ease of use and widespread computing to everyone.
The inclusion of web-browsers
I hate to repeat arguments, but... again, what leads you to believe that this was a contribution? It was being done before MS did it. Netscape was available with new computers before IE was, but MS killed that because it was allowing customers to grow at a pace and along a course that MS could neither keep up with nor control.
Like it or not, MS filled a serious demand in the computer world. They broke lots of laws, squashed lots people, and did it the closed way, but...
I'm not going to argue that MS filled that demand, but the point that has to be made is that the demand would have been filled by someone else if not for MS. Squashing competitors, killing products, and stiffling the growth of an industry isn't something that MS should be lauded for just because they ended up with a lot of cash and customers. And, they may have made more money than any other company, but if their prospertity came at the cost of the jobs and businesses of thousands (easily tens of thousands) of competing developers, so they've cost us as Americans more than they've made us.
Writing MS off because we don't like they way they do business isn't "a serious mistake", it's the responsible course of action.
volume at the expense of clarity, he claims that this volume suffers neither fault.
It's early, and when I read this, I thought he meant that the book had neither volume nor clarity.
Looking at the new emulator, it was just posted 2 days ago (5/1/01). My guess is that they'll release the source real soon now, but it probably wasn't quite ready for prime time just yet.
The GPL doesn't say anything about giving them a few days. If the product was ready for release, then it should *all* be ready, and if it's not, then nothing should have been released. Even if they release the source soon, they require users to agree to a EULA which is not compatible with the GPL, and is not the license for that software, which is illegal. I do believe that Corel did that with their Linux distro not so very long ago. The community reminded them of their error, and they fixed it.
The problem is that several companies don't seem to understand that the GPL is a license, and that its terms are every bit as binding as any license that they pay for. If they need reminders, we've got plenty of them.
I think your math is off, and doubt that you've dealt with MS servers recently. Let me give you my take on the situation:
- Lets say your "entrepeneur" wants support for his Linux, so he buys Red Hat Linux Professional Server. For $180 bucks he can provide file sharing, printer sharing, domain logons for Windows, mail service, etc for all of the users in his office.
- Windows 2000 professional might make a fine desktop (still crashes too often for my taste), but it will in no way provide you with a server. For that you will have to buy Windows 2000 Server with client licenses for each user. If this is a *really* small office, you might get away with 10 CAL's, and pay around $1000. You're a lot more likely to need 25 for even a small office and pay closer to $3000. On top of that, you still have to buy Exchange for email, and probably MS SBS, too. Throw in a few thousand more dollars. The cost difference for the OS is not insignificant. It could easily pay for several visits from a consultant.
- IBM's SBS isn't free, but neither is Microsoft's.
- A linux box can be set up in short order by almost any technically apt person. You probably have one on staff if you work with any number of computers, but you may not if that's not your feild. So you hire a consultant to set it up. Once done, Linux installations rarely need baby sitting. A configured box can usually be left on its own to do its job.
- Windows is not easy to "use and administer", and if you think it is, you haven't used MS SBS. For instance, adding users in SBS will create users with a default login script. However, that login script will be overwritten every time you add a user. Managers want to use MS SBS because (this is a quote from a customer) "I don't have to do things the NT way, this is simpler". What's that I hear? The NT way is too complicated? However, when working for this particular customer, we had to do most things the "NT way" because the SBS way screwed them up. Overwriting the default login script meant loosing any local changes they'd made. The only option was to add users with the standard NT user manager. I have other examples, but listing them would take so long, I'd never finish this post. And as an aside... I've *never* seen an RPM install hose a system, so I don't know what you're referring to in your "user error" comment.
- Whether Windows or Linux, your need for a dedicated admin will depend on the number of users and tasks handled by the system. It's about the same for both, and most small businesses can get by with outsourced support. The difference is that Windows will need it more often. Having serviced customers with Windows and Linux boxen, I am qualified to make that statement, and I have the experience (as opposed to your speculation) to back it up.
- It's easier and faster for me to install PostgreSQL or MySQL on Linux than MSSQL on Windows. And don't forget that MS's default installation was responsible for *many* hacks in the last several months, including Wells Fargo.
In summary, people don't buy and use MS because it's better. They buy and use MS because of FUD like yours. Them that have consultants that they *trust* frequently choose Linux.
I didn't see anyone really blame IBM yet, although I suspect people will.
The analogy I was thinking of was blaming Alfred Nobel for blowing people up with dynamite (he invented it for use in mining)
1: Microsoft tech support and Psychic Friends Network are about equal in their ability to provide support for Microsoft Products.
2: Slashdot readers have them both beat all to hell.
My guess is that with your change things appear faster, but in the end you've actually hurt your overall throughput (the amount of computational work you actually get done in a given time).
You almost got that one right. IIRC, increasing the time slice increases throughput, the price you pay is increased latency.
You also have failed to realize that perception is reality. If it seems faster, then it IS faster.
First, I feel obligated to state that I think Python is totally tits. I love it. Thank you.
Reading "What's New in Python 2.1", I'm curious about nested scopes. Given that name lookup has always been the cause of poor performance in Python (or so I've heard), it would initially seem that introducing nested scopes would further reduce the speed at which Python scripts run if you use this feature.
Is there any information available about Python's performance with this addition?
If you guys are open to including an open source ksh, I think it'd make a lot of people very happy.
David Korn recently did a slashdot interview, and his answer to the first question should best express why an authentic ksh should replace the pdksh included now. See this document.
But, calling it an industry standard is probably taking that a bit too far.
Says who? Linux is shipping on more servers than any other UNIX variant. More software is being developed on Linux than on any other UNIX variant.
No, it's not the standard PC platform, but the previous comment doesn't indicate that it was... He just said that Linux is the industry standard UNIX. It may not be the most powerful or mature, but from a lot of people's point of view, it's the standard.
When Red Hat 7.0 was released, there was NO known exploit for that LPD. It was discovered weeks after the release of 7.0. As always, errata was released in a timely fashion.
Red Hat's gcc 2.96 was hardly useless. I've used it to build everything on my system, and it's been totally stable. The only thing that 2.96 wouldn't build correctly was the 2.2 kernel. Guess what, that's not a new problem! IIRC, egcs also wouldn't compile kernels correctly. Assembly code isn't defined as well as, say, C. egcs broke the previous assembler semantics, and the kernel had to be fixed. This is pretty much the same issue with 2.96. The 2.4 kernel has semantics that the 2.96 compiler will accept, and the new compiler is officially used to build kernels now.
C++ support for GCC has never been that good for exactly those reasons. If Red Hat had used any version of GCC other than egcs 1.1.2 (from 6.2) their C++ binaries would not have been compatible with the older, 6.x versions or the 8.x series which will most likey include gcc 3.0.
Based on that unavoidable problem, and their need to support the Alpha platform, Red Hat's engineers decided to use a version of gcc from CVS, and have done a lot of work to make sure that it's stable. Red Hat 7.0 has been rock steady for me. None of the components have suffered because of gcc 2.96rh. It produces stable binaries. Reports of Red Hat's demise are greatly exaggerated.
But, it appears (see other post) the consensus is to downgrade GCC
This consensus is generally reached by the uninformed. You will only cause yourself headache by doing this.
Know of any place I could look for more information? Like how the server decides what programs are given privileges and how those programs request privileges from the server?
The limitation that you're running into is this: you can't SUID a library in UNIX. I understand that you could do this in Multics, but I believe that it was never a feature of UNIX since it's too dangerous. Especially in a threaded application, one could do all kinds of nasty things while the library's running to get unauthorized access. (It might actually work if the library ran as a separate process with its own resources, but I'm going way off topic. Bring this up on a kernel development list and see what kind of responce you get.)
In order for PAM to access any password database, including the shadow file, the application running must be able to access that database. If you don't care about non root users accessing your password database, then you shouldn't be using shadow files. There are other things to consider as well.
Authentication isn't the only reason that a server needs to run as root. A typical login sequence for an interactive daemon works like:
* daemon starts as root
* bind to port (may require root, depending on service)
* accept connection
* read username and password
* compare username and password to authentication db (probably requires root)
* drop root privileges, switching to the authenticated user (definitely requires root)
The last step is an important one. If you want your daemons to be able to switch UID, they have to be run as root. The alternative is to run the imap server as a non-root user with access to the password database, or proxy the username and password to an external authentication daemon, but then all of the mailboxes must be readable and writable by the imap user. It can be done, but then each daemon needs its own hack to get everything working.
Since you can't switch UID's without root, and you can't drop root until after authentication, you're stuck doing authentication as root. You would *think* that this would be a small enough section of code that it wouldn't be commonly exploitable, but it is all the same. </me sighs>
We fully support data, directory and system interop with UNIX, Linux, Novell, Mac, IBM mainframes through our base OS protocol support as well as through products like Services for UNIX, Interix, Services for NetWare, MetaDirectory and Host Integration Server.
Well, I'm glad this was first so we can get this out of the way. Doug defends interoperability by stating that MS offers products that will provide it. However, their OS's alone provide interoperability only with other, similar (read: MS) products. If you want interoperability with, say, a UNIX system, you have to pay for additional software licenses. Their customers are criticizing them not because interoperability isn't available, but because they feel like they have to pay for MS licenses for every machine they run, whether it actually uses MS software or not.
Further, Dave goes on to defend MS tactics of 'Embrace and (Extend|Extinguish)' as protecting their IP. If we consider MS's undocumented (for our intents and purposes) extentions to Kerberos, we can clearly see why customers complain about 'vendor lock-in'. In order for a Kerberos ticket to be valid in a Windows domain network, it must contain the additional information that a Windows server will place there. That's valid. That's why there's a chunk of a Kerberos ticket left undefined by the standard. However, because that information isn't documented, only a Windows Kerberos server can be used for authentication. That is not interoperability. That does not give your customers the ability to choose the best solution for the job.
If your ISP used NT servers as terminal servers (as opposed to Portmasters, MAX's, Cisco, etc.), then they *would* have to pay for a client license for every line that would be in use at max capacity. If they're just running NT as a Radius server, then they only have to pay for client licenses for each of the terminal servers that connect to them. An ISP using NT would normally have to pay fairly hefty client license fees. Figure out how many users will be checking their mail, how many web pages are being viewed and how many radius clients are connecting simultaniously at maximum capacity, and pay client licenses for all of them.
In the Alaska Airlines situation you describe, the clients in question are connecting directly to the NT servers and using their resources. According to MS, that means they have to pay for client licenses.