I should have been clearer in my post. While the disk performance in normal operation may be comparable hardware RAID does have several key advantages. The first being that with anything other than RAID 0 or 1 the cpu hit from the RAID driver during heavy use will be high even on a high end machine.
No, it won't. Even a 500Mhz P3 has a RAID5 checksumming speed of ~1GB/sec. A current low-end CPU (eg: 1.6Ghz Pentium E2140) has a RAID5 checksumming speed of around 4GB/sec.
Suffice to say your average SATA array that's unlikely to even get much over 150MB/sec isn't going to put much load on any remotely modern CPU (at least not from the checksumming).
This is fine if you aren't using that machine for anything else but if it is a desktop machine or a server that is required to do anything else other than serve files then it does cause a problem. This is even more apparent if the array becomes degraded as it will take significantly more cpu time to perform all the parity calculations to rebuild the array.
I'm not sure why you think parity calculations when the disk is rebuilding are any different from parity calculations when the disk is being written to normally, but they're not.
The real hit you take from software RAID is to the bus bandwidth, not the CPU. Most amateurs compare a hardware RAID controller to software RAID on a dinky little 32bit/33Mhz PCI bus (and remember that those onboard SATA ports are probably hanging off a regular 32bit/33Mhz PCI bus), that's why they frequently conclude software RAID is slower ("especially during rebuilds"). When you have a system with ample bus bandwidth the situation is quite different.
If you do need to rebuild the array then a decent RAID card will handle it without putting a heavy load on the server and significantly without affecting disk perfomance.
This is impossible. There will always be a performance hit during a rebuild, no matter whether your RAID logic is running on the system CPU or the embedded CPU on a RAID controller. Further, that performance hit comes from the greater number of IOPS necessary while the array is rebuilding and has nothing to do with "parity calculations" (which even a 10+ year old commodity CPU can do faster than any normal (not to mention most abnormal) arrays could ever hope to be).
In a server that is in constant use this is a key point. Rebuilding a RAID 5 array in software often reduces the disk performance to very low levels, effectively denying access to data until the rebuild is complete.
The difference you are seeing is almost certainly because most hardware RAID controllers throttle rebuilds by default to be relatively slow, so "normal" disk access suffers as little as possible. Software RAID can do this as well - although most HOWTOs tell you to bump the rebuild speed up as high as possible, which is probably why you see the results you do. The downside, of course, that the longer rebuild means a bigger window where your array's performance - and more importantly, reliability - are degraded.
If your data is valuable enough to use raid, then hardware raid is the only way to go. If you have a problem with the os the data is lost.
Rubbish. Software RAID arrays can generally be migrated between machines that aren't even the same architecture, a simple "OS problem" isn't going to damage them.
Yes migration of will probably happen over time, but that is why there is no immediate adoption of Vista, there is no killer feature and the hardware requirements are more onerous.
In this it is really no different than 2000/XP and NT4. There was no real "killer feature" there, either, and the hardware requirements were relatively high (from usable on a 32MB, ~100Mhz Pentium to a 64MB (2000) 128MB (XP) ~300Mhz P2). The closest was Active Directory and, in particular, Group Policy - but even with those two enticements people were running NT4 Domains well into the '00s.
Indeed, I expect the migration from NT4 to Windows 2000/XP and the upcoming migration from XP to Vista/Server 2008 have a great deal of similarity, both in concept and magnitude, and I expect it to take a similarly long timeframe to finish.
I did some 2000 (and even Win98) to XP upgrades in 2003 where the hardware, though older was still capable of running XP within a corporate environment quite well, this is likely not going to be possible with Vista, because the minimum specs are much higher than what I see on a normal corporate desktop (and I'm not assuming that business users will want Aero, 512Mb RAM seems about the norm at the moment (with 256 still common) and importantly low end and low capacity HDD's and processors (After all its much cheaper to pick up new clients with minimum specs)).
A significant proportion of PCs bought in the last 12 - 18 months will have shipped with 1G RAM, which is all you need for Vista in a typical business setting. Even 512M is usable, although definitely "slow".
HDD size and processor power are basically irrelevant. Anything from 1GHz (ca. 1999) and up is fine and 20GB+ drives - even on budget machines - have been commonplace for years.
Interestingly, I also doubt that the migration will be piecemeal like it could be from 2000 to XP, an environment where either client could co-exist peacefully and without issue, I have it on moderately decent authority* that dropping a Vista box in as a straight replacement for an XP box and not making any changes in other areas (AD primarily, whether group policy or user profiles and scripts etc..)is not possible, whilst in a 2000 to XP transition it was (XP being fully backward compatible but without too many fundamental changes, you simply didn't get some of XP's other benefits with regard to group policy etc..)
The only really big potential hiccup I've seen in my (admittedly cursory) look into Vista migrations is that it can't use existing roaming profiles. This is certainly not an insignificant problem, but if you're doing things "properly" with Folder Redirection, it's manageable.
However, I just don't expect fast migrations to Vista, and I doubt anyone else whose job isn't in marketing does either. The simple truth is that Windows hit the "good enough" stage in the early 2000s and convincing people there are genuine business improvements to be had from upgrading from there will be difficult. Windows Server 2008 might have some new "Vista-only" features that make the upgrade more compelling, but we'll have to wait and see.
Why would a business want to buy additional hardware for this upgrade?
They wouldn't - but that wasn't the issue. The issue was the implication that hardware for Vista is (more) expensive.
They wouldn't need to if they didn't upgrade.
Actually they probably will, once the lease on their existing hardware runs out.
If a business has thousands of desktops from a few years ago, say 1.8Ghz CPU, 512Mb RAM, 40Gb hard disk etc... a bog standard corporate desktop, its not cheap to replace all that kit in one go with something vista will run on, its much easier (and more cost efficient) to upgrade as stuff breaks and then at some distant point in the future upgrade to vista because all your equipment is capable of running it and it fits in with the other infrastructure life cycles that you are working within.
Which is exactly how a migration to Vista will be done, just like it was exactly how the migrations to XP, 2000, NT, etc, etc were handled before that.
Why anyone thinks it odd that businesses aren't migrating to Vista immediately - or tries to imply it has problems because of that - is beyond me. Most companies, running a ~3yr hardware cycle, won't even be _considering_ a Vista rollout until mid 2008 at the earliest. This is completely and utterly independent of any technical issues - real or imagined - that exist.
Microsoft knows this (which is why aberrations are worth talking about). Anyone with any real experience in corporate IT knows this. The only people who seem not to know it are the anti-Microsoft trolls on Slashdot who, for some reason, think because corporate IT departments are following the same rollout methodology they've been using for the last couple of decades, it means Vista has problems.
Vista may become viable as hardware becomes cheaper [...]
Huh ? An entry level <US$400 PC from Dell will run Vista fine. How much cheaper do you think a new PC can get ?
While there are numerous genuine reasons why an upgrade to Vista isn't compelling, there are far more stupid ones - and "expensive hardware" (along with "DRM") would have to be one of the dumbest.
My "cursory googles" only show up references to Peter Gutmann's paper, which is both mostly FUD and mostly refuted - and added to *that* most of the pages based off Gutmann's paper are extrapolating to conclusions even he didn't draw.
Either accept it as a given for the sake of argument, or don't bother responding. This is the last I have to say on the matter.
Suit yourself, but you're still wrong.
A protected path DRM implementation has to be operational at all times by definition. If it isn't operational, the path isn't protected.
No, it only needs to be operational when "protected" media is being played. If the content isn't DRM-encumbered, the path doesn't *need* to be protected. This is exactly how Vista works - if the media isn't DRM encumbered, the DRM system isn't active.
Incidentally, it is trivially simple to disprove your assertion, simply by playing HD content without DRM on a Vista machine through an analogue output like VGA. The media plays at full resolution, which it would not do if the DRM system is always active (since the output path isn't "protected").
You'd be somewhat correct in the case of the old style windows media DRM; that could (in theory) be written in a modular way similar to decoding plugins. However, it would still be taking up drive space without giving any benefit to the owner of the drive.
The benefit is being able to play media they otherwise wouldn't. You may not consider that a benefit, but people who want to watch the movies certainly will.
If Microsoft stood up to Sony/whoever backs HD DVD, they'd have to either give up the draconian DRM or the Windows platform. Guess which they'd choose?
Windows, unquestionably. Most people will be playing their DRM-encumbered media - like they do now with their regular media - through games consoles and cheap set-top players.
I'm not sure where this (apparently common) idea that Microsoft has significant sway over the media companies comes from. It certainly doesn't follow any rational analysis, given that Microsoft's presence in the "content creation" market is essentially nonexistent and even in the "content player" market is quite small - the vast majority of content is consumed through appliances like set-top players, games consoles, iPods, and the like, not PCs running Windows (or anything else, for that matter).
Regardless, we'll all still be able to see the content, because all DRM is cracked if it protects anything worth having.
Of course - which strengthens my argument about the irrelevancy of DRM even more.
The iMac shipped with three replacements for a floppy drive. The first was the Ethernet port...remember that the "i" in the name stood for "Internet."
Sorry, but in 1998 internet connections - especially broadband ones - were *far* from ubiquitous.
The other was the USB port--flashdrives were well on their way to becoming common when the iMac hit the market.
The first flash drives weren't on the market until 2000, so they sure as hell weren't "common" in 1998. Further, since Windows didn't have built-in support for them until Windows 2000 and Windows ME, they weren't "common" until a couple of years after that. It would have been brave indeed to walk around any time before about 2002 with the assumption your USB key would just "plug and play" with the majority of computers you'd encounter.
The whole idea was that data would move across the pipe more and more, so the floppy was not necessary.
Indeed, there was nothing wrong with the idea - the problem was that it was half a decade too early.
The success of the iMac, and the way we work now, shows that was an accurate prediction.
The complaints and subsequent boom in the late 90s for USB floppy drives suggest lots of people thought the iMac's lack of any easy and compatible method to transfer data was a problem.
No, I'm pointing out that "walking into the store" and "taking stuff off the shelves" - like "connecting to a WAP" and "using services via that WAP" - are different things. Permission to do one does not in any way imply permission to do the other.
The implicit permission is established by a social norm, which takes into account that the person who is supposed to follow that norm must have a reasonable chance to recognize what is expected of him. If your garden looks just like the park that it borders on and there is no sign or fence to tell people that it is private property, then no social norm and no law forbids people to enter your garden. If the entrance to your house looks like an art gallery front and people wander in through the open door, they are not breaking any rules and they are firmly within the social norm as well.
Indeed. So which "social norm" is it that condones using other people's property and paid-for services - likely incurring a direct and immediate negative impact to them - without their permission ? In what way is this like walking into an art gallery or through someone's unfenced garden ?
Apparently it does. People like you seem to have a problem with the fact that the firmly established social norms around computers don't always involve human intervention and often rely on machine-readable rather than human-readable information.
If it doesn't involve humans, it's not a "social norm". Not unless there's been some pretty big AI breakthroughs in the last couple of days. When "permission" is "granted" to complete a TCP connection, for example, tells you nothing about the intent of the person who setup the server on the other end.
It is standard that the absence of access controls implies permission to use the service.
No, it's not. By that logic, it's OK for a systems administrator with root-level access to read other people's email, because there's an "absence of access controls" for him.
That is not only a sensible and long established social norm - where legislators have a clue it's also codified in law. Apparently you don't just want people who install computers with open services to be exempt from having to follow the established norms, you want to give them the right to lash out at others for not tiptoeing around their misconfigured access points.
No, I want them to be able to have the same expectations that others will respect their property they can with everything else they own or pay for. I want to live in a society where I don't need legislation to stop people acting like selfish arseholes.
But even if it weren't, DRM - by definition - must have a negative impact on users. It takes up resources that could otherwise be put to use in the interest of the user.
Only when it's being used - and since the alternative to not using it is not being able to see the content at all, it's difficult to see how that is a "negative impact".
In the long list of non-arguments against Vista that gets constantly rehashed on Slashdot, "DRM" is probably the biggest (with "high hardware requirements" a very close second).
What is worse: a broken HTTP link in a mail, or a mail with an obsolete Word document attached? At least in the first case you know something has changed.
A broken link, by far. At least with an obselete Word document you have a decent chance of being able to find something that will open it - a broken HTTP link gives you absolutely nothing.
Much of the things (at least most inter-office document passing) people do with mail attachments should be done by keeping those documents under version control, and pointing in the mail to the document in the repository (and maybe a specific version).
I agree, but the problem is those solutions still suck from a usability perspective - which is why people are using email like they do.
Microsoft can optimize resource loading and speed up the parsing of the Registry, but its not going to achieve OS X-like speed without major changes to the architecture of Windows.
Windows has been loading services and drivers asynchronously since the first release of Windows XP, in 2001 - and I'm sure you'll be happy to say Windows XP wasn't a "major change to the architecture of Windows".
Proud of a company that uses weird tactics and the power of its monopoly to totally crush its competition? A company that (to better profit) tries to pervert the idea of the free and open software (and destroy it)?
A company that behaves like every other one, just with more impact than most.
You're looking at it backwards. The real (and honest) question is, how long until the Linux and OSX cheerleaders realize "Yep, Vista isn't as bad as all that!"?
It doesn't have to tell me anything, because the intent is irrelevant if a normal interaction doesn't give me a chance to know that I'm not welcome.
My "normal" interaction with a car that has the keys in the ignition doesn't give me a chance to know I can't take it, either.
Looking like a publicly accessible wireless LAN implies permission to connect, just as looking like a store implies permission to enter.
Let me know with how taking a few things off the shelves at that store and walking out goes.
That would be a catastrophic thing to tell people, because we interact with other people's property based solely on implicit or automatic permission all the time. You could hardly set foot in a city otherwise.
Untrue. We rarely interact with other people's property without their explicit or implicit (through well-established social norms) permission - and in those rare situations where neither are present, we almost always err on the side of caution. Which is why everyone knows you can walk into a pub without a worry (assuming you're of age), but has no problem understanding they're not allowed to walk behind the bar and fix their own drinks.
Apparently, however, the whole world changes when dealing with computers. Which is why wardriving the neighbourhood for an unsecured WAP to leech porn through is A-OK, just because it'll DHCP you an IP address.
It's even clearer with wireless networks, because you are taking a piece of PUBLIC property and refuse to cooperate with other people who are in the same space.
Complete and utter tripe.
It is not hard to clearly state your intent, but you insist on reducing the usefulness of the technology just so that you can keep using a public frequency band without bothering about the conventions of the medium. That is truly impolite and disrespectful.
You are beating a straw man (heck, you're beating a whole pack of them). Nothing in any of my arguments suggests reducing the utility of intentionally free wifi networks and *especially* of the "public frequency band" (this latter part being the most ridiculous, given all wifi networks use the same frequencies - so you're saying private ones are abusing it as well).
GTFO. It's the most relevant point in the list. That's why it's there three times.
It's completely and utterly irrelevant. If you don't have DRM-encumbered media, it simply doesn't apply. If you *do* have DRM-encumbered media, Vista isn't imposing any more restrictions than any other device you use to play that media will.
Finally, DRM is an attribute of the media, not the player. In no way is the ability of Vista to play DRM-encumbered media a "bad" thing.
And I seriously want to know, what is it about Vista that compels you to upgrade to it?
Right now ? The search, UAC and being able to use "Switch User" while part of a domain. MCE is also very nice.
In the future, the updates to the internals will mean my future machines with 4+ cores, 4+GB RAM and beefy video cards will perform better than they would with XP. I also expect there are improvements in the network manageability, but I haven't had time to properly evaluate them.
It should also be more stable and easier for Microsoft to maintain, making it more reliable - but since a) I don't have problems with XP's stability or reliability and b) don't have any practical way to measure that, I wouldn't consider them "compelling" per se.
I'm arguing that it is reasonable to assume that you're welcome to have your computer automatically connect to open wireless access points, because to a computer they look exactly like intentionally open access points.
And I'm pointing out that when your computer automatically connects to a WAP that tells you _nothing_ about the intent of the person who owns that network.
I also explained that requiring explicit access permission by law is a strategic mistake because it leaves more people vulnerable in the long run by hiding the fact that in even more places the law does not even attempt to protect these people from passive eavesdropping.
Except it doesn't because it promotes the idea that people should assume if they don't have permission to use someone else's property, they shouldn't. Like I keep saying, it's a damn shame we now need legislation to achieve this.
I should NOT have to explicitly stop someone using my stuff, they should assume they need to ask permission first. It's called being polite and having respect.
No, that's a biased choice. There are very few free public phones, for example. There are many intentionally free wireless networks however.
No, there aren't - at least if you define "many" in a meaningful fashion to mean a significant proportion of all WAPs or want to try and define "intentionally free" as "unsecured" (or even more laughably, try to include all those cafes that let you use their wifi on the proviso you buy coffee as "intentionally free").
The social norm with computers is generally that if you can access it remotely without providing credentials, access is allowed. An access point is a computer, and it is frequently installed and used by businesses, cities and individuals according to that social norm.
To most people, an "access point" is not a computer, any more than a DVD player is. Further, as I've been trying to get across, access to a WAP tells you nothing about the intent of the person who owns it, and especially nothing about their intent with regards to other things connected to that network.
The sole purpose of an Access Point (what's in a name) is to provide access to a network.
Indeed, but a network and the resources on it (or accessible via it) are different things.
There is no assumption, there is a (perfectly formilized even) request for access to a (publicly advertised) network. The answer can be either yes or no. But when the answer it turns out to be yes, why should I still assume it actually meant no?
Because it tells you nothing about the intent of the person who owns the network. Any more than the TCP/IP RFCs tell you anything about the intent of someone who owns a server.
That exactly the same with a webserver, a doorbell and borrowing a car, you communicate a request, and may or may not receive an answer. But when you do recieve an answer and it happens to be yes you recieved permission.
Your problem is that you are inferring more from the answer than that answer is able to give.
By your logic, hacking into a server is perfectly ok. After all, if it completes the TCP/IP negotiation, that implies you can do whatever you want once that negotiation is complete.
You first, the http specification (you know, the webserver stuff) is found in
What ? I made neither assertion nor implication that would require me to do so. You're the only one insisting a technical description of protocol negotation(s) are an indication of a person's intent.
You personally will have a very hard time arguing just that when someone links you to this very slashdot discussion.
Why ? A handful of nerds saying they run open WAPs are not in any way a meaningful representation of society in general.
Put pulling up the user manual for a AP will pretty much do the same in most cases. Failing to read it should not be and excuse.
Running an unsecured WAP should not be seen as an invitation to use someone else's network resources without permission, any more than a car with the keys in the ignition should be seen as an invitation for joyriding.
No, that is not what i'm arguing. What i'm arguing is that I ask permission, through a well established procedure and when this permission is granted it is logical to assume the permission is granted.
So I assume you're also ok with things like brute force dictionary attacks on SSH servers ? Or, indeed, other forms of cracking that don't require subverting standard protocols ?
Well that would be today then, in my neighbourhood you can usually just ring a bell and ask if you can make a phone call. I really doubt any of my neighbours would have a problem with that, I sure would glad to help you. That's just it, asked permission, got permission, encountered no fence, did not bypass anything clearly intended to keep me out, nobody asked me to leaf, did nothing wrong.
Different situation. You are confirming a person's intent, not leveraging a machine's compliance to a technical specification.
This applies to your estate, why shouldn't it work the same for a network?
No, it doesn't. Even if trespassing laws in your location say you are allowed to walk onto private property, I'm sure you're not allowed to use their phone, eat their food, sleep in their bed, etc.
Which means that a minimal indication that the owner of the network does not want me around should be enough, just like even the most stupid fence is enough and I will respect that. It also means there is a serious difference between accessing the network to get to the internet and actually accessing the computer that may be on the network.
Actually it doesn't, because the network and the computers are separate things. But I'm pretty sure from what you've said so far, that you'd also consider it OK to copy anything off a network share on one of those machines.
That is also very true when accessing private property, you can go there, but you can't be looking inside the cabinets or read the mail.
But apparently using someone else's internet connection through an unsecured WAP without their permission is ok ?
Not necessarily. You can have a password-protected website via HTTP authentication -- just like you can have a password-protected access point.
Indeed you can, but I fail to see how this is relevant. Password-protecting a website both a) takes effort on the part of the person doing so and b) clearly indicates the user does not want the content to be publically available.
Your distinction makes sense in the context of informally agreed-upon netiquette, but the law doesn't go far enough to make that distiction. That's what I was pointing out -- that the law really is too vague in this matter.
The law will care about the intent of the owner of the network and, to a lesser degree, the owner of the machine accessing the WAP. Since there is no way to definitively ascertain the intent of the network's owner simple based on whether or not their WAP is unsecured and you cannot reasonably assume that an unsecured WAP means the owner specifically intends it for public use, the "law" will expect most people to either a) not connect to it at all, or b) disconnect immediately if their computer "automatically" connects.
This is no different to any other situation involving other people's private "property". If you can't definitively ascertain their intent you should err on the side of caution - and that's what the law expects you to do.
Slashdot Mirror
I should have been clearer in my post. While the disk performance in normal operation may be comparable hardware RAID does have several key advantages. The first being that with anything other than RAID 0 or 1 the cpu hit from the RAID driver during heavy use will be high even on a high end machine.
No, it won't. Even a 500Mhz P3 has a RAID5 checksumming speed of ~1GB/sec. A current low-end CPU (eg: 1.6Ghz Pentium E2140) has a RAID5 checksumming speed of around 4GB/sec.
Suffice to say your average SATA array that's unlikely to even get much over 150MB/sec isn't going to put much load on any remotely modern CPU (at least not from the checksumming).
This is fine if you aren't using that machine for anything else but if it is a desktop machine or a server that is required to do anything else other than serve files then it does cause a problem. This is even more apparent if the array becomes degraded as it will take significantly more cpu time to perform all the parity calculations to rebuild the array.
I'm not sure why you think parity calculations when the disk is rebuilding are any different from parity calculations when the disk is being written to normally, but they're not.
The real hit you take from software RAID is to the bus bandwidth, not the CPU. Most amateurs compare a hardware RAID controller to software RAID on a dinky little 32bit/33Mhz PCI bus (and remember that those onboard SATA ports are probably hanging off a regular 32bit/33Mhz PCI bus), that's why they frequently conclude software RAID is slower ("especially during rebuilds"). When you have a system with ample bus bandwidth the situation is quite different.
If you do need to rebuild the array then a decent RAID card will handle it without putting a heavy load on the server and significantly without affecting disk perfomance.
This is impossible. There will always be a performance hit during a rebuild, no matter whether your RAID logic is running on the system CPU or the embedded CPU on a RAID controller. Further, that performance hit comes from the greater number of IOPS necessary while the array is rebuilding and has nothing to do with "parity calculations" (which even a 10+ year old commodity CPU can do faster than any normal (not to mention most abnormal) arrays could ever hope to be).
In a server that is in constant use this is a key point. Rebuilding a RAID 5 array in software often reduces the disk performance to very low levels, effectively denying access to data until the rebuild is complete.
The difference you are seeing is almost certainly because most hardware RAID controllers throttle rebuilds by default to be relatively slow, so "normal" disk access suffers as little as possible. Software RAID can do this as well - although most HOWTOs tell you to bump the rebuild speed up as high as possible, which is probably why you see the results you do. The downside, of course, that the longer rebuild means a bigger window where your array's performance - and more importantly, reliability - are degraded.
I just know how to spend wisely; $150 - $200 on a proper raid card [...]
Is a complete waste of money. Software RAID will be faster, more reliable and more flexible.
The only reason to use hardware RAID is for the transparency, easier initial setup and (maybe, depending on your needs or hardware) hot-swap.
If your data is valuable enough to use raid, then hardware raid is the only way to go. If you have a problem with the os the data is lost.
Rubbish. Software RAID arrays can generally be migrated between machines that aren't even the same architecture, a simple "OS problem" isn't going to damage them.
With a software RAID, you do lose some performance [...]
No, you don't - quite the opposite in fact (unless your hardware is seriously underpowered).
Yes migration of will probably happen over time, but that is why there is no immediate adoption of Vista, there is no killer feature and the hardware requirements are more onerous.
In this it is really no different than 2000/XP and NT4. There was no real "killer feature" there, either, and the hardware requirements were relatively high (from usable on a 32MB, ~100Mhz Pentium to a 64MB (2000) 128MB (XP) ~300Mhz P2). The closest was Active Directory and, in particular, Group Policy - but even with those two enticements people were running NT4 Domains well into the '00s.
Indeed, I expect the migration from NT4 to Windows 2000/XP and the upcoming migration from XP to Vista/Server 2008 have a great deal of similarity, both in concept and magnitude, and I expect it to take a similarly long timeframe to finish.
I did some 2000 (and even Win98) to XP upgrades in 2003 where the hardware, though older was still capable of running XP within a corporate environment quite well, this is likely not going to be possible with Vista, because the minimum specs are much higher than what I see on a normal corporate desktop (and I'm not assuming that business users will want Aero, 512Mb RAM seems about the norm at the moment (with 256 still common) and importantly low end and low capacity HDD's and processors (After all its much cheaper to pick up new clients with minimum specs)).
A significant proportion of PCs bought in the last 12 - 18 months will have shipped with 1G RAM, which is all you need for Vista in a typical business setting. Even 512M is usable, although definitely "slow".
HDD size and processor power are basically irrelevant. Anything from 1GHz (ca. 1999) and up is fine and 20GB+ drives - even on budget machines - have been commonplace for years.
Interestingly, I also doubt that the migration will be piecemeal like it could be from 2000 to XP, an environment where either client could co-exist peacefully and without issue, I have it on moderately decent authority* that dropping a Vista box in as a straight replacement for an XP box and not making any changes in other areas (AD primarily, whether group policy or user profiles and scripts etc..)is not possible, whilst in a 2000 to XP transition it was (XP being fully backward compatible but without too many fundamental changes, you simply didn't get some of XP's other benefits with regard to group policy etc..)
The only really big potential hiccup I've seen in my (admittedly cursory) look into Vista migrations is that it can't use existing roaming profiles. This is certainly not an insignificant problem, but if you're doing things "properly" with Folder Redirection, it's manageable.
However, I just don't expect fast migrations to Vista, and I doubt anyone else whose job isn't in marketing does either. The simple truth is that Windows hit the "good enough" stage in the early 2000s and convincing people there are genuine business improvements to be had from upgrading from there will be difficult. Windows Server 2008 might have some new "Vista-only" features that make the upgrade more compelling, but we'll have to wait and see.
What about all the companies that are at that point in the cycle now... ?
That's why I said "most", not "all" (or "none").
Why would a business want to buy additional hardware for this upgrade?
They wouldn't - but that wasn't the issue. The issue was the implication that hardware for Vista is (more) expensive.
They wouldn't need to if they didn't upgrade.
Actually they probably will, once the lease on their existing hardware runs out.
If a business has thousands of desktops from a few years ago, say 1.8Ghz CPU, 512Mb RAM, 40Gb hard disk etc... a bog standard corporate desktop, its not cheap to replace all that kit in one go with something vista will run on, its much easier (and more cost efficient) to upgrade as stuff breaks and then at some distant point in the future upgrade to vista because all your equipment is capable of running it and it fits in with the other infrastructure life cycles that you are working within.
Which is exactly how a migration to Vista will be done, just like it was exactly how the migrations to XP, 2000, NT, etc, etc were handled before that.
Why anyone thinks it odd that businesses aren't migrating to Vista immediately - or tries to imply it has problems because of that - is beyond me. Most companies, running a ~3yr hardware cycle, won't even be _considering_ a Vista rollout until mid 2008 at the earliest. This is completely and utterly independent of any technical issues - real or imagined - that exist.
Microsoft knows this (which is why aberrations are worth talking about). Anyone with any real experience in corporate IT knows this. The only people who seem not to know it are the anti-Microsoft trolls on Slashdot who, for some reason, think because corporate IT departments are following the same rollout methodology they've been using for the last couple of decades, it means Vista has problems.
Vista may become viable as hardware becomes cheaper [...]
Huh ? An entry level <US$400 PC from Dell will run Vista fine. How much cheaper do you think a new PC can get ?
While there are numerous genuine reasons why an upgrade to Vista isn't compelling, there are far more stupid ones - and "expensive hardware" (along with "DRM") would have to be one of the dumbest.
Even a cursory google would confirm this for you.
My "cursory googles" only show up references to Peter Gutmann's paper, which is both mostly FUD and mostly refuted - and added to *that* most of the pages based off Gutmann's paper are extrapolating to conclusions even he didn't draw.
Either accept it as a given for the sake of argument, or don't bother responding. This is the last I have to say on the matter.
Suit yourself, but you're still wrong.
A protected path DRM implementation has to be operational at all times by definition. If it isn't operational, the path isn't protected.
No, it only needs to be operational when "protected" media is being played. If the content isn't DRM-encumbered, the path doesn't *need* to be protected. This is exactly how Vista works - if the media isn't DRM encumbered, the DRM system isn't active.
Incidentally, it is trivially simple to disprove your assertion, simply by playing HD content without DRM on a Vista machine through an analogue output like VGA. The media plays at full resolution, which it would not do if the DRM system is always active (since the output path isn't "protected").
You'd be somewhat correct in the case of the old style windows media DRM; that could (in theory) be written in a modular way similar to decoding plugins. However, it would still be taking up drive space without giving any benefit to the owner of the drive.
The benefit is being able to play media they otherwise wouldn't. You may not consider that a benefit, but people who want to watch the movies certainly will.
If Microsoft stood up to Sony/whoever backs HD DVD, they'd have to either give up the draconian DRM or the Windows platform. Guess which they'd choose?
Windows, unquestionably. Most people will be playing their DRM-encumbered media - like they do now with their regular media - through games consoles and cheap set-top players.
I'm not sure where this (apparently common) idea that Microsoft has significant sway over the media companies comes from. It certainly doesn't follow any rational analysis, given that Microsoft's presence in the "content creation" market is essentially nonexistent and even in the "content player" market is quite small - the vast majority of content is consumed through appliances like set-top players, games consoles, iPods, and the like, not PCs running Windows (or anything else, for that matter).
Regardless, we'll all still be able to see the content, because all DRM is cracked if it protects anything worth having.
Of course - which strengthens my argument about the irrelevancy of DRM even more.
The iMac shipped with three replacements for a floppy drive. The first was the Ethernet port...remember that the "i" in the name stood for "Internet."
Sorry, but in 1998 internet connections - especially broadband ones - were *far* from ubiquitous.
The other was the USB port--flashdrives were well on their way to becoming common when the iMac hit the market.
The first flash drives weren't on the market until 2000, so they sure as hell weren't "common" in 1998. Further, since Windows didn't have built-in support for them until Windows 2000 and Windows ME, they weren't "common" until a couple of years after that. It would have been brave indeed to walk around any time before about 2002 with the assumption your USB key would just "plug and play" with the majority of computers you'd encounter.
The whole idea was that data would move across the pipe more and more, so the floppy was not necessary.
Indeed, there was nothing wrong with the idea - the problem was that it was half a decade too early.
The success of the iMac, and the way we work now, shows that was an accurate prediction.
The complaints and subsequent boom in the late 90s for USB floppy drives suggest lots of people thought the iMac's lack of any easy and compatible method to transfer data was a problem.
You're being intentionally dense.
No, I'm pointing out that "walking into the store" and "taking stuff off the shelves" - like "connecting to a WAP" and "using services via that WAP" - are different things. Permission to do one does not in any way imply permission to do the other.
The implicit permission is established by a social norm, which takes into account that the person who is supposed to follow that norm must have a reasonable chance to recognize what is expected of him. If your garden looks just like the park that it borders on and there is no sign or fence to tell people that it is private property, then no social norm and no law forbids people to enter your garden. If the entrance to your house looks like an art gallery front and people wander in through the open door, they are not breaking any rules and they are firmly within the social norm as well.
Indeed. So which "social norm" is it that condones using other people's property and paid-for services - likely incurring a direct and immediate negative impact to them - without their permission ? In what way is this like walking into an art gallery or through someone's unfenced garden ?
Apparently it does. People like you seem to have a problem with the fact that the firmly established social norms around computers don't always involve human intervention and often rely on machine-readable rather than human-readable information.
If it doesn't involve humans, it's not a "social norm". Not unless there's been some pretty big AI breakthroughs in the last couple of days. When "permission" is "granted" to complete a TCP connection, for example, tells you nothing about the intent of the person who setup the server on the other end.
It is standard that the absence of access controls implies permission to use the service.
No, it's not. By that logic, it's OK for a systems administrator with root-level access to read other people's email, because there's an "absence of access controls" for him.
That is not only a sensible and long established social norm - where legislators have a clue it's also codified in law. Apparently you don't just want people who install computers with open services to be exempt from having to follow the established norms, you want to give them the right to lash out at others for not tiptoeing around their misconfigured access points.
No, I want them to be able to have the same expectations that others will respect their property they can with everything else they own or pay for. I want to live in a society where I don't need legislation to stop people acting like selfish arseholes.
It is impacting non DRM files.
No, it's not.
But even if it weren't, DRM - by definition - must have a negative impact on users. It takes up resources that could otherwise be put to use in the interest of the user.
Only when it's being used - and since the alternative to not using it is not being able to see the content at all, it's difficult to see how that is a "negative impact".
In the long list of non-arguments against Vista that gets constantly rehashed on Slashdot, "DRM" is probably the biggest (with "high hardware requirements" a very close second).
No.
Yes.
The fact that a Microsoft presentation program can't do this on the latest Microsoft operating system is pathetic.
It (apparently) wasn't a "Microsoft presentation program", it was a HP Craptastic BundleWare DVD Player.
What is worse: a broken HTTP link in a mail, or a mail with an obsolete Word document attached? At least in the first case you know something has changed.
A broken link, by far. At least with an obselete Word document you have a decent chance of being able to find something that will open it - a broken HTTP link gives you absolutely nothing.
Much of the things (at least most inter-office document passing) people do with mail attachments should be done by keeping those documents under version control, and pointing in the mail to the document in the repository (and maybe a specific version).
I agree, but the problem is those solutions still suck from a usability perspective - which is why people are using email like they do.
Microsoft can optimize resource loading and speed up the parsing of the Registry, but its not going to achieve OS X-like speed without major changes to the architecture of Windows.
Windows has been loading services and drivers asynchronously since the first release of Windows XP, in 2001 - and I'm sure you'll be happy to say Windows XP wasn't a "major change to the architecture of Windows".
Proud of a company that uses weird tactics and the power of its monopoly to totally crush its competition? A company that (to better profit) tries to pervert the idea of the free and open software (and destroy it)?
A company that behaves like every other one, just with more impact than most.
The problem is his balls are terribly small. Steve is too scared of Microsoft and his own user base to release OSX for all computers.
Steve has no fear with regards to Microsoft when it comes to releasing OS X for regular PCs - why would he ?
Steve's fear is of companies like Dell and HP, because they'll sell cheaper (and more configurable) Macs than he does.
You're looking at it backwards. The real (and honest) question is, how long until the Linux and OSX cheerleaders realize "Yep, Vista isn't as bad as all that!"?
About the same time Windows 7 is released.
It doesn't have to tell me anything, because the intent is irrelevant if a normal interaction doesn't give me a chance to know that I'm not welcome.
My "normal" interaction with a car that has the keys in the ignition doesn't give me a chance to know I can't take it, either.
Looking like a publicly accessible wireless LAN implies permission to connect, just as looking like a store implies permission to enter.
Let me know with how taking a few things off the shelves at that store and walking out goes.
That would be a catastrophic thing to tell people, because we interact with other people's property based solely on implicit or automatic permission all the time. You could hardly set foot in a city otherwise.
Untrue. We rarely interact with other people's property without their explicit or implicit (through well-established social norms) permission - and in those rare situations where neither are present, we almost always err on the side of caution. Which is why everyone knows you can walk into a pub without a worry (assuming you're of age), but has no problem understanding they're not allowed to walk behind the bar and fix their own drinks.
Apparently, however, the whole world changes when dealing with computers. Which is why wardriving the neighbourhood for an unsecured WAP to leech porn through is A-OK, just because it'll DHCP you an IP address.
It's even clearer with wireless networks, because you are taking a piece of PUBLIC property and refuse to cooperate with other people who are in the same space.
Complete and utter tripe.
It is not hard to clearly state your intent, but you insist on reducing the usefulness of the technology just so that you can keep using a public frequency band without bothering about the conventions of the medium. That is truly impolite and disrespectful.
You are beating a straw man (heck, you're beating a whole pack of them). Nothing in any of my arguments suggests reducing the utility of intentionally free wifi networks and *especially* of the "public frequency band" (this latter part being the most ridiculous, given all wifi networks use the same frequencies - so you're saying private ones are abusing it as well).
It's difficult to comprehend how someone could call Windows 95 and 98SE "OK" and say NT4 "sucks".
GTFO. It's the most relevant point in the list. That's why it's there three times.
It's completely and utterly irrelevant. If you don't have DRM-encumbered media, it simply doesn't apply. If you *do* have DRM-encumbered media, Vista isn't imposing any more restrictions than any other device you use to play that media will.
Finally, DRM is an attribute of the media, not the player. In no way is the ability of Vista to play DRM-encumbered media a "bad" thing.
And I seriously want to know, what is it about Vista that compels you to upgrade to it?
Right now ? The search, UAC and being able to use "Switch User" while part of a domain. MCE is also very nice.
In the future, the updates to the internals will mean my future machines with 4+ cores, 4+GB RAM and beefy video cards will perform better than they would with XP. I also expect there are improvements in the network manageability, but I haven't had time to properly evaluate them.
It should also be more stable and easier for Microsoft to maintain, making it more reliable - but since a) I don't have problems with XP's stability or reliability and b) don't have any practical way to measure that, I wouldn't consider them "compelling" per se.
I'm arguing that it is reasonable to assume that you're welcome to have your computer automatically connect to open wireless access points, because to a computer they look exactly like intentionally open access points.
And I'm pointing out that when your computer automatically connects to a WAP that tells you _nothing_ about the intent of the person who owns that network.
I also explained that requiring explicit access permission by law is a strategic mistake because it leaves more people vulnerable in the long run by hiding the fact that in even more places the law does not even attempt to protect these people from passive eavesdropping.
Except it doesn't because it promotes the idea that people should assume if they don't have permission to use someone else's property, they shouldn't. Like I keep saying, it's a damn shame we now need legislation to achieve this.
I should NOT have to explicitly stop someone using my stuff, they should assume they need to ask permission first. It's called being polite and having respect.
No, that's a biased choice. There are very few free public phones, for example. There are many intentionally free wireless networks however.
No, there aren't - at least if you define "many" in a meaningful fashion to mean a significant proportion of all WAPs or want to try and define "intentionally free" as "unsecured" (or even more laughably, try to include all those cafes that let you use their wifi on the proviso you buy coffee as "intentionally free").
The social norm with computers is generally that if you can access it remotely without providing credentials, access is allowed. An access point is a computer, and it is frequently installed and used by businesses, cities and individuals according to that social norm.
To most people, an "access point" is not a computer, any more than a DVD player is. Further, as I've been trying to get across, access to a WAP tells you nothing about the intent of the person who owns it, and especially nothing about their intent with regards to other things connected to that network.
The sole purpose of an Access Point (what's in a name) is to provide access to a network.
Indeed, but a network and the resources on it (or accessible via it) are different things.
There is no assumption, there is a (perfectly formilized even) request for access to a (publicly advertised) network. The answer can be either yes or no. But when the answer it turns out to be yes, why should I still assume it actually meant no?
Because it tells you nothing about the intent of the person who owns the network. Any more than the TCP/IP RFCs tell you anything about the intent of someone who owns a server.
That exactly the same with a webserver, a doorbell and borrowing a car, you communicate a request, and may or may not receive an answer. But when you do recieve an answer and it happens to be yes you recieved permission.
Your problem is that you are inferring more from the answer than that answer is able to give.
By your logic, hacking into a server is perfectly ok. After all, if it completes the TCP/IP negotiation, that implies you can do whatever you want once that negotiation is complete.
You first, the http specification (you know, the webserver stuff) is found in
What ? I made neither assertion nor implication that would require me to do so. You're the only one insisting a technical description of protocol negotation(s) are an indication of a person's intent.
You personally will have a very hard time arguing just that when someone links you to this very slashdot discussion.
Why ? A handful of nerds saying they run open WAPs are not in any way a meaningful representation of society in general.
Put pulling up the user manual for a AP will pretty much do the same in most cases. Failing to read it should not be and excuse.
Running an unsecured WAP should not be seen as an invitation to use someone else's network resources without permission, any more than a car with the keys in the ignition should be seen as an invitation for joyriding.
No, that is not what i'm arguing. What i'm arguing is that I ask permission, through a well established procedure and when this permission is granted it is logical to assume the permission is granted.
So I assume you're also ok with things like brute force dictionary attacks on SSH servers ? Or, indeed, other forms of cracking that don't require subverting standard protocols ?
Well that would be today then, in my neighbourhood you can usually just ring a bell and ask if you can make a phone call. I really doubt any of my neighbours would have a problem with that, I sure would glad to help you. That's just it, asked permission, got permission, encountered no fence, did not bypass anything clearly intended to keep me out, nobody asked me to leaf, did nothing wrong.
Different situation. You are confirming a person's intent, not leveraging a machine's compliance to a technical specification.
This applies to your estate, why shouldn't it work the same for a network?
No, it doesn't. Even if trespassing laws in your location say you are allowed to walk onto private property, I'm sure you're not allowed to use their phone, eat their food, sleep in their bed, etc.
Which means that a minimal indication that the owner of the network does not want me around should be enough, just like even the most stupid fence is enough and I will respect that. It also means there is a serious difference between accessing the network to get to the internet and actually accessing the computer that may be on the network.
Actually it doesn't, because the network and the computers are separate things. But I'm pretty sure from what you've said so far, that you'd also consider it OK to copy anything off a network share on one of those machines.
That is also very true when accessing private property, you can go there, but you can't be looking inside the cabinets or read the mail.
But apparently using someone else's internet connection through an unsecured WAP without their permission is ok ?
Not necessarily. You can have a password-protected website via HTTP authentication -- just like you can have a password-protected access point.
Indeed you can, but I fail to see how this is relevant. Password-protecting a website both a) takes effort on the part of the person doing so and b) clearly indicates the user does not want the content to be publically available.
Your distinction makes sense in the context of informally agreed-upon netiquette, but the law doesn't go far enough to make that distiction. That's what I was pointing out -- that the law really is too vague in this matter.
The law will care about the intent of the owner of the network and, to a lesser degree, the owner of the machine accessing the WAP. Since there is no way to definitively ascertain the intent of the network's owner simple based on whether or not their WAP is unsecured and you cannot reasonably assume that an unsecured WAP means the owner specifically intends it for public use, the "law" will expect most people to either a) not connect to it at all, or b) disconnect immediately if their computer "automatically" connects.
This is no different to any other situation involving other people's private "property". If you can't definitively ascertain their intent you should err on the side of caution - and that's what the law expects you to do.