Instead of buying pills at the pharmacy we could download recipes from a secure site and produce it at home.
Even that's pretty far fetched. Unless you're talking the *really* simple to manufacture stuff where you basically grind something up and stick it into a capsule.
A semi-valid example would be to look at how easy/difficult it is to make recreational drugs. Some of the drugs made from chemicals (I'm thinking PCP) are extremely hazardous to manufacture without proper tools/ventilation. Other drugs like cocaine require organic plant sources which are not easily duplicated.
Perhaps ComputerWorld is partial to Microsoft. The more I become familiar with tech industry news, the more apparent it becomes that various news outlets have a tendency to be very credulous with the companies they are most familiar with. Other companies tend to have their PR very much sliced and diced and taken with a grain of salt.
For another example, see the recent PC Magazine where they fawn all over Vista and DirectX 10 as being "products of the year". Maybe it would be more credible at the end of 2007 to award them such kind words, but it stretches credibility past the breaking point to do so in 2006.
I'm so glad that they send me that rag for free. Because there's no way in hell I'd pay to receive it anymore.
(Virtualization is much more interesting and important then what benefits the marketing droids think DirectX 10 will bring to the masses.)
You have an emergency of some sort and must have access to your account, but forgot/lost your token, the battery died, whatever. Is there a secondary mechanism that will allow you to access your account which does not rely on the use of the security token? If so, you've just doubled the institution's cost of doing business with no net benefit to the institution.
You mean like physically going to a bank branch and presenting some sort of credentials? (And jumping through whatever hoops are needed.)
Seriously though, what possible emergency would require account access in more then 24-48 hours that could not be handled by keeping a spare credit card in reserve?
ROWE works best in environments where you have a lot of little tasks with very small overlap (highly parallel in nature). For our group, most projects involve between 10 and 60 hours with about a dozen sub-tasks. Which means that each of us will typically be the only person working on a particular project, but that others can easily step in and finish other sub-tasks when we're behind. We'll also typically be juggling anywhere between 2-4 projects at the same time, in various states of completion.
Our coordinator typically does "leveling" every 1-3 days where we take a quick look at what tasks are due soon, what projects need priority in the queue, and what expected ETAs are. He also sets the final decision on what gets priority and what can be shuffled off to another day.
An example would be my current week where I just finished up a 60-80 hour job. As far as we know, there's nothing else big in the pipeline until after the holidays. So after taking a day or two to recuperate and run errands, I get to put on one of my other hats and work on system administration tasks. I might end up doing another small 8 hour job next week, but I should be able to spend most of my time setting up the new DB and web servers.
We've been doing ROWE-style scheduling for almost 7 years now. Having 100% telecommuters helps in changing the focus from seat-time to result-oriented. Off-hand, I'd say that 25%-33% of our workforce works from offsite at least once a week and about 20% work offsite the majority of the time (or full-time). And I'm constantly working on implementing technologies that allow us that flexibility without giving up security.
Communication tech is pretty critical. Home workers need a dedicated phone number (either dedicated line, distinct ring, or cell phone). Broadband is also a requirement. A corporate chat server (Jabber/XMPP), e-mail server (IMAP, WebMail, and POP3) are needed to allow for instant communication without using the phone or using e-mail for unsynchronized communication. Some workers can make do with dedicated desktop machines at home, others will require laptops, others can simply remote into their work machines.
Another tool that is very useful is some sort of project / task tracking system. There needs to be a way that people working on a project (and mgmt) can see where a project stands. Version control systems (Subversion, etc) are also very important because they decentralize file storage while keeping people in sync.
I don't know that you ultimately save money with telecommuters. It's typically a very large gain for the employee because they're not wasting 30-120 minutes per day in traffic. But if the company needs to buy a laptop every few years, pay for broadband, pay for other communication tools, etc, the cost savings can be marginal. The budget for a remote worker will be somewhere between $200/mo and $500/mo, depending on technical needs. But frankly, I think that's a reasonable price to spend to get a huge morale boost.
(I guess it depends on what you pay for office space and how much wastage time you think there is due to office distractions.)
I've worked in such an environment as well for the past 7 years. It helps that our group is project/task oriented rather then time-clock oriented. As long as the projects / tasks are getting done on schedule, without breaking the clock or working unreasonable hours, everyone is comfortable.
My day tends towards working 8:30-noon, then a break for lunch, another 3-4 hours in the afternoon, a break for dinner, and maybe an hour during the evenings or on the weekend. It also helps that I telecommute the majority of the time (our entire group does). Some weeks are 30 hour weeks, some are 50 hours.
Communication is handled via phone, IM (Jabber server in-house), e-mail and a project log that keeps track of what is/isn't done yet on a particular project. Our projects only run for 1-2 weeks at a time and we're all usually working on 3-5 projects at the same time. Our projects are also all small enough chunks that there's not much sharing of work between people (we try to keep things in a single person's hands from start to finish).
We also keep all files in Subversion (prior to this year we used VSS + SourceOffSite). Makes it easy to work disconnected and keep everything synchronized.
I see that it's still not easy to run IPCop v1.4 (2.4 linux kernel) under Xen v3 (2.6 kernel). Any word on when they'll be starting up v1.5 (with a 2.6 kernel)?
Mmm... 8 cores in 2009? Maybe by 2010. I figure 8/16 core chips will require the move to 45nm. (Unless we're talking non-x86.) We are barely getting into quad-core in 2006 and it won't really be available until 2007. And RAM speeds are still in the 3-4GB/s range (what's next? DDR3? DDR4?).
Dunno if I'd call it excessively expensive. It would be nicer at $100 instead of $150, but almost all of the other cell phones on the market are also $100-$600. We're just used to getting them for close to free as part of a long-term contract. (I just bought a Motorola Q... so I don't consider $150 for a basic phone to be overly pricey.)
And I'd consider a screen to be a must-have feature in a cell phone. First off, it lets you see what number you're dialing. Second, it lets you see who is calling before burning up airtime minutes on the call. The BW article (print version) had a fairly flattering image of the phone showing what looked like 24-36 point fonts on the screen. But that might've been a mock-up or photoshopped image.
I'm hoping that they find a successful niche at their price point. Maybe it would convince the bigger players that there's a worthwhile market for simpler cell phones in the market.
Two years ago, though and you could've gotten a 2GHz Athlon64 single-core and still outfitted it with 2GB of RAM. Maybe only 80GB drives (although I think the sweet spot was 120-160GB back then). Prices and performance have been pretty static for the past 2 years.
I still stand by my statement that if you don't need Dell's hand-holding you're better off doing it yourself with quality non-proprietary components. Especially for a game box that you might want to upgrade every so often (CPU/MB/RAM in 2006, better video card in 2007, more RAM in 2008, bigger HD in 2009, rinse-repeat).
When I buy a low-end Dell, I might end up with a case, MB and PSU that are proprietary. (Have they gotten better at this?) I'd much rather spend the extra $50 on a quality case and PSU that I can use across 2-3 motherboard swaps over the next 10-15 years. Cases like the Antec (p160, p180b, Overture II, Sonata II) or Lian-Li (PC-6077, PC-7077, PC-6070, PC-73SL) models that are flexible and sturdy. Some of the newer cases are even designed to be used for both BTX and ATX, but it's mostly a gimmick (BTX seems to be a non-starter).
PSUs are a bit more of a crap-shoot for re-use, although things might settle down now that PSUs have dual PCIe power connectors on them and are shipping standard with 24+8 connectors. I won't say that there's not something around the corner that will make modern PSU supplies out-of-date, but (3) 12V rails and lots of connectors should cover the foreseeable future. (I have perfectly good ATX PSUs that just don't supply enough 12V amperage to drive modern systems.) So a good PSU *should* last 6-8 years, through 2+ MB swaps.
until then we get to read about AMD fanboys talking about how much better this is then Intels current line up. Completly ignoring that fact that Intel has plans for the next 18 months as well.
As long as the two of them remain mostly neck-n-neck (as they are right now if you ignore the extreme ends of the product lineups), I'm happy.
If you don't need VLANs or managed switches, try the "smart" switches which fill the niche between completely unmanaged and fully managed switches. Then you can get things like a 48-port gigabit smart switch for around $1500. Which gives you 40 ports for end-users and 8-ports for uplink or backbone use. Even some of the less expensive "smart" switches support VLANs, but you have to configure using a web browser.
They aren't the fastest things in the world, but at least they do trunking.
(Heck, I have a 16-port SMC smart gigabit switch that I picked up for around $240. Gigabit is definitely within reach.)
For $1050, you could get a custom-built box that would do a lot better then that unit.
Athlon64 X2 or Core 2 Duo - $200
2GB RAM - $250 (RAM is pricey right now)
Asus motherboard ($80)
Some video card ($150)
120-200GB HD (hell, buy 2, $120)
Quality Antec or Lian Li case w/ PSU ($100)
WinXP OEM ($100-$130)
17" flat screen ($200)
Okay... $1200... but that's only if you don't already have a display.
If you're concerned about compatibility issues, go to some place like MWave or Monarch and order a motherboard bundle. They'll even put it together and test the CPU / RAM / MB for you. No muss, no fuss.
Unless you absolutely need Dell's hand-holding... do it yourself. You'll get a screamer that is as upgradeable as you desire without any proprietary parts. (Hell, I've got an ATX mid-tower case here that is 10 years old this month and it's still being used. I think it's on its 3rd innards.)
had three episodes like that in just a few months. Even though it was somebody else's fault in each case, I had to ask myself why I was getting in trouble so much. It isn't enough to be in the right: you want to be safe. And always pushing the limits is not safe.
I'll assume that you've taken a defensive driving course since then...
Excellent point about watching more then just the car in front of you. I'll specifically change speeds to pass (or be passed by) large obstacle-type vehicles on the highway. I can't stand being stuck behind a vehicle I can't see through or around. Always feel like I'm about 0.5sec away from doom in those situations.
I always thought it was a flat '2 second' distance.
Try 5 seconds and you'd be closer to the mark. Granted, 5 seconds is difficult to maintain during rush hour, but in regular traffic it should be attainable. Two seconds seems like no where near enough time to react to situations without having to panic.
(But then I used to work for UPS and still have their defensive driving notes.)
I'm not sure how they can do this. MMORPGs generaly have a clause in the EULA stating that all virtual goods belong to them and have no intrinsic value. If the government decides that they do in fact have value, what happens when a server goes down just after you recieved a valuable item and you are rolled back to before you got it? Can you sue the game's owner for the value of the item? Can these games survive if a hardware failure could result in massive lawsuits against them?
OTOH, MMORPG producers have pretty consistently ignored the sale of items / characters via 3rd parties. Or at least they've not been diligent enough about stamping it out and keeping it to a dull rumble. SOE even went so far with their services as to setup a "trusted" site where you could sell items / characters.
So I see this as karmic revenge for not policing their properties and cracking down on the account / item sales. And I am highly amused.
(If you can't tell, I've always been against being able to buy an advantage in worlds like EQ, EQ2, WoW. It attracts people who have no interest in playing or interacting with others and rewards all sorts of nasty behavior.)
Is it really that bad? Firefox and Thunderbird generally consume 50-75MB for me. Memory is pretty cheap now, so if XUL is what is needed to quickly develop powerful and flexible apps, then I'm all for it.
It's only cheap until you run into the limits of what the machine is capable of having installed. At which point, you're boned because the developers were lazy about making sure their applications are sleek and svelte.
As we are looking at moving to a 3 year rollover on hardware most of the hardware will not be Vista ready for at least the next two years, by which times there will be at least 2 service packs and numerous packs for the inevitable MSism in the OS.
3-year rollover for PC hardware is a waste unless you're doing CPU-intensive work. It was a reasonable lifespan figure back in the mid-90s, when performance and capacity was doubling every 12-15 months but isn't realistic anymore. Excluding dual-core, PC performance has barely doubled over the past 3-4 years. (If that... dual-core and quad-core are the first exciting things to hit the streets since 2003.)
A machine from 2002-2003 (1.5GHz or faster) with adequate amounts of RAM (512MB up to 2GB) is still good enough to run Vista or WinXP or Win2000. At least, as long as you're not doing heavy number crunching and are just manipulating documents. You can probably squeeze another 1-2 years out of a machine that is already 4-5 years old. And if it was a multi-CPU machine, you might even get more life out of it before it feels too slow to be useful. There are very few companies out there where older machines can't be shuffled to less-demanding users to gain a few more years of useful life.
Which brings me around to what multi-core means for lifespan. If a single-core machine can last 4-5 years before feeling too slow to be useful (and I'm typing this on a 5 year old laptop), then dual-core (for an addition $75/machine) should easily get you lifespans in the 8-10 year range.
Hell now that I think about it, I got rid of the last NT 4.0 machine just two months ago. Unless your corporation is very small you keep PCs around until they die or become so obsolete they can no longer run the programs you need them to. In this case we had an active directory upgrade so we had to get rid of all the NT 4.0 machines as they were no longer going to work with the upgrade.
Lucky you. I'm still trying to retire Win98 machines. Although, if things go well, we'll be done by next summer. Hopefully we can still buy WinXP by then.
(And small corporations are even more likely to keep things running until they break.)
Slashdot Mirror
Instead of buying pills at the pharmacy we could download recipes from a secure site and produce it at home.
Even that's pretty far fetched. Unless you're talking the *really* simple to manufacture stuff where you basically grind something up and stick it into a capsule.
A semi-valid example would be to look at how easy/difficult it is to make recreational drugs. Some of the drugs made from chemicals (I'm thinking PCP) are extremely hazardous to manufacture without proper tools/ventilation. Other drugs like cocaine require organic plant sources which are not easily duplicated.
Perhaps ComputerWorld is partial to Microsoft. The more I become familiar with tech industry news, the more apparent it becomes that various news outlets have a tendency to be very credulous with the companies they are most familiar with. Other companies tend to have their PR very much sliced and diced and taken with a grain of salt.
For another example, see the recent PC Magazine where they fawn all over Vista and DirectX 10 as being "products of the year". Maybe it would be more credible at the end of 2007 to award them such kind words, but it stretches credibility past the breaking point to do so in 2006.
I'm so glad that they send me that rag for free. Because there's no way in hell I'd pay to receive it anymore.
(Virtualization is much more interesting and important then what benefits the marketing droids think DirectX 10 will bring to the masses.)
You have an emergency of some sort and must have access to your account, but forgot/lost your token, the battery died, whatever. Is there a secondary mechanism that will allow you to access your account which does not rely on the use of the security token? If so, you've just doubled the institution's cost of doing business with no net benefit to the institution.
You mean like physically going to a bank branch and presenting some sort of credentials? (And jumping through whatever hoops are needed.)
Seriously though, what possible emergency would require account access in more then 24-48 hours that could not be handled by keeping a spare credit card in reserve?
Guilty here as well... because Yahoo! seems to always be up and ping'able. I guess if they were really annoyed, they'd firewall off ping responses.
ROWE works best in environments where you have a lot of little tasks with very small overlap (highly parallel in nature). For our group, most projects involve between 10 and 60 hours with about a dozen sub-tasks. Which means that each of us will typically be the only person working on a particular project, but that others can easily step in and finish other sub-tasks when we're behind. We'll also typically be juggling anywhere between 2-4 projects at the same time, in various states of completion.
Our coordinator typically does "leveling" every 1-3 days where we take a quick look at what tasks are due soon, what projects need priority in the queue, and what expected ETAs are. He also sets the final decision on what gets priority and what can be shuffled off to another day.
An example would be my current week where I just finished up a 60-80 hour job. As far as we know, there's nothing else big in the pipeline until after the holidays. So after taking a day or two to recuperate and run errands, I get to put on one of my other hats and work on system administration tasks. I might end up doing another small 8 hour job next week, but I should be able to spend most of my time setting up the new DB and web servers.
We've been doing ROWE-style scheduling for almost 7 years now. Having 100% telecommuters helps in changing the focus from seat-time to result-oriented. Off-hand, I'd say that 25%-33% of our workforce works from offsite at least once a week and about 20% work offsite the majority of the time (or full-time). And I'm constantly working on implementing technologies that allow us that flexibility without giving up security.
Communication tech is pretty critical. Home workers need a dedicated phone number (either dedicated line, distinct ring, or cell phone). Broadband is also a requirement. A corporate chat server (Jabber/XMPP), e-mail server (IMAP, WebMail, and POP3) are needed to allow for instant communication without using the phone or using e-mail for unsynchronized communication. Some workers can make do with dedicated desktop machines at home, others will require laptops, others can simply remote into their work machines.
Another tool that is very useful is some sort of project / task tracking system. There needs to be a way that people working on a project (and mgmt) can see where a project stands. Version control systems (Subversion, etc) are also very important because they decentralize file storage while keeping people in sync.
I don't know that you ultimately save money with telecommuters. It's typically a very large gain for the employee because they're not wasting 30-120 minutes per day in traffic. But if the company needs to buy a laptop every few years, pay for broadband, pay for other communication tools, etc, the cost savings can be marginal. The budget for a remote worker will be somewhere between $200/mo and $500/mo, depending on technical needs. But frankly, I think that's a reasonable price to spend to get a huge morale boost.
(I guess it depends on what you pay for office space and how much wastage time you think there is due to office distractions.)
I've worked in such an environment as well for the past 7 years. It helps that our group is project/task oriented rather then time-clock oriented. As long as the projects / tasks are getting done on schedule, without breaking the clock or working unreasonable hours, everyone is comfortable.
My day tends towards working 8:30-noon, then a break for lunch, another 3-4 hours in the afternoon, a break for dinner, and maybe an hour during the evenings or on the weekend. It also helps that I telecommute the majority of the time (our entire group does). Some weeks are 30 hour weeks, some are 50 hours.
Communication is handled via phone, IM (Jabber server in-house), e-mail and a project log that keeps track of what is/isn't done yet on a particular project. Our projects only run for 1-2 weeks at a time and we're all usually working on 3-5 projects at the same time. Our projects are also all small enough chunks that there's not much sharing of work between people (we try to keep things in a single person's hands from start to finish).
We also keep all files in Subversion (prior to this year we used VSS + SourceOffSite). Makes it easy to work disconnected and keep everything synchronized.
I see that it's still not easy to run IPCop v1.4 (2.4 linux kernel) under Xen v3 (2.6 kernel). Any word on when they'll be starting up v1.5 (with a 2.6 kernel)?
S939 was killed because of the shift to DDR2. A downside of having the memory controller on the CPU.
Pretty sure the next-gen quad-core AMD CPUs will be compatible with the AM2 socket. At least, they've said so as recently as Sep 2006 interviews.
Mmm... 8 cores in 2009? Maybe by 2010. I figure 8/16 core chips will require the move to 45nm. (Unless we're talking non-x86.) We are barely getting into quad-core in 2006 and it won't really be available until 2007. And RAM speeds are still in the 3-4GB/s range (what's next? DDR3? DDR4?).
(I'm not as optimistic as you are...)
That or Pimped School Buses I can't decide which.
Didn't we already do those back in the 60s and 70s?
Dunno if I'd call it excessively expensive. It would be nicer at $100 instead of $150, but almost all of the other cell phones on the market are also $100-$600. We're just used to getting them for close to free as part of a long-term contract. (I just bought a Motorola Q... so I don't consider $150 for a basic phone to be overly pricey.)
And I'd consider a screen to be a must-have feature in a cell phone. First off, it lets you see what number you're dialing. Second, it lets you see who is calling before burning up airtime minutes on the call. The BW article (print version) had a fairly flattering image of the phone showing what looked like 24-36 point fonts on the screen. But that might've been a mock-up or photoshopped image.
I'm hoping that they find a successful niche at their price point. Maybe it would convince the bigger players that there's a worthwhile market for simpler cell phones in the market.
Aye, I did miss the 2004 information.
Two years ago, though and you could've gotten a 2GHz Athlon64 single-core and still outfitted it with 2GB of RAM. Maybe only 80GB drives (although I think the sweet spot was 120-160GB back then). Prices and performance have been pretty static for the past 2 years.
I still stand by my statement that if you don't need Dell's hand-holding you're better off doing it yourself with quality non-proprietary components. Especially for a game box that you might want to upgrade every so often (CPU/MB/RAM in 2006, better video card in 2007, more RAM in 2008, bigger HD in 2009, rinse-repeat).
When I buy a low-end Dell, I might end up with a case, MB and PSU that are proprietary. (Have they gotten better at this?) I'd much rather spend the extra $50 on a quality case and PSU that I can use across 2-3 motherboard swaps over the next 10-15 years. Cases like the Antec (p160, p180b, Overture II, Sonata II) or Lian-Li (PC-6077, PC-7077, PC-6070, PC-73SL) models that are flexible and sturdy. Some of the newer cases are even designed to be used for both BTX and ATX, but it's mostly a gimmick (BTX seems to be a non-starter).
PSUs are a bit more of a crap-shoot for re-use, although things might settle down now that PSUs have dual PCIe power connectors on them and are shipping standard with 24+8 connectors. I won't say that there's not something around the corner that will make modern PSU supplies out-of-date, but (3) 12V rails and lots of connectors should cover the foreseeable future. (I have perfectly good ATX PSUs that just don't supply enough 12V amperage to drive modern systems.) So a good PSU *should* last 6-8 years, through 2+ MB swaps.
And... now we're completely off-topic.
until then we get to read about AMD fanboys talking about how much better this is then Intels current line up. Completly ignoring that fact that Intel has plans for the next 18 months as well.
As long as the two of them remain mostly neck-n-neck (as they are right now if you ignore the extreme ends of the product lineups), I'm happy.
Competition is good.
Cheaper multi-core for everyone!
If you don't need VLANs or managed switches, try the "smart" switches which fill the niche between completely unmanaged and fully managed switches. Then you can get things like a 48-port gigabit smart switch for around $1500. Which gives you 40 ports for end-users and 8-ports for uplink or backbone use. Even some of the less expensive "smart" switches support VLANs, but you have to configure using a web browser.
They aren't the fastest things in the world, but at least they do trunking.
(Heck, I have a 16-port SMC smart gigabit switch that I picked up for around $240. Gigabit is definitely within reach.)
Pretty sure the PCIe bus can supply 1GByte/s to a 10Gbit card. I know low-end DDR2 RAM is capable of around 3GB/s of sustained data transfers.
And... I forget what the speeds are for HT...
One thing I saw in the Dec 11 2006 BusinessWeek was a cell phone that is *just* a phone (although it does have a 12-15 number memory feature). It's called the Jitterbug from GreatCall.
For $1050, you could get a custom-built box that would do a lot better then that unit.
Athlon64 X2 or Core 2 Duo - $200
2GB RAM - $250 (RAM is pricey right now)
Asus motherboard ($80)
Some video card ($150)
120-200GB HD (hell, buy 2, $120)
Quality Antec or Lian Li case w/ PSU ($100)
WinXP OEM ($100-$130)
17" flat screen ($200)
Okay... $1200... but that's only if you don't already have a display.
If you're concerned about compatibility issues, go to some place like MWave or Monarch and order a motherboard bundle. They'll even put it together and test the CPU / RAM / MB for you. No muss, no fuss.
Unless you absolutely need Dell's hand-holding... do it yourself. You'll get a screamer that is as upgradeable as you desire without any proprietary parts. (Hell, I've got an ATX mid-tower case here that is 10 years old this month and it's still being used. I think it's on its 3rd innards.)
had three episodes like that in just a few months. Even though it was somebody else's fault in each case, I had to ask myself why I was getting in trouble so much. It isn't enough to be in the right: you want to be safe. And always pushing the limits is not safe.
I'll assume that you've taken a defensive driving course since then...
Safe Driving Tips
I think the real issue is mass vs contact area of the tires doing the braking.
A lightly loaded 18-wheeler, using all methods of braking, has a *very* short stopping distance.
Excellent point about watching more then just the car in front of you. I'll specifically change speeds to pass (or be passed by) large obstacle-type vehicles on the highway. I can't stand being stuck behind a vehicle I can't see through or around. Always feel like I'm about 0.5sec away from doom in those situations.
I always thought it was a flat '2 second' distance.
Try 5 seconds and you'd be closer to the mark. Granted, 5 seconds is difficult to maintain during rush hour, but in regular traffic it should be attainable. Two seconds seems like no where near enough time to react to situations without having to panic.
(But then I used to work for UPS and still have their defensive driving notes.)
I'm not sure how they can do this. MMORPGs generaly have a clause in the EULA stating that all virtual goods belong to them and have no intrinsic value. If the government decides that they do in fact have value, what happens when a server goes down just after you recieved a valuable item and you are rolled back to before you got it? Can you sue the game's owner for the value of the item? Can these games survive if a hardware failure could result in massive lawsuits against them?
OTOH, MMORPG producers have pretty consistently ignored the sale of items / characters via 3rd parties. Or at least they've not been diligent enough about stamping it out and keeping it to a dull rumble. SOE even went so far with their services as to setup a "trusted" site where you could sell items / characters.
So I see this as karmic revenge for not policing their properties and cracking down on the account / item sales. And I am highly amused.
(If you can't tell, I've always been against being able to buy an advantage in worlds like EQ, EQ2, WoW. It attracts people who have no interest in playing or interacting with others and rewards all sorts of nasty behavior.)
Is it really that bad? Firefox and Thunderbird generally consume 50-75MB for me. Memory is pretty cheap now, so if XUL is what is needed to quickly develop powerful and flexible apps, then I'm all for it.
It's only cheap until you run into the limits of what the machine is capable of having installed. At which point, you're boned because the developers were lazy about making sure their applications are sleek and svelte.
As we are looking at moving to a 3 year rollover on hardware most of the hardware will not be Vista ready for at least the next two years, by which times there will be at least 2 service packs and numerous packs for the inevitable MSism in the OS.
3-year rollover for PC hardware is a waste unless you're doing CPU-intensive work. It was a reasonable lifespan figure back in the mid-90s, when performance and capacity was doubling every 12-15 months but isn't realistic anymore. Excluding dual-core, PC performance has barely doubled over the past 3-4 years. (If that... dual-core and quad-core are the first exciting things to hit the streets since 2003.)
A machine from 2002-2003 (1.5GHz or faster) with adequate amounts of RAM (512MB up to 2GB) is still good enough to run Vista or WinXP or Win2000. At least, as long as you're not doing heavy number crunching and are just manipulating documents. You can probably squeeze another 1-2 years out of a machine that is already 4-5 years old. And if it was a multi-CPU machine, you might even get more life out of it before it feels too slow to be useful. There are very few companies out there where older machines can't be shuffled to less-demanding users to gain a few more years of useful life.
Which brings me around to what multi-core means for lifespan. If a single-core machine can last 4-5 years before feeling too slow to be useful (and I'm typing this on a 5 year old laptop), then dual-core (for an addition $75/machine) should easily get you lifespans in the 8-10 year range.
Hell now that I think about it, I got rid of the last NT 4.0 machine just two months ago. Unless your corporation is very small you keep PCs around until they die or become so obsolete they can no longer run the programs you need them to. In this case we had an active directory upgrade so we had to get rid of all the NT 4.0 machines as they were no longer going to work with the upgrade.
Lucky you. I'm still trying to retire Win98 machines. Although, if things go well, we'll be done by next summer. Hopefully we can still buy WinXP by then.
(And small corporations are even more likely to keep things running until they break.)