Disclaimer: I'm not the original poster, I was the one clarifying his point.
There are issues with blocking outgoing port 25.
Issue one is one of configurability and roaming. Right now there's no standardized way of passing on SMTP server information to clients, which means a user roaming on two or more ISPs cannot easily switch without going into the preferences of every blasted mail app they use and change the settings. This might be described as an annoyance, and is probably achievable via some funky scripting and configuration of a local, forwarding, SMTP server, but that kind of defeats the purpose.
Issue two is one of privacy and network transparency. You may want to route email via a particular relay (for example, one that supports Alternic hosts, or a work relay that supports internal-only FQDNs.) Yes, you can ask for that relay run on a port that isn't 25, but that's still relatively unusual, so unusual in fact that many mail clients still do not support anything but port 25.
Issue three again goes back to network transparency: what if the service you're trying to contact is not an SMTP server?
"Now wait!", I pretend to hear you reply, "I don't give a crap because none of these apply to me, and they can all be fixed with complicated Perl scripts, or convoluted network configurations, and by upgrading to the latest versions of the latest email programs regardless of whether the one you use, that you like, that you've been using for years, supports what you want." Well, I know many who'd argue this.
Likewise, I suspect the same people would argue that it's ok to draw on someone's livingroom wall, because the wall's owner can always paint over the marks, or put down wallpaper if the pen has created grooves. And it'll not cost anything because everyone has spare paint just lying around.
In other words, someone somewhere is taking something that's previously been ok, causing a problem, and expecting everyone else to make the effort to get things to work the way they were.
"But, hold on" says the straw man I'm arguing with. "Nobody has to draw on a livingroom wall, whereas we have to deal with spam." I'd say dealing with spam is a "want to" rather than a "have to", but more importantly I question whether taking email out of the hands of the users has been particularly successful at eliminating spam. I would argue it hasn't. I'd argue that nothing has been more destructive to the integrity of the Internet and nothing has helped spammers more than this type of anti-spam system. And when I say "nothing", I pretty much include spam itself, unless you see anti-spam as caused by spam of course.
The bottom line really is that blocking ports breaks things. You can come up with an argument for saying that that breaking is simply necessary, that the circumstances simply require it, just as we're all (well most of us) against killing people but (most of us) see the need for the occasional war, or at any rate self defense killing. But people do end up dying in wars, and legitimate activities do get hampered when you arbitrarily block ports.
Why am I still with Earthlink? Because of all of the issues that can affect my broadband, this - the blocking of outgoing 25 - is minor. Most importantly to me is that they do not blocking any incoming ports, they do not ban me from running my own network (though they refuse to support it, but that's fine) or servers, and they let me use any OS I want.
I'm having difficulty working out why the device in the linked to article is refered to as a "mantrap":
In particular, Forchia knew that the store was equipped with a device known as a "mantrap," which is a small hallway between two steel doors separating the customer area of the store from the employee area. Forchia testified that the mantrap, which is activated twenty-four hours a day, is designed to detect when there are more than two pressure points on the mat covering the floor of the enclosed hallway. If more than one individual walks through the mantrap at the same time, a silent alarm is triggered, calling the alarm company and notifying the police. Realizing that she could trigger the mantrap alarm without Denton knowing, Forchia walked into the mantrap and waited for Denton to follow her in. The two then proceeded through the mantrap into the employee area of the store together. Forchia's plan worked, and, when the police phoned the store a short time later, she was able to use "yes" and "no" answers to advise them that a robbery was in progress.
It appears to be an alarm system, or specifically the person detector part of that alarm. I can't see in what way it traps anyone.
The United States most certainly is a democracy. It's not a constitutional democracy, that is, it is not obligated by the constitution to be one, but the key test - that the legislature is answerable to the governed - is passed.
It could be more or less democratic than it is, but in boolean terms, it is one.
In order to cease being democratic, you'd have to abolish the link between the governed and the legislature. You'd have to abolish clause 3, the 17th amendment, and change the constitutions of a fair number of the states to abolish popular representation in their legislatures.
In a sense, informally, the US has a modified version of that system for Presidential races. American voters get to choose the right/left candidates from a cast of tens (in the primaries), and then get to choose which one they want as the actual President.
It's not perfect, but the notion that the choice is completely between two candidates who have come from nowhere tends to be exaggerated by most who argue that.
I just went through Frances too (see journal) and my employer has decided to pay all the salaried people and not pay all the hourly for the days the company was closed (three - Thursday, Friday, and Tuesday); the worst is that the hourly people are the worst paid in the company, so this affects them even more than it would affect anyone else. Needless to say, there's a lot of complaining, and when the economy starts to turn and get a little healthier, I suspect the company will lose a lot of valuable staff.
The positive side is that there is something you can do about this that doesn't involve lawyers - contact FEMA. FEMA has an emergency assistance program that covers lost wages through hurricane related office closures. My employer is "helpfully" encouraging its hourly staff to do this. I don't think it's ethical, especially as my employer isn't in the line of work where it'll lose business because of this (it'll force everyone to have stress attacks for the next month to catch up, but if you think anything's going to be delayed...) but there is that option.
What about only allowing people who are subject to the laws of the United States vote? Wouldn't that make more sense than the smaller group who pay taxes? Or did I miss something, and the laws of the United States only apply to taxpayers?
A nit-pick, but on an issue that annoys the F out of me:
This is one of the reasons why we're a republic, or a representative democracy.
Despite the insistance on certain trying-to-be-clever people using the word "Republic" to mean "not a democracy", there is no connection with one another. The US is a Republic because it has a head of state that isn't head of state by birth (ie is not a King), not because of the electoral system used to put him there. The US is also a Democracy because the legislature is answerable to the people governed.
If Bush had been directly elected, this country would still be a republic. It would be more democratic, but the simple boolean "Is the US a democracy" would be unchanged (ie, yes, it is.)
If one wanted to come up with arguments for or against the US being a democracy, you'd have to go for constitutional limitations (except that, as it currently stands with the member states of the US, the constitution is answerable, it can be changed if the governed feel strongly enough), or laws against certain groups voting, but the most obvious group that's disenfranchised right now - ex-felons - would be argued to have forfeited that right "voluntarily" by many.
So to answer the issue directly, you're correct in saying it's one of the reasons why we're a representative democracy, but it's not one of the reasons we're a republic. We're a republic because we reject monarchism.
Apple's not attacking Real for supporting AAC, they're attacking Real for supporting FairPlay. And their attacks are probably contrary to the DMCA in any case.
QT is licensed under the GPL, so whether Trolltech intends this or not, you most certainly can distribute it without the QT Commercial License, whether you do so commercially or otherwise.
As an example, most Linux Distributions, commercial and non-commercial include QT as part of their distributions. This includes RedHat, SuSE, Mandrake, and Linspire. None of them, to the best of my knowledge, have agreed to anything more than the GPL when it comes to QT.
I would suggest though that TrollTech almost certainly know that, and are making the common mistake of equating proprietary and commercial. In the quote you make, I'm assuming the second bullet point is refering to non-commercial proprietary software.
Wrong, OS X out of the box will create an Admin account which does not have access to the System or any other use account. Root is disabled by default and for the average Joe hard to enable. If I move something into my user folder, no other user can click on it, in fact they wont even be able to get to it or see it.
What I said was completely correct. The default set up of Mac OS X results in a default user being set up who has access to practically the entire disk (with a small number of honourable exceptions, the small number being roughly equal to that in XP, 2000, and NT's default user.)
That default user is, as you say, an administrator. Just as the default user in NT based OSes is.
I'm hoping we're talking semantics here, but when you write:
I disagree with that. It is quite possible (and dealt with in a gazillion different OS licenses) to distribute source for free, yet require license payments if said source, or a derivative thereof, is used in a commercial offering.
I'm hoping you mean "is used in a proprietary offering". This is certainly the case, the most famous example being Trolltech's QT.
However, an offering can be commercial and open source. A company making money from, say, online play, might be perfectly happy with an entirely open source game, depending on the nature of the game and the number of "secrets" that'd have to be shipped with it.
Inferior compared to what? Compared to IrDA, which it largely replaced? How? Do you miss IrDA's directionality?
Bluetooth is very good at what it's designed to do, allow you to connect devices without cables, without having to line things up, just get them together, register them the first time you do this, and then act as if the things are plugged in.
I can't think of any other technologies in widespread use that do this so well. IrDA's directionality made it awkward, you ended up having to line things up, you couldn't just start syncing the cellphone in your pocket to the address book on your PC. It's actually easier to use USB cables. When Clinton lied, no-one died.
In terms of other wireless technologies, DECT is too task specific, and 802.11a/b/g is too much of an Ethernet replacement. You don't want your keyboard to talk to a machine on the other side of the room, and protocols to link devices via 802.11 would be awkward, to say the least.
It would be nice to see some actual reasons why you feel Bluetooth fails and "good riddance" should it be on its way to death.
You're the troll, mate. I responded to someone claiming that the $1,299 iMac is, somehow, comparable in price to an "IBM clone".
If you have a legitimate argument to back that up, let's hear it. However, the fact is I can price similar spec'd real IBMs for less than $1,299, and IBM clones for a few hundred dollars.
And, yes, as I pointed out earlier, this does include comparable optical drives (indeed, you'd be hard pressed to buy a combo drive right now, your IBM equivalent would most likely have a DVD burner compared to the $1,299 iMac's combo drive). FSB is a performance issue, I'm certain my example PC runs, in practice, at a comparable speed. Neither the IBM nor the iMac includes wireless, but the IBM can be upgraded to support 802.11g for next to nothing, it's $99 for the iMac.
This isn't about marketshare, it's about marketshare in the relevent market - namely, what blackhats are using. And, as you might expect, it's even more complicated than that.
IIS is generally installed on more Windows machines than Apache. So, in terms of relevent market (machines blackhats like using), it does have a higher marketshare.
As long as a computer has a single remote root exploit, it's directly vulnerable. As long as a computer allows users to manage files - their own, other people's, etc - it's vulnerable to psychological attacks. The most targeted platform ends up being the one the blackhat programmers are most familiar with, or perhaps the ones they hate most, for obvious reasons.
Back when the Amiga was a popular machine, it was still dwalfed by the over-all marketshare of the PC. Yet the Amiga was innundated with viruses. Why? Because blackhats had Amigas, PCs were boring office machines, but people bought Amigas to play games. Amigas had the marketshare in the relevent market. Now Windows PCs have the marketshare in the relevent market.
Why not GNU/Linux, you might ask? Why has GNU/Linux so far not suffered much from these exploits when Windows has, and when GNU/Linux is a favourite of anyone technically inclined?
Answer, in this case, it bucks the trend, not by being more secure, but by a variety of things that work in its favour. There's no single GNU/Linux distribution so no monoculture that can easily spread a virus; people "into" GNU/Linux tend to update it regularly, not so much because of security concerns (though many do), but because a high percentage are first adopters. GNU/Linux users tend to be more technical, meaning they're less inclined to blindly trust attachments, more likely to hear about bugs and install patches, and more likely to set up sane security measures. None of these have anything inherently to do with the security of GNU/Linux.
OS X is unpopular with virus writers because virus writers do not own Macs. It's that simple. Apple releases security updates on a regular basis (proving there's nothing inherently secure about the code), OS X, like XP, in its default, user-encouraged-to-set-it-up-this-way-and-leave-we ll-alone configuration, gives the user full access to almost all areas of the disk, files can be "run" by double clicking (and quite right too!) In this case, it's all about marketshare.
Right. Now take that same machine, upgrade the video card to a GeForce FX 5200
Covered in my comment. It's about $50, as someone else commented.
the hard drive to 80 gig (from 40)
Ok, let's trick that out and add that. IBM doesn't do it, but that's about $50-100 from Staples.
CDR/DVD drive to a DVD-RW
Nope, the low end iMac comes with a CDRW-DVD combo. You'd be adding less than $100 at current prices for a DVD+/-RW right now though.
We're still looking at well below $1,300.
and it's hardly integrated, so it's hardly a fair comparison.
Again, I covered that in my comment. It's EXPANDABLE, therefore the IBM represents better value. It's also the case that because it's not integrated, and because of that expandability, IBM's should, by rights, cost MORE than Apple's.
Furthermore, pricing the cost of buying your own hardware and putting it in isn't fair, and it's silly besides. Who wants to buy half a computer, especially if you're in the market for an all-in-one?
Tell that to all the people saying 256Mb isn't a problem, because everyone can upgrade straight away using the ultra-cheap RAM prices.
In any case, you're missing the point. The person I responded to said that the iMac is somehow comparable in price now to IBM clones. Fact is, you can get an IBM clone for a hell of a lot less than that, and the "real thing" - with the added benefit expandability - for a fraction less.
This is not a cheap machine, when compared to others of similar spec or just when looking at the entry level price regardless of spec. It's a little absurd for people to claim otherwise. By all means claim it's beautiful, claim grandma can't cope with more than one wire, (though two box vs three box doesn't seem a significant difference to me, grandma would be better off with a laptop), claim OS X is the best OS in the world (I like it very much too), but quit the pretense that this is somehow "low cost" or "competitive" - no bloody way is it either.
Don't know about running Windows, but my three year old $200 Wal*Mart box is running fine. It's one of these machines that's up 24/7 (in my Florida appartment.) Never had a problem with it, though I did upgrade the graphics card because the built in one wasn't particularly good. The thing has a 800MHz VIA C3, which 3 years ago wasn't top of the line but was reasonable for a low end machine. It runs fairly cool, which is a good thing.
Why do you ask? Do you believe there's a significant difference between the commodity parts used in $200 machines and the ones used in Dells?
There are three iMacs announced today, and only one has that particular spec.
The other two have 17" displays. I priced up IBM (yes, IBM, evil expensive IBM) hardware roughly matching the spec (but not an AIO) here. So, yeah, the top-of-the-line iMac might be semi-reasonable value, but I don't think any of the others are.
You failed to miss the fact that this isn't a 32bit machine, which you compared it to. Price out a comparable athlon 64 system though and you'd probably hit a similar price.
Right now, there's no practical difference between a 32 bit machine and a 64 bit machine for the consumer desktop, especially when we're comparing against a machine with a 2G upper RAM limit. If we're not comparing upper RAM limits, then we're comparing performance, and I seriously doubt a 2.8GHz P4 is any slower, in practice, than a 1.6GHz or 1.8GHz G5, especially when comparing the two running essentially 32 bit code.
Back when we were comparing the internally 16 bit 8086 to the interally 32 bit 68000, it directly mattered because programmers found it easier to work with nice simple 32 bit integers to access memory than resort to paging and other such tricks. I doubt that 64 bit pointers are going to be anything like as important in the next half-decade, especially now that virtually everyone programs at a higher level.
It's kind of funny. It's expensive because it's an AIO, but it's not really an AIO, because it doesn't come with enough RAM, so it's a safe bet that the first thing a buyer will have to do is buy some. Either that, or the buyer will assume Macs are slow and underpowered, that they can't play perfectly ordinary games, let alone do at any reasonable speed the big multimedia operations the machine is sold to do, and will switch to Windows for their next machine.
And the kicker is "AIO" actually ought to be cheaper. You're sacrificing expandability for an everything-in-one-unit design, and that design ought
to mean cost savings (cases, bulk buying/guaranteed all unit sales, etc) for the manufacturer. But it isn't.
It's as inexpensive as a IBM clone and worth more in value.
I hope you're being sarcastic! It's exactly the same price as its predecessor. And IBM clones start at about $200 these days, $300 including monitor.
Indeed, forget clones, try pricing an IBM at around this spec. An A50 with a 2.8GHz P4 ($599) and 17" monitor ($459) is $1,058.00. "Aha!" I pretend to hear you cry, "It's not an AIO, and it doesn't have a decent graphics card!" Well, yeah, it's not an AIO (that's an advantage to the IBM, and it ought to mean the AIO is cheaper anyway), and you can buy a $100 graphics card for it and still be under.
The 256Mb of RAM is my major beef with the machine (not that I particularly like the pricing either.) OS X really needs much more than that to run smoothly running anything but the most trivial applications. Even "modern" games have problems in half a gig.
Slashdot Mirror
There are issues with blocking outgoing port 25.
Issue one is one of configurability and roaming. Right now there's no standardized way of passing on SMTP server information to clients, which means a user roaming on two or more ISPs cannot easily switch without going into the preferences of every blasted mail app they use and change the settings. This might be described as an annoyance, and is probably achievable via some funky scripting and configuration of a local, forwarding, SMTP server, but that kind of defeats the purpose.
Issue two is one of privacy and network transparency. You may want to route email via a particular relay (for example, one that supports Alternic hosts, or a work relay that supports internal-only FQDNs.) Yes, you can ask for that relay run on a port that isn't 25, but that's still relatively unusual, so unusual in fact that many mail clients still do not support anything but port 25.
Issue three again goes back to network transparency: what if the service you're trying to contact is not an SMTP server?
"Now wait!", I pretend to hear you reply, "I don't give a crap because none of these apply to me, and they can all be fixed with complicated Perl scripts, or convoluted network configurations, and by upgrading to the latest versions of the latest email programs regardless of whether the one you use, that you like, that you've been using for years, supports what you want." Well, I know many who'd argue this.
Likewise, I suspect the same people would argue that it's ok to draw on someone's livingroom wall, because the wall's owner can always paint over the marks, or put down wallpaper if the pen has created grooves. And it'll not cost anything because everyone has spare paint just lying around.
In other words, someone somewhere is taking something that's previously been ok, causing a problem, and expecting everyone else to make the effort to get things to work the way they were.
"But, hold on" says the straw man I'm arguing with. "Nobody has to draw on a livingroom wall, whereas we have to deal with spam." I'd say dealing with spam is a "want to" rather than a "have to", but more importantly I question whether taking email out of the hands of the users has been particularly successful at eliminating spam. I would argue it hasn't. I'd argue that nothing has been more destructive to the integrity of the Internet and nothing has helped spammers more than this type of anti-spam system. And when I say "nothing", I pretty much include spam itself, unless you see anti-spam as caused by spam of course.
The bottom line really is that blocking ports breaks things. You can come up with an argument for saying that that breaking is simply necessary, that the circumstances simply require it, just as we're all (well most of us) against killing people but (most of us) see the need for the occasional war, or at any rate self defense killing. But people do end up dying in wars, and legitimate activities do get hampered when you arbitrarily block ports.
Why am I still with Earthlink? Because of all of the issues that can affect my broadband, this - the blocking of outgoing 25 - is minor. Most importantly to me is that they do not blocking any incoming ports, they do not ban me from running my own network (though they refuse to support it, but that's fine) or servers, and they let me use any OS I want.
So people can run their own email servers and thus manage their own email, but they need to route outgoing email via Earthlink's.
It could be more or less democratic than it is, but in boolean terms, it is one.
In order to cease being democratic, you'd have to abolish the link between the governed and the legislature. You'd have to abolish clause 3, the 17th amendment, and change the constitutions of a fair number of the states to abolish popular representation in their legislatures.
It's not perfect, but the notion that the choice is completely between two candidates who have come from nowhere tends to be exaggerated by most who argue that.
I just went through Frances too (see journal) and my employer has decided to pay all the salaried people and not pay all the hourly for the days the company was closed (three - Thursday, Friday, and Tuesday); the worst is that the hourly people are the worst paid in the company, so this affects them even more than it would affect anyone else. Needless to say, there's a lot of complaining, and when the economy starts to turn and get a little healthier, I suspect the company will lose a lot of valuable staff.
The positive side is that there is something you can do about this that doesn't involve lawyers - contact FEMA. FEMA has an emergency assistance program that covers lost wages through hurricane related office closures. My employer is "helpfully" encouraging its hourly staff to do this. I don't think it's ethical, especially as my employer isn't in the line of work where it'll lose business because of this (it'll force everyone to have stress attacks for the next month to catch up, but if you think anything's going to be delayed...) but there is that option.
People respect McCain. That doesn't mean they agree with him. Politics is much more than just whether a candidate is a fundamentally decent person.
What about only allowing people who are subject to the laws of the United States vote? Wouldn't that make more sense than the smaller group who pay taxes? Or did I miss something, and the laws of the United States only apply to taxpayers?
If Bush had been directly elected, this country would still be a republic. It would be more democratic, but the simple boolean "Is the US a democracy" would be unchanged (ie, yes, it is.)
If one wanted to come up with arguments for or against the US being a democracy, you'd have to go for constitutional limitations (except that, as it currently stands with the member states of the US, the constitution is answerable, it can be changed if the governed feel strongly enough), or laws against certain groups voting, but the most obvious group that's disenfranchised right now - ex-felons - would be argued to have forfeited that right "voluntarily" by many.
So to answer the issue directly, you're correct in saying it's one of the reasons why we're a representative democracy, but it's not one of the reasons we're a republic. We're a republic because we reject monarchism.
Apple's not attacking Real for supporting AAC, they're attacking Real for supporting FairPlay. And their attacks are probably contrary to the DMCA in any case.
As an example, most Linux Distributions, commercial and non-commercial include QT as part of their distributions. This includes RedHat, SuSE, Mandrake, and Linspire. None of them, to the best of my knowledge, have agreed to anything more than the GPL when it comes to QT.
I would suggest though that TrollTech almost certainly know that, and are making the common mistake of equating proprietary and commercial. In the quote you make, I'm assuming the second bullet point is refering to non-commercial proprietary software.
Sorry. I meant to write "over-arr"... ;-)
That default user is, as you say, an administrator. Just as the default user in NT based OSes is.
However, an offering can be commercial and open source. A company making money from, say, online play, might be perfectly happy with an entirely open source game, depending on the nature of the game and the number of "secrets" that'd have to be shipped with it.
Bluetooth is very good at what it's designed to do, allow you to connect devices without cables, without having to line things up, just get them together, register them the first time you do this, and then act as if the things are plugged in.
I can't think of any other technologies in widespread use that do this so well. IrDA's directionality made it awkward, you ended up having to line things up, you couldn't just start syncing the cellphone in your pocket to the address book on your PC. It's actually easier to use USB cables. When Clinton lied, no-one died.
In terms of other wireless technologies, DECT is too task specific, and 802.11a/b/g is too much of an Ethernet replacement. You don't want your keyboard to talk to a machine on the other side of the room, and protocols to link devices via 802.11 would be awkward, to say the least.
It would be nice to see some actual reasons why you feel Bluetooth fails and "good riddance" should it be on its way to death.
If you have a legitimate argument to back that up, let's hear it. However, the fact is I can price similar spec'd real IBMs for less than $1,299, and IBM clones for a few hundred dollars.
And, yes, as I pointed out earlier, this does include comparable optical drives (indeed, you'd be hard pressed to buy a combo drive right now, your IBM equivalent would most likely have a DVD burner compared to the $1,299 iMac's combo drive). FSB is a performance issue, I'm certain my example PC runs, in practice, at a comparable speed. Neither the IBM nor the iMac includes wireless, but the IBM can be upgraded to support 802.11g for next to nothing, it's $99 for the iMac.
IIS is generally installed on more Windows machines than Apache. So, in terms of relevent market (machines blackhats like using), it does have a higher marketshare.
As long as a computer has a single remote root exploit, it's directly vulnerable. As long as a computer allows users to manage files - their own, other people's, etc - it's vulnerable to psychological attacks. The most targeted platform ends up being the one the blackhat programmers are most familiar with, or perhaps the ones they hate most, for obvious reasons.
Back when the Amiga was a popular machine, it was still dwalfed by the over-all marketshare of the PC. Yet the Amiga was innundated with viruses. Why? Because blackhats had Amigas, PCs were boring office machines, but people bought Amigas to play games. Amigas had the marketshare in the relevent market. Now Windows PCs have the marketshare in the relevent market.
Why not GNU/Linux, you might ask? Why has GNU/Linux so far not suffered much from these exploits when Windows has, and when GNU/Linux is a favourite of anyone technically inclined?
Answer, in this case, it bucks the trend, not by being more secure, but by a variety of things that work in its favour. There's no single GNU/Linux distribution so no monoculture that can easily spread a virus; people "into" GNU/Linux tend to update it regularly, not so much because of security concerns (though many do), but because a high percentage are first adopters. GNU/Linux users tend to be more technical, meaning they're less inclined to blindly trust attachments, more likely to hear about bugs and install patches, and more likely to set up sane security measures. None of these have anything inherently to do with the security of GNU/Linux.
OS X is unpopular with virus writers because virus writers do not own Macs. It's that simple. Apple releases security updates on a regular basis (proving there's nothing inherently secure about the code), OS X, like XP, in its default, user-encouraged-to-set-it-up-this-way-and-leave-we ll-alone configuration, gives the user full access to almost all areas of the disk, files can be "run" by double clicking (and quite right too!) In this case, it's all about marketshare.
We're still looking at well below $1,300.
Again, I covered that in my comment. It's EXPANDABLE, therefore the IBM represents better value. It's also the case that because it's not integrated, and because of that expandability, IBM's should, by rights, cost MORE than Apple's. Tell that to all the people saying 256Mb isn't a problem, because everyone can upgrade straight away using the ultra-cheap RAM prices.In any case, you're missing the point. The person I responded to said that the iMac is somehow comparable in price now to IBM clones. Fact is, you can get an IBM clone for a hell of a lot less than that, and the "real thing" - with the added benefit expandability - for a fraction less.
This is not a cheap machine, when compared to others of similar spec or just when looking at the entry level price regardless of spec. It's a little absurd for people to claim otherwise. By all means claim it's beautiful, claim grandma can't cope with more than one wire, (though two box vs three box doesn't seem a significant difference to me, grandma would be better off with a laptop), claim OS X is the best OS in the world (I like it very much too), but quit the pretense that this is somehow "low cost" or "competitive" - no bloody way is it either.
Why do you ask? Do you believe there's a significant difference between the commodity parts used in $200 machines and the ones used in Dells?
The other two have 17" displays. I priced up IBM (yes, IBM, evil expensive IBM) hardware roughly matching the spec (but not an AIO) here. So, yeah, the top-of-the-line iMac might be semi-reasonable value, but I don't think any of the others are.
Back when we were comparing the internally 16 bit 8086 to the interally 32 bit 68000, it directly mattered because programmers found it easier to work with nice simple 32 bit integers to access memory than resort to paging and other such tricks. I doubt that 64 bit pointers are going to be anything like as important in the next half-decade, especially now that virtually everyone programs at a higher level.
And the kicker is "AIO" actually ought to be cheaper. You're sacrificing expandability for an everything-in-one-unit design, and that design ought to mean cost savings (cases, bulk buying/guaranteed all unit sales, etc) for the manufacturer. But it isn't.
Indeed, forget clones, try pricing an IBM at around this spec. An A50 with a 2.8GHz P4 ($599) and 17" monitor ($459) is $1,058.00. "Aha!" I pretend to hear you cry, "It's not an AIO, and it doesn't have a decent graphics card!" Well, yeah, it's not an AIO (that's an advantage to the IBM, and it ought to mean the AIO is cheaper anyway), and you can buy a $100 graphics card for it and still be under.
The 256Mb of RAM is my major beef with the machine (not that I particularly like the pricing either.) OS X really needs much more than that to run smoothly running anything but the most trivial applications. Even "modern" games have problems in half a gig.