I spent 21 years in the Navy as an electronics tech/electronics material officer. Along the way I completed an Electronics BS and a Computer Engineering MS via night school. I retired in 94. So, the IT world was limited to specific purpose mini-computer systems, dumb terminal/server combos, and a few desktops while I was on active duty. The maintenance folks did most of the systems maintenance. My military training was excellent, but it was for what is now lost art, discrete component repair and electromechancical device repair.
When I left the military, I went into federal agency engineering support and worked primarily with the FAA and the IRS. Both agencies have serious shortcomings with the knowledge level of their government employee computer specialists. There was not a lot of new blood or folks who had continued their educations. The computer specialist field is filled with senior people that started their careers as main frame tape apes/system operators. Unfortunately, they now make decisions about system engineering and acquisition and are for the most part saved by a few knowledgable standouts and a large cadre of better trained contractors.
I will say that the IRS has a dedicated and extremely competant crew of COBOL and Assembler programmers that seem to get the extremely complicated tax systems running correctly every year in spite of what Congress might do to change the tax codes.
I have over 29 years experience working in the electronics industry and as a system developer/implementer. The most recent 10 years were spent as a professional project or program manager on government system acquisitions.
A project/program manager is hired to control three factors of a project, scope, schedule, and budget. In most projects, the PM that was around for contract negotiations has been canned by upper management along with several of his or her successors before the project is completed. The reason is that no matter how savvy the manager is, contracts are captured by proposal teams based on delivering the cheapest bid based on a proposal promising pie in the sky technology. The promises in many cases can never be fulfilled regardless of how much money or time is sprinkled on the resulting development effort.
True project schedules capture reality, but they are built by capturing and carefully analyzing all of the tasks and assets required to deliver a product. This includes personnel, materials, time, and monetary resources required for completing those tasks. After all of the info is captured and the critical path through tasks is plotted, a scientific risk assessment must be accomplished to determine the probability of task completion. Usually working with an 80% probability of completion gives a reasonable calendar date for accomplishign a task.
Unfortunately, contracted milestone dates are usually on paper before this type of assessment is ever accomplished. The result is that virtually all projects/programs are badly over budget, under scoped, or don't have a realistic schedule until the project is almost complete.
I can't defend dolts that end up in project or program management with no concept of the technology they are delivering. However, I will say that a PM is the buffer zone between marketing and engineering that tries to deliver the goods. Being a PM requires a set of skills that are not the same as those required in a developer. Most developers do not make good project managers because they are too close to the details of a technology. Likewise, most project managers would not be competent to take over for their developers. However, a good PM does have to be savvy enough to know when technology or the laws of physics are being exceeded.
While I think it is a particularly gutless stand (since I'm not a business person), I understand the AMD position. They produce an excellent product that just barely stays competitive no matter what type of technical innovation or advantange that the product has. I believe the company is afraid that if they don't suck up to Microsoft, any drivers for AMD friendly chipsets or features might just conveniently disappear from future Microsoft operating systems.
I have to disagree about the satellite connection. I moved to the boonies where I lost my DSL line. I believed the hype and paid the bucks for a 2 way satellite system. First, the $69.00 per month is $30.00 per month more than I paid for DSL. More importantly, throughput is worse than a 56K modem on the uplink, and while the downlink can zoom at times, it is really slow during peak hours when the system is loaded. I've had it drop to below 80K on numerous occasions. Additionally, even in a strong signal area you lose a lot with just a little cloud cover. If it's windy, the rocking of your dish will also play havoc with reception, and finally, if you rely on a ping intensive app such as pcAnywhere, it will not work on the system due to the latency issues involved with a 44,000 mile signal path. I've been a frustrated customer since October and can't get out of my promo deal for another 8 months. Dial-up is slow, but its only 21.95 a month.
I have a DirecPC two way system resold by Earthlink. I'm assuming that it and the Starband systems are similar. I bought the system after I moved to the boonies, not to be on the bleeding edge, but to remotely administer a small network and to maintain a few websites. All of the advertisements lead the consumer to believe that this is a substitute for cable or DSL. This is not the case.
It only took about two months of pain and less than helpful tech support for me to learn to crank the 56K modem back up for running remote desktops or FTPing files.
I'll admit that I get fast downloads, but my best upload time is about 22Kbps, and the latency issues play havoc with ping intensive applications. Additionally, the system dies several times a day, requiring a reboot of a Win2K host machine to get it running again.
I wish that DirecPC had built the system to use an ethernet connection with the modem, rather than the USB one. I think USB works well, but an ethernet connection might have led to the hacking of a smart modem that would allow offloading the satcom protocol from the host PC in the manner of DSL router/switches (and to the freedom to use the OS of your choice).
Slashdot Mirror
I spent 21 years in the Navy as an electronics tech/electronics material officer. Along the way I completed an Electronics BS and a Computer Engineering MS via night school. I retired in 94. So, the IT world was limited to specific purpose mini-computer systems, dumb terminal/server combos, and a few desktops while I was on active duty. The maintenance folks did most of the systems maintenance. My military training was excellent, but it was for what is now lost art, discrete component repair and electromechancical device repair. When I left the military, I went into federal agency engineering support and worked primarily with the FAA and the IRS. Both agencies have serious shortcomings with the knowledge level of their government employee computer specialists. There was not a lot of new blood or folks who had continued their educations. The computer specialist field is filled with senior people that started their careers as main frame tape apes/system operators. Unfortunately, they now make decisions about system engineering and acquisition and are for the most part saved by a few knowledgable standouts and a large cadre of better trained contractors. I will say that the IRS has a dedicated and extremely competant crew of COBOL and Assembler programmers that seem to get the extremely complicated tax systems running correctly every year in spite of what Congress might do to change the tax codes.
I have over 29 years experience working in the electronics industry and as a system developer/implementer. The most recent 10 years were spent as a professional project or program manager on government system acquisitions.
A project/program manager is hired to control three factors of a project, scope, schedule, and budget. In most projects, the PM that was around for contract negotiations has been canned by upper management along with several of his or her successors before the project is completed. The reason is that no matter how savvy the manager is, contracts are captured by proposal teams based on delivering the cheapest bid based on a proposal promising pie in the sky technology. The promises in many cases can never be fulfilled regardless of how much money or time is sprinkled on the resulting development effort.
True project schedules capture reality, but they are built by capturing and carefully analyzing all of the tasks and assets required to deliver a product. This includes personnel, materials, time, and monetary resources required for completing those tasks. After all of the info is captured and the critical path through tasks is plotted, a scientific risk assessment must be accomplished to determine the probability of task completion. Usually working with an 80% probability of completion gives a reasonable calendar date for accomplishign a task.
Unfortunately, contracted milestone dates are usually on paper before this type of assessment is ever accomplished. The result is that virtually all projects/programs are badly over budget, under scoped, or don't have a realistic schedule until the project is almost complete.
I can't defend dolts that end up in project or program management with no concept of the technology they are delivering. However, I will say that a PM is the buffer zone between marketing and engineering that tries to deliver the goods. Being a PM requires a set of skills that are not the same as those required in a developer. Most developers do not make good project managers because they are too close to the details of a technology. Likewise, most project managers would not be competent to take over for their developers. However, a good PM does have to be savvy enough to know when technology or the laws of physics are being exceeded.
While I think it is a particularly gutless stand (since I'm not a business person), I understand the AMD position. They produce an excellent product that just barely stays competitive no matter what type of technical innovation or advantange that the product has. I believe the company is afraid that if they don't suck up to Microsoft, any drivers for AMD friendly chipsets or features might just conveniently disappear from future Microsoft operating systems.
I have to disagree about the satellite connection. I moved to the boonies where I lost my DSL line. I believed the hype and paid the bucks for a 2 way satellite system. First, the $69.00 per month is $30.00 per month more than I paid for DSL. More importantly, throughput is worse than a 56K modem on the uplink, and while the downlink can zoom at times, it is really slow during peak hours when the system is loaded. I've had it drop to below 80K on numerous occasions. Additionally, even in a strong signal area you lose a lot with just a little cloud cover. If it's windy, the rocking of your dish will also play havoc with reception, and finally, if you rely on a ping intensive app such as pcAnywhere, it will not work on the system due to the latency issues involved with a 44,000 mile signal path. I've been a frustrated customer since October and can't get out of my promo deal for another 8 months. Dial-up is slow, but its only 21.95 a month.
I have a DirecPC two way system resold by Earthlink. I'm assuming that it and the Starband systems are similar. I bought the system after I moved to the boonies, not to be on the bleeding edge, but to remotely administer a small network and to maintain a few websites. All of the advertisements lead the consumer to believe that this is a substitute for cable or DSL. This is not the case.
It only took about two months of pain and less than helpful tech support for me to learn to crank the 56K modem back up for running remote desktops or FTPing files.
I'll admit that I get fast downloads, but my best upload time is about 22Kbps, and the latency issues play havoc with ping intensive applications. Additionally, the system dies several times a day, requiring a reboot of a Win2K host machine to get it running again.
I wish that DirecPC had built the system to use an ethernet connection with the modem, rather than the USB one. I think USB works well, but an ethernet connection might have led to the hacking of a smart modem that would allow offloading the satcom protocol from the host PC in the manner of DSL router/switches (and to the freedom to use the OS of your choice).