No, passengers do not STFU. A child I can tell to STFU and have a small chance they'll listen if they are at least five. I don't have that option with adults.
Why is it that people seem to insist that passengers actually quiet down in the car?
Then again, only times a cell phone call is needed while driving in my family, the under twelve kid makes the call, not the driver.
Knowing the way laws tend to work, we'll likely see an outright ban on all cell phone usage in an operating vehicle.
Overall, I'm highly skeptical of the methods of this so called study. They are basing their claim of damages caused on an estimate which they admit has no data to support it. Their measure of impairment when last I heard about this was a laboratory setting that did not appear to be conducive to the real world observed behavior of people who stop talking for a while, sometimes mid-sentence, while they control the vehicle. The nature of the conversation even can have a strong effect on results. A casual talk about the weather is one thing. A detailed analysis of a subject is something else. The reporting seems as bad as most so called journalism on statistical studies.
Obviously you don't drive around with others much. Anyone who has never driven on their own before. Anyone who has the emotional maturity of someone too young to get a license (even if they are physically old enough for one), many self centered people, those who aren't really paying attention to where you are driving, etc. will tend to not follow such cues as you imply.
A recent case I saw had a couple arguing strongly in a vehicle. No cell phones involved. Loss of control resulted in a fatality.
I've seen many times where the adult passenger does not realize that I'm trying to do something requiring my full attention and they kept talking to me despite my pointedly ignoring them while I handled the vehicle.
Next, I imagine people will try to ban radios, music players, etc for similar reasons. We can go back to the bad days of driver fatigue where long stretches of straight roads resulted in the driver being lulled into a state where they were much more impaired than they talk about here.
When you type at 80+ WPM, who cares how you type as long as it doesn't cause you any pain? If it means that I use my right hand for the B key, then I'll do that, even if some typing instructor thinks I should use the left hand.
So called proper typing practice when I learned didn't have any control keys. Then when I got a C-64, that keyboard layout was so different compared to modern standards that you really couldn't compare. Keyboards aren't very standard these days. I live with it, but enjoy the fact that there has been some standardization over the past fifteen years.
I type heavily on my macbook with the new style keyboard, and find it is one of the best laptop keyboards I've ever used. The keys stand up very well to prolonged typing and don't feel like they will start sliding around slightly which is one of the habits I've seen of most previous model keyboards of different laptops.
I'm usually typing full paragraphs, not coding, so my typing habits may be enough different to make a difference to you.
Many Sun keyboards. I own a type 7 Sun keyboard, and with the exception of the left hand double row of special Sun keys that are pretty worthless under Windows (stop, again, copy, paste, etc.) it is a pretty standard keyboard to me. They come in two flavors for type 7, so called "PC layout" and "Unix layout". I ended up getting the PC layout because I wanted a PC layout keyboard I wouldn't have any problems adjusting to.
A side effect is that the diamond key acts as a Windows key in Windows, and a CMD key in Mac. According to xev, it sends a Meta event.
Maybe my fingers are just too long, but I find no problem hitting the control key with my left pinky and then rapidly hitting any key that the left hand is expected to type (and one or two that some think the right hand should type).
This whole "control must be next to the A key" nonsense is amusing to me because the first Unix keyboards I used were so nonstandard by recent viewpoints that I saw non alpha-numeric keys all over the place. DEC had one layout, Sun another, HP a third. I can't even remember what SGI used in that time. They may have gone to a "standard" layout by the time I touched one. The standardization has been useful to me because I no longer have to stop and study every keyboard layout in order to get it right. I can just sit down at a semi-random computer keyboard and type at a reasonable speed. Even most laptops are close enough to the same that I don't have to look at them before using them, unless I'm using something like a Fn-key.
Yes, some people like control near where the caps lock key goes today. Reduce its size, move it around, I probably wouldn't notice. Move the control key and I'll notice and find it highly objectionable. I use it, my fingers have the position memorized so much that even when I move from laptop keyboard to full size keyboard, I only require a minimal adjustment.
On my keyboard, the key that triggers the "Windows key" function and the Mac "command" function appears to actually send a meta keystroke. Remember, Escape, Meta, Alt, Control, Shift are your friends.
So goes the rumor, except my experience is that people talk in about the same volume they do otherwise. Some people half yell whenever they talk in a friendly manner. Others talk quietly.
If I want to hear loud voices, I just sit outside and listen on a quiet day for two people walking outside. They are nearly shouting at each other just because the sidewalk is too narrow for them to walk side by side, or face each other as they talk. I've had many a time I can hear the conversation half a block away.
I just don't believe that volume is the reason people are so annoyed with cell phone usage.
A common theme of a lot of the replies seems to be that black hat behavior is the only way to learn computer security. Far from it. I don't need to have broken into an insecure network connection without permission to understand the problems of sending passwords in the clear. Often, it takes a little imagination, a bit of reasoning, and a bit of technical skill -- the same skills I often suggest for system administrators.
The best security analysts I've worked with are so strictly white hat that they've managed to get policies in place that prohibit black or gray hats from working in security in the companys I've been in. Is it perfect? No. Some people managed to mostly hide their historical black hat behavior. Once it was learned, a quiet black mark was placed against them and they were gently eased away from security work. There are enough good security professionals who have no history of breaking into computers without permission of the owner to fill the jobs requiring that level of technical skill.
Most of the places I would work at have long standing policies that forbid the use of even gray hats in security. It doesn't matter if they are employees or contractors or consultants. If it is learned that you have a black hat record, you are out of security.
Seem harsh? Maybe, but it sure beats the alternative of hiring yet another pretend reformer.
While there was a moral victory for science educators at showing the issues in trying to restrict science education, most discussions I've read of the outcome of the trial point out that the end result was removing from textbooks significant material that was considered offensive.
This changed dramatically in 1957 with Sputnik. There was a brief rush to teach more advanced technical subjects. Beyond that date, I have anecdotes rather than more solid information about the state of education.
In math education, many people deride and criticize the New Math movement for focusing on correctness of technique over the answer. This despite the fact that in advanced math, all emphasis was on the technique. A sign error in a multiplication in a calculus class would likely lose a point or two, but would be unlikely to cost you all points in the problem if you showed understanding of the calculus involved. It also helped result in geometry being taught as a mathematics course instead of an engineering course (with theorems and proofs).
Yet despite that, New Math is often cited as the end of advanced math in schools.
I will agree that elementary math education has significant issues. I had extreme objections to the math that the public schools tried to teach in the past five years. I objected strongly to the fact that geometry was changed from a mathematics course to an engineering course (no work on theorem proof and studying math as a system of making proofs from axioms and previously proven theorems.)
Some more details that weren't mentioned in the summary. Since I saw the news story several times from different sources, I didn't bother reading the above article.
The student in question didn't just refuse to behave once, but multiple times that day. Each time, the on site police officer was called in to deal with the student after the student refused to follow the legitimate directions of the teacher to put away the cell phone. The police officer tried to demand that the student surrender their cell phone. The student lied and claimed they didn't have one, at which point the officer left.
It sounded like the school gave every opportunity for the student to resolve the situation by following directions.
Part of what SOX did is to significantly increase the penalties for auditors certifying something that they shouldn't. In other words, it no longer will take an Enron to get a major auditing group in fairly serious trouble.
SOX is interesting partly because compliance isn't required at time of audit. What is required is an implementable plan to come into compliance. Then, you have to show you are making real progress on it, or be able to document previously unforseen issues that delayed compliance with that control. This makes a certain amount of business sense. If a major application that is SOX impacted cannot comply with SOX, you can't really shut down the business until you can implement a replacement. It would defeat much of the purpose of SOX. You can however, hold the management's feet to the fire until they do fix the issues.
What SOX and related audit laws do is provide a method to show if the fraudsters are lying or not, and to give regulators real power over them if they catch them before they cause significant harm to others.
Yes, there has been some transition period while people get used to the SOX requirements and how to handle them in addition to the normal requirements. I've seen increased attention lately, as well as better checks and balances to help detect if someone is trying to deceive the external auditors, or act in collusion with a corrupt auditor.
There's another area I've seen SOX help a great deal, computer security. Prior to SOX, there was a strong tendency in management to say "all security risk is outweighed by the potential gain." In other words, there was an unrealistic expectation of security vulnerabilities never being exploited, or if they were, being outweighed by the advantage of not spending the time to improve security.
Now, with the SOX computer requirements, management is forced to monitor computer security as one of many elements of operations. As a result, there is more attention to how servers are configured, and from what I've heard, server reliability ends up improving. People are forced to review the servers, and when something needs fixing, there is a little bit of the "while you're fixing X to comply with audits, fix these other bugs by fixing X in this particular way."
In one example repeated over and over, an application had non-compliant authentication systems. Bringing the authentication into compliance allowed the system to be integrated with the corporate authentication method, reducing helpdesk costs and improving the user experience (fewer passwords to manage).
Auditing has seen a real increase. SOX is far from the only cause, especially in the example you give. Other audits and rules may also be responsible for some of the increased auditing requirements.
Disclaimer: I work in the group that responds to computer security audits, gathering technical information for the auditors, as well as working on implementation of security policies.
What is this break time of which you speak? Do you perhaps refer to the mythical lunch break that I hear people describe? I heard that people could actually stop working while they ate, rather than gobbling food down while on mute or in the two minutes between calls.
Try that netstat command on a well configured, SAN enabled HP-UX server. Then when you are done picking your jaw up off the floor, pipe it through your favorite pager so you can see just how many network services are enabled on this "locked down server".
I help maintain the official list of blessed network services at $work. It's long, as in dozens of entries long. These are just the ones that very significant numbers of systems would run, regardless of application. Even Solaris and Linux boxes have huge numbers of additional ports you wouldn't expect to make them part of an enterprise environment. To name a few:
ssh security management tool (corporate mandate) syslog (so we can send log files to central log server) availability monitoring tool capacity management tool SAN management tool 4-5 backup daemon ports (Netbackup, I'm looking at you here). configuration management daemons (e.g. cfengine or tivoli, sometimes you are forced to run both) Hardware management tools and daemons. SMTP daemon
Notice, I haven't mentioned some popular ones like NFS servers, web servers, etc.
Servers run huge numbers of ports today also. You can't just call a server a system that is locked down to run few network services. In fact, the server status causes it to often run network listeners that I don't expect on the desktop (SAN management tools, enterprise hardware management tools, etc.)
Base 10 is quite possibly the worst thing to happen to our learning system. It is also the reason why I find the metric system so incredibly annoying as to be completely unusable for real life. There is absolutely nothing good about base 10 except that we happen to have five digits coming out of each of our two hands, making it easy to multiply by nine on our fingers.
Base 60 is actually one of the friendlier bases I've seen. Evenly divisible by 2, 3, 4, 5, giving it a huge range of even subdivisions. (12 is the other base that I find very easy to work with for natural usage).
One of the better ideas I've seen for bringing open source software in house is to set up internal repositories so people aren't constantly getting it from an outside source. So if one gets fifty requests for say firefox, bring it in house once, and have someone volunteer to update the software as needed. Further requests are then assigned to obtain it from the internal repository. It provides for several key benefits: * Reduced number of versions for easier threat management * Reduced risk of hostile code because it is only brought in from the outside rarely (hopefully by someone who is more familiar with the tool and thus knows how to take reasonable precautions), and then that version being internal is registered for tracking.
Back when I supported CAD work, CRTs were one of our banes because of the lack of color fidelity. We weren't even trying to be picky, we just wanted green to be roughly green rather than roughly brown.
This came up with those huge D roll plotters trying to print the diagrams.
Of course it is not a valid argument to take a highly specialized field and generalize it to the general purpose computing. That's worse than trying to claim that Gentoo Linux is a good general purpose OS for the general consumer.
I have an iBook G4 and a recent MacBook, Matte screen and Glossy screen. For most of my purposes, I don't really care. Both alleviate my primary objection to CRTs, refresh rates. Neither gives me the glare issues some refer to. In fact, stepping outside, both are reasonably usable, though the glossy is slightly more usable if I am able to slightly shade the display. Sitting under a nearby tree, the Glossy is much more usable.
I'm not a professional graphics person. I'm closer to a general user, the kind that Apple is primarily aiming for. But all these claims that no one would like a glossy screen is just bogus. I've used both, in many cases, switching rapidly back and forth over a few hours, and often used both in conditions that people claim are horrible for glossy. I just don't have the problems described.
Too small swap partitions is the bane of my existence right now, and might help explain how an application is causing me to run out of swap space on Linux every week.
We're not even at one tenth of eventual application dataset size, and we've exceeded the RAM capabilities of Linux and the application design.
It doesn't help that as Linux finally joins the ranks of modern operating systems and you can use the swap device to write kernel panics, that you need a little more than RAM as swap (assuming you don't tweak the dump parameters from the default, but then again, this is Linux, not an enterprise ready OS.)
HP-UX ceased to be cutting edge around the time of 10.0. Even 11.0 always felt like "10.x with all the patches pre-loaded", which is exactly what HP branded it as.
AIX doesn't just rest on LPARs, they still claim mksysb and their LVM as the best and only equivalent around. (Sarcasm for the humor impaired).
ZFS isn't yet stable enough for me to use in the enterprise. Every six months, my company looks at it again, and rejects it again. Still not ready, some key part missing, should be delivered Real Soon Now.
There are great things in Solaris 10, but I still can't get ZFS to do what I want and Sun to agree that it is ready for me to use.
One to four security alerts per week requiring us to issue an internal alert and mandate to patch our RHEL boxes does not qualify as "few and far between". And these are only the ones that require the OS to be patched. The ones that are against applications don't get reviewed by me prior to release, just the OS ones. If I do get sent an application issue, I kick it back for application review.
Performance is too aimed at what application you are evaluating, but security is my primary area so I'll speak up.
Linux security alerts too often require kernel patches. Due to major vendors not supporting even minor kernel patches, this requires these alerts go unremediated for months, or lose vendor support from common vendors of middleware, neither of which is an attractive option to a large business.
When writing audit compliance tools for four flavors of Unix: RHEL, Solaris, AIX, HP-UX, we often find RHEL to be the most limiting in providing a reasonable way to check an audit control. Each OS has its time being the odd OS out, but Linux far more often just doesn't have a standard tool available (logins(1m) command comes to mind, a great, if seldom used tool for auditing user configurations and password settings). Yes, we are able to find workarounds, but it does require more effort than it should.
I've also noticed that Linux has required more care and feeding in other security related areas. In fact, the main areas Linux used to be ahead are areas where the other OS's either have caught up, or the workaround was surprisingly simple and global to all the other OS's to the point that Linux became the special case, albeit in a good way.
There is one area where I still consider Linux a better solution from a security perspective. I am not aware of a vendor supported tool on Solaris, AIX, or HP-UX to provide for file system encryption that creates a mountable encrypted filesystem. RHEL provides such a tool with cryptsetup, though documentation is very sparse, it is at least quite usable. While there has been talk of ZFS and encrypted filesystems, I haven't seen one yet, and the Veritas/Symantec does not appear to provide such yet.
Slashdot Mirror
No, passengers do not STFU. A child I can tell to STFU and have a small chance they'll listen if they are at least five. I don't have that option with adults.
Why is it that people seem to insist that passengers actually quiet down in the car?
Then again, only times a cell phone call is needed while driving in my family, the under twelve kid makes the call, not the driver.
Knowing the way laws tend to work, we'll likely see an outright ban on all cell phone usage in an operating vehicle.
Overall, I'm highly skeptical of the methods of this so called study. They are basing their claim of damages caused on an estimate which they admit has no data to support it. Their measure of impairment when last I heard about this was a laboratory setting that did not appear to be conducive to the real world observed behavior of people who stop talking for a while, sometimes mid-sentence, while they control the vehicle. The nature of the conversation even can have a strong effect on results. A casual talk about the weather is one thing. A detailed analysis of a subject is something else. The reporting seems as bad as most so called journalism on statistical studies.
Obviously you don't drive around with others much. Anyone who has never driven on their own before. Anyone who has the emotional maturity of someone too young to get a license (even if they are physically old enough for one), many self centered people, those who aren't really paying attention to where you are driving, etc. will tend to not follow such cues as you imply.
A recent case I saw had a couple arguing strongly in a vehicle. No cell phones involved. Loss of control resulted in a fatality.
I've seen many times where the adult passenger does not realize that I'm trying to do something requiring my full attention and they kept talking to me despite my pointedly ignoring them while I handled the vehicle.
Next, I imagine people will try to ban radios, music players, etc for similar reasons. We can go back to the bad days of driver fatigue where long stretches of straight roads resulted in the driver being lulled into a state where they were much more impaired than they talk about here.
When you type at 80+ WPM, who cares how you type as long as it doesn't cause you any pain? If it means that I use my right hand for the B key, then I'll do that, even if some typing instructor thinks I should use the left hand.
So called proper typing practice when I learned didn't have any control keys. Then when I got a C-64, that keyboard layout was so different compared to modern standards that you really couldn't compare. Keyboards aren't very standard these days. I live with it, but enjoy the fact that there has been some standardization over the past fifteen years.
I type heavily on my macbook with the new style keyboard, and find it is one of the best laptop keyboards I've ever used. The keys stand up very well to prolonged typing and don't feel like they will start sliding around slightly which is one of the habits I've seen of most previous model keyboards of different laptops.
I'm usually typing full paragraphs, not coding, so my typing habits may be enough different to make a difference to you.
Many Sun keyboards. I own a type 7 Sun keyboard, and with the exception of the left hand double row of special Sun keys that are pretty worthless under Windows (stop, again, copy, paste, etc.) it is a pretty standard keyboard to me. They come in two flavors for type 7, so called "PC layout" and "Unix layout". I ended up getting the PC layout because I wanted a PC layout keyboard I wouldn't have any problems adjusting to.
A side effect is that the diamond key acts as a Windows key in Windows, and a CMD key in Mac. According to xev, it sends a Meta event.
Maybe my fingers are just too long, but I find no problem hitting the control key with my left pinky and then rapidly hitting any key that the left hand is expected to type (and one or two that some think the right hand should type).
This whole "control must be next to the A key" nonsense is amusing to me because the first Unix keyboards I used were so nonstandard by recent viewpoints that I saw non alpha-numeric keys all over the place. DEC had one layout, Sun another, HP a third. I can't even remember what SGI used in that time. They may have gone to a "standard" layout by the time I touched one. The standardization has been useful to me because I no longer have to stop and study every keyboard layout in order to get it right. I can just sit down at a semi-random computer keyboard and type at a reasonable speed. Even most laptops are close enough to the same that I don't have to look at them before using them, unless I'm using something like a Fn-key.
Yes, some people like control near where the caps lock key goes today. Reduce its size, move it around, I probably wouldn't notice. Move the control key and I'll notice and find it highly objectionable. I use it, my fingers have the position memorized so much that even when I move from laptop keyboard to full size keyboard, I only require a minimal adjustment.
On my keyboard, the key that triggers the "Windows key" function and the Mac "command" function appears to actually send a meta keystroke. Remember, Escape, Meta, Alt, Control, Shift are your friends.
So goes the rumor, except my experience is that people talk in about the same volume they do otherwise. Some people half yell whenever they talk in a friendly manner. Others talk quietly.
If I want to hear loud voices, I just sit outside and listen on a quiet day for two people walking outside. They are nearly shouting at each other just because the sidewalk is too narrow for them to walk side by side, or face each other as they talk. I've had many a time I can hear the conversation half a block away.
I just don't believe that volume is the reason people are so annoyed with cell phone usage.
A common theme of a lot of the replies seems to be that black hat behavior is the only way to learn computer security. Far from it. I don't need to have broken into an insecure network connection without permission to understand the problems of sending passwords in the clear. Often, it takes a little imagination, a bit of reasoning, and a bit of technical skill -- the same skills I often suggest for system administrators.
The best security analysts I've worked with are so strictly white hat that they've managed to get policies in place that prohibit black or gray hats from working in security in the companys I've been in. Is it perfect? No. Some people managed to mostly hide their historical black hat behavior. Once it was learned, a quiet black mark was placed against them and they were gently eased away from security work. There are enough good security professionals who have no history of breaking into computers without permission of the owner to fill the jobs requiring that level of technical skill.
Most of the places I would work at have long standing policies that forbid the use of even gray hats in security. It doesn't matter if they are employees or contractors or consultants. If it is learned that you have a black hat record, you are out of security.
Seem harsh? Maybe, but it sure beats the alternative of hiring yet another pretend reformer.
Much earlier. Tennessee v Scopes.
While there was a moral victory for science educators at showing the issues in trying to restrict science education, most discussions I've read of the outcome of the trial point out that the end result was removing from textbooks significant material that was considered offensive.
This changed dramatically in 1957 with Sputnik. There was a brief rush to teach more advanced technical subjects. Beyond that date, I have anecdotes rather than more solid information about the state of education.
In math education, many people deride and criticize the New Math movement for focusing on correctness of technique over the answer. This despite the fact that in advanced math, all emphasis was on the technique. A sign error in a multiplication in a calculus class would likely lose a point or two, but would be unlikely to cost you all points in the problem if you showed understanding of the calculus involved. It also helped result in geometry being taught as a mathematics course instead of an engineering course (with theorems and proofs).
Yet despite that, New Math is often cited as the end of advanced math in schools.
I will agree that elementary math education has significant issues. I had extreme objections to the math that the public schools tried to teach in the past five years. I objected strongly to the fact that geometry was changed from a mathematics course to an engineering course (no work on theorem proof and studying math as a system of making proofs from axioms and previously proven theorems.)
Some more details that weren't mentioned in the summary. Since I saw the news story several times from different sources, I didn't bother reading the above article.
The student in question didn't just refuse to behave once, but multiple times that day. Each time, the on site police officer was called in to deal with the student after the student refused to follow the legitimate directions of the teacher to put away the cell phone. The police officer tried to demand that the student surrender their cell phone. The student lied and claimed they didn't have one, at which point the officer left.
It sounded like the school gave every opportunity for the student to resolve the situation by following directions.
Part of what SOX did is to significantly increase the penalties for auditors certifying something that they shouldn't. In other words, it no longer will take an Enron to get a major auditing group in fairly serious trouble.
SOX is interesting partly because compliance isn't required at time of audit. What is required is an implementable plan to come into compliance. Then, you have to show you are making real progress on it, or be able to document previously unforseen issues that delayed compliance with that control. This makes a certain amount of business sense. If a major application that is SOX impacted cannot comply with SOX, you can't really shut down the business until you can implement a replacement. It would defeat much of the purpose of SOX. You can however, hold the management's feet to the fire until they do fix the issues.
What SOX and related audit laws do is provide a method to show if the fraudsters are lying or not, and to give regulators real power over them if they catch them before they cause significant harm to others.
Yes, there has been some transition period while people get used to the SOX requirements and how to handle them in addition to the normal requirements. I've seen increased attention lately, as well as better checks and balances to help detect if someone is trying to deceive the external auditors, or act in collusion with a corrupt auditor.
There's another area I've seen SOX help a great deal, computer security. Prior to SOX, there was a strong tendency in management to say "all security risk is outweighed by the potential gain." In other words, there was an unrealistic expectation of security vulnerabilities never being exploited, or if they were, being outweighed by the advantage of not spending the time to improve security.
Now, with the SOX computer requirements, management is forced to monitor computer security as one of many elements of operations. As a result, there is more attention to how servers are configured, and from what I've heard, server reliability ends up improving. People are forced to review the servers, and when something needs fixing, there is a little bit of the "while you're fixing X to comply with audits, fix these other bugs by fixing X in this particular way."
In one example repeated over and over, an application had non-compliant authentication systems. Bringing the authentication into compliance allowed the system to be integrated with the corporate authentication method, reducing helpdesk costs and improving the user experience (fewer passwords to manage).
Auditing has seen a real increase. SOX is far from the only cause, especially in the example you give. Other audits and rules may also be responsible for some of the increased auditing requirements.
Disclaimer: I work in the group that responds to computer security audits, gathering technical information for the auditors, as well as working on implementation of security policies.
What is this break time of which you speak? Do you perhaps refer to the mythical lunch break that I hear people describe? I heard that people could actually stop working while they ate, rather than gobbling food down while on mute or in the two minutes between calls.
Try that netstat command on a well configured, SAN enabled HP-UX server. Then when you are done picking your jaw up off the floor, pipe it through your favorite pager so you can see just how many network services are enabled on this "locked down server".
I help maintain the official list of blessed network services at $work. It's long, as in dozens of entries long. These are just the ones that very significant numbers of systems would run, regardless of application. Even Solaris and Linux boxes have huge numbers of additional ports you wouldn't expect to make them part of an enterprise environment. To name a few:
ssh
security management tool (corporate mandate)
syslog (so we can send log files to central log server)
availability monitoring tool
capacity management tool
SAN management tool
4-5 backup daemon ports (Netbackup, I'm looking at you here).
configuration management daemons (e.g. cfengine or tivoli, sometimes you are forced to run both)
Hardware management tools and daemons.
SMTP daemon
Notice, I haven't mentioned some popular ones like NFS servers, web servers, etc.
Servers run huge numbers of ports today also. You can't just call a server a system that is locked down to run few network services. In fact, the server status causes it to often run network listeners that I don't expect on the desktop (SAN management tools, enterprise hardware management tools, etc.)
Base 10 is quite possibly the worst thing to happen to our learning system. It is also the reason why I find the metric system so incredibly annoying as to be completely unusable for real life. There is absolutely nothing good about base 10 except that we happen to have five digits coming out of each of our two hands, making it easy to multiply by nine on our fingers.
Base 60 is actually one of the friendlier bases I've seen. Evenly divisible by 2, 3, 4, 5, giving it a huge range of even subdivisions. (12 is the other base that I find very easy to work with for natural usage).
One of the better ideas I've seen for bringing open source software in house is to set up internal repositories so people aren't constantly getting it from an outside source. So if one gets fifty requests for say firefox, bring it in house once, and have someone volunteer to update the software as needed. Further requests are then assigned to obtain it from the internal repository. It provides for several key benefits:
* Reduced number of versions for easier threat management
* Reduced risk of hostile code because it is only brought in from the outside rarely (hopefully by someone who is more familiar with the tool and thus knows how to take reasonable precautions), and then that version being internal is registered for tracking.
Back when I supported CAD work, CRTs were one of our banes because of the lack of color fidelity. We weren't even trying to be picky, we just wanted green to be roughly green rather than roughly brown.
This came up with those huge D roll plotters trying to print the diagrams.
Of course it is not a valid argument to take a highly specialized field and generalize it to the general purpose computing. That's worse than trying to claim that Gentoo Linux is a good general purpose OS for the general consumer.
I have an iBook G4 and a recent MacBook, Matte screen and Glossy screen. For most of my purposes, I don't really care. Both alleviate my primary objection to CRTs, refresh rates. Neither gives me the glare issues some refer to. In fact, stepping outside, both are reasonably usable, though the glossy is slightly more usable if I am able to slightly shade the display. Sitting under a nearby tree, the Glossy is much more usable.
I'm not a professional graphics person. I'm closer to a general user, the kind that Apple is primarily aiming for. But all these claims that no one would like a glossy screen is just bogus. I've used both, in many cases, switching rapidly back and forth over a few hours, and often used both in conditions that people claim are horrible for glossy. I just don't have the problems described.
Too small swap partitions is the bane of my existence right now, and might help explain how an application is causing me to run out of swap space on Linux every week.
We're not even at one tenth of eventual application dataset size, and we've exceeded the RAM capabilities of Linux and the application design.
It doesn't help that as Linux finally joins the ranks of modern operating systems and you can use the swap device to write kernel panics, that you need a little more than RAM as swap (assuming you don't tweak the dump parameters from the default, but then again, this is Linux, not an enterprise ready OS.)
HP-UX ceased to be cutting edge around the time of 10.0. Even 11.0 always felt like "10.x with all the patches pre-loaded", which is exactly what HP branded it as.
AIX doesn't just rest on LPARs, they still claim mksysb and their LVM as the best and only equivalent around. (Sarcasm for the humor impaired).
ZFS isn't yet stable enough for me to use in the enterprise. Every six months, my company looks at it again, and rejects it again. Still not ready, some key part missing, should be delivered Real Soon Now.
There are great things in Solaris 10, but I still can't get ZFS to do what I want and Sun to agree that it is ready for me to use.
One to four security alerts per week requiring us to issue an internal alert and mandate to patch our RHEL boxes does not qualify as "few and far between". And these are only the ones that require the OS to be patched. The ones that are against applications don't get reviewed by me prior to release, just the OS ones. If I do get sent an application issue, I kick it back for application review.
Performance is too aimed at what application you are evaluating, but security is my primary area so I'll speak up.
Linux security alerts too often require kernel patches. Due to major vendors not supporting even minor kernel patches, this requires these alerts go unremediated for months, or lose vendor support from common vendors of middleware, neither of which is an attractive option to a large business.
When writing audit compliance tools for four flavors of Unix: RHEL, Solaris, AIX, HP-UX, we often find RHEL to be the most limiting in providing a reasonable way to check an audit control. Each OS has its time being the odd OS out, but Linux far more often just doesn't have a standard tool available (logins(1m) command comes to mind, a great, if seldom used tool for auditing user configurations and password settings). Yes, we are able to find workarounds, but it does require more effort than it should.
I've also noticed that Linux has required more care and feeding in other security related areas. In fact, the main areas Linux used to be ahead are areas where the other OS's either have caught up, or the workaround was surprisingly simple and global to all the other OS's to the point that Linux became the special case, albeit in a good way.
There is one area where I still consider Linux a better solution from a security perspective. I am not aware of a vendor supported tool on Solaris, AIX, or HP-UX to provide for file system encryption that creates a mountable encrypted filesystem. RHEL provides such a tool with cryptsetup, though documentation is very sparse, it is at least quite usable. While there has been talk of ZFS and encrypted filesystems, I haven't seen one yet, and the Veritas/Symantec does not appear to provide such yet.