A Senator/Congressman will only be able to fire so many aides for sneaking in legislation before the public will say "maybe the problem isn't with the aides."
Similarly, an entire House (forgive me if I get the terminology wrong, I'm not American) can only sneak through so much stupid legislation, ship young men off to war without a clear reason, set up prison camps which bypass the Geneva Convention and make excuses for their own soldiers torturing Iraqi prisoners before the public will say "maybe the problem isn't in Iraq".
I'll probably get modded down for this, but I've got karma to burn and I still think the point stands.
I've yet to find anything which has 100% compatability. Everything I've found depends on a plugin installed for Outlook, and the plugins tend to vary in terms of how good they are.
One thing they frequently don't offer, which Exchange/Outlook does, is realtime updating of free/busy information. It doesn't sound like much, but without it you've got a massive race condition every single time someone schedules a meeting. The first manager to get tripped up by this will have IT's guts for garters.
Actually, that's not strictly true. Scalix claims 100% compatability, complete with support for HA, clustering with Exchange and various other things, but I've looked into that and found the licensing costs for that particular edition to be comparable to Exchange itself. Additionally, while it claims to have great integration with Active Directory, it doesn't integrate with any other directory services. So you'd still need an AD domain, which right now implies at least two Windows servers and the appropriate licensing.
I am finally forced to agree that OOo is no replacement for corporate uses of MS Office. It's not OOo's problem, though; it's MSO's problem.
Except that Microsoft hold all the power, and will do until such time as some of the large-scale highly publicised OO.o migration plans come to fruition.
While it's technically Microsoft's problem, don't for one minute imagine that they consider it to be a problem.
What is stopping you from creating your own SNMP agents ? Unless you some some esoteric piece of hardware, it is fairly easy to create one.
A few things:
1. I'm the IT manager in a company which develops embedded networked devices. We've got all sorts of weird and wonderful stuff.
2. The same thing which stops me putting Net-SNMP onto a server on a non-firewalled connection. SNMPv1 offers very poor security, and ZenOSS doesn't support SNMPv3 security.
3. I trust a network probe to confirm that a service is still working properly a lot further than I trust an SNMP agent which is essentially going to be doing little more than ensuring that the process is running and listening on the port - while the SNMP agent may be technically correct in its assesment, it hasn't actually tried to get any sense out of the process.
At my current employer, I'm using Nagios (which is great as far as it goes, but needs a lot of configuration) and looking towards ZenOSS (which requires a lot less configuration because it can get everything natively through SNMP where available, but isn't anywhere near so flexible in terms of defining your own tests if SNMP isn't an option),
Ideally I'd have a solution which offered the SNMP support and out of the box functionality of something like ZenOSS, while at the same time being dead easy to extend to run scripts and check things which can't easily be checked with SNMP. I don't have a lot of time to invest in the initial setup - obviously I'm prepared to invest more time in extending where necessary, but SNMP support should be as simple as "name the host and the SNMP community".
Why can't you merge changes back to BSD-licensed code?
Nothing. But if I merge changes into a piece of BSD-licensed code, there's nothing to stop my competitor 20 miles down the road downloading this piece of code complete with my changes, making a few minor tweaks to it, keeping those tweaks private and selling it - even though I provided a lot of the code which his product is based on.
With the GPL, he can still make tweaks to it but he's got to make those tweaks freely available to anyone he distributes code to. Many choose to simply submit these tweaks as patches upstream rather than maintain their own fork of the software - not really a lot of point in being that anal when the license explicity allows your customer to do that anyway.
What we're seeing happening now - particularly in the embedded space - is that manufacturers are taking the free stuff in Linux, tweaking the kernel and submitting changes where necessary but keeping the majority of their proprietary code logically separate in userland so they don't have to GPL it. Hence why you can have a router based on Linux which is technically open source, but the clever stuff (eg. removing the complication from configuring iptables with a web app, a means of holding firewall rules and some glue to turn these into iptables commands) remains private.
I'm talking about people like the late lamented John Peel
Now that is very true. A legend in his own lifetime. He must have listened to loads of crap, as you say - but the difference was that was his entire livelihood, so he did have the time to do it.
Here in the UK, there is 1 (yes one) company which pretty much has a stranglehold on about 90% of the bricks & mortar large computer stores. While smaller ones exist, I'm only talking about the large places with a reasonable selection that tend to be located in out of town industrial areas, since that's the fairest comparison with CompUSA and BestBuy.
It's the closest thing we have to Best Buy or CompUSA.
It's called PC World, and it's run by a company called the Dixons Stores Group.
Let's just say: £80 network cards (that was apparently the cheapest), £20 USB cables (again, the cheapest), £1,200 PCs. They stock cheaper ones but I've never met anyone who ever bought one - I suspect those who are savvy enough don't go there in the first place, and those who aren't are generally told "You want to send email? Well, in that case you'll need this...". And extended warranties which cost 70% of the value of the PC, yet are serviced by spotty 16 year olds who wouldn't know a PC if it dropped on their head.
I think you'll probably find that the reliability issue with laptop hard drives has more to do with them spending most of their working life being thrown around in a laptop than any inherent problem.
You can't do that. The terms for all of Microsoft's site licenses state "this is an UPGRADE, and as such can only be applied to other versions of our OS. The full licensed version must still be purchased at some point for every system in your business, either preinstalled on the PC as an OEM copy or purchased as the boxed retail product".
It's a fine line, and it's an entirely artificial distinction as the CDs you get when you buy a site license aren't anything particularly special - they don't check, for instance, that you already have a valid OS to upgrade from. But breaking that rule is a very good way to have the BSA come down heavily on you, and that will almost certainly cost an awful lot more than just buying the right licenses in the first place.
(For the purposes of simplicity, I'm overlooking the fact that most of those licenses are quite obviously worded in the most awful way in order to set you up for failure).
Just sounds like some young buck made a bad decision without due dilligence and is passing the buck to Dell.
So sounds like someone without a lot of knowledge and poor problem resolution skills.
That or astroturfing FUD. IME, it's practically impossible to find a SCSI controller which can't be made to work like a charm in Linux so the OP is, IMO, either lying or incompetent.
Funny, I'm running a couple of x8xx series servers (which also use an LSI-based RAID controller) and they work just fine.
The only thing that's closed-source is the program which configures, manages and gets diagnostics out of the RAID controller. This, I accept, is a problem because without it you can't usefully monitor things like disk failure. But you can certainly load the OS without anything proprietary.
All very true, but Postgres' defaults are extremely conservative, and as a result tend to produce a system which works beautifully, but somewhat slowly.
In the Windows world, it's relatively unusual to have to do that much tweaking to a piece of software to get the most out of it.
rpm -e --force e2fsprogs-libs will do it nicely. Of course it's trivially easy to fix with a rescue CD, but apparently ESR hasn't noticed the rescue option when installing.
It would also be trivially easy to setup a list of absolutely vital, do-not-remove, no-really-we-mean-it software which cannot be removed even with rpm --force, and hack rpm to read this list every time it's invoked with the -e switch. Then if you try rpm -e --force this-vitally-important-thing, a big red hairy error comes up saying that you really really don't want to do that.
Document where this file is so if you're a real nutcase you can remove stuff from the list of "packages not to be removed under any circumstances" and it's problem solved.
Of course, if all the packages were perfect in the first place, then there would be no need for such a file. The very fact that you find yourself trying --force would serve as a warning. But I've seen packages which you do have to do that for a fair few times now.
You tried Redhat lately? I'd say they basically lost it a few years ago, shortly after spinning off fedora (and the warning signs had been there for some time before).
Mandrake and YellowDog had done a lot to obviate the most glaring deficiences in rpm with urpmi and yum, but things were still a long way from being perfect.
Now as a Linux user, I find that whenever I want to use anything RedHat based (eg. Fedora, CentOS) frustrating because it's missing fundamental packages which I need. That's OK, I can live with that and roll my own. But then dependencies are missing, and it's frequently quite difficult to find out what "cannot link lib-obscure-random-string" is fixed with which package. Then you find that you're downloading the latest version of a given tool, which was developed on the assumption that you've got the latest version of every library. Except you don't, as your OS was put together and tested 6 months ago and hasn't moved since. So it's back to compiling,
The only thing which amazes me is that ESR has either put up with this for so long or has never been seriously hurt by it.
I'd say ESR is taking a very reasonable approach. Software needs to work - or if it's not quite perfect, it shouldn't be so broken as to be awkward/impossible to fix without source code hacking. At most, it should be a few bits of relatively straightforward configuration away. And for desktop software (I'll make an exception for fundamentally server side stuff like Apache, because the target audience there is supposed to know what it's doing), that configuration should be click-click-click, not click, bring up terminal, spend half an hour wading through a manpage, tweak a config file, pray it worked.
In any event, it should be difficult if not impossible to fundamentally break a system as an end user unless you're putting it to some extreme abuse.
ESR has clearly decided that RedHat is unsuitable for him because it doesn't meet these requirements. Good for him. IMO, I could have told him that 5 years ago.
As others have said, it's only news because it's someone well known in the community. If it was almost anyone else, there would be a collective yawn.
Exactly. "Unsupported" is a magic word used by tech support departments so they can wash their hands of the problem.
Your OS keeps crashing? You're running an unsupported application. Go away and don't come back until you've fixed that.
Still crashing? You must be running unsupported hardware. What's the exact make and model number of every single component in your PC? You don't know? Go away and come back when you do.
Hardware vendors are just as bad:
Your hard disk appears to have failed? Sorry, you're running an unsupported operating system. Go away.
Your power supply has exploded? Sorry, we only support people who don't actually ring up requiring support.
Your power supply has caught fire, destroying your house and all your belongings? [click]
I have my doubts that this article can possibly be true. Remember that the music is only being sold because the big labels of this world - the RIAA (do they cover Canada as well?) say that it may be.
As soon as any of the music stores start selling RIAA-covered music without DRM, expect the RIAA to come down on them like a ton of lawyers.
Slashdot Mirror
A Senator/Congressman will only be able to fire so many aides for sneaking in legislation before the public will say "maybe the problem isn't with the aides."
Similarly, an entire House (forgive me if I get the terminology wrong, I'm not American) can only sneak through so much stupid legislation, ship young men off to war without a clear reason, set up prison camps which bypass the Geneva Convention and make excuses for their own soldiers torturing Iraqi prisoners before the public will say "maybe the problem isn't in Iraq".
I'll probably get modded down for this, but I've got karma to burn and I still think the point stands.
I don't think they're talking about Tom and Jerry here.
I've yet to find anything which has 100% compatability. Everything I've found depends on a plugin installed for Outlook, and the plugins tend to vary in terms of how good they are.
One thing they frequently don't offer, which Exchange/Outlook does, is realtime updating of free/busy information. It doesn't sound like much, but without it you've got a massive race condition every single time someone schedules a meeting. The first manager to get tripped up by this will have IT's guts for garters.
Actually, that's not strictly true. Scalix claims 100% compatability, complete with support for HA, clustering with Exchange and various other things, but I've looked into that and found the licensing costs for that particular edition to be comparable to Exchange itself. Additionally, while it claims to have great integration with Active Directory, it doesn't integrate with any other directory services. So you'd still need an AD domain, which right now implies at least two Windows servers and the appropriate licensing.
I am finally forced to agree that OOo is no replacement for corporate uses of MS Office. It's not OOo's problem, though; it's MSO's problem.
Except that Microsoft hold all the power, and will do until such time as some of the large-scale highly publicised OO.o migration plans come to fruition.
While it's technically Microsoft's problem, don't for one minute imagine that they consider it to be a problem.
What is stopping you from creating your own SNMP agents ? Unless you some some esoteric piece of hardware, it is fairly easy to create one.
A few things:
1. I'm the IT manager in a company which develops embedded networked devices. We've got all sorts of weird and wonderful stuff.
2. The same thing which stops me putting Net-SNMP onto a server on a non-firewalled connection. SNMPv1 offers very poor security, and ZenOSS doesn't support SNMPv3 security.
3. I trust a network probe to confirm that a service is still working properly a lot further than I trust an SNMP agent which is essentially going to be doing little more than ensuring that the process is running and listening on the port - while the SNMP agent may be technically correct in its assesment, it hasn't actually tried to get any sense out of the process.
At my current employer, I'm using Nagios (which is great as far as it goes, but needs a lot of configuration) and looking towards ZenOSS (which requires a lot less configuration because it can get everything natively through SNMP where available, but isn't anywhere near so flexible in terms of defining your own tests if SNMP isn't an option),
Ideally I'd have a solution which offered the SNMP support and out of the box functionality of something like ZenOSS, while at the same time being dead easy to extend to run scripts and check things which can't easily be checked with SNMP. I don't have a lot of time to invest in the initial setup - obviously I'm prepared to invest more time in extending where necessary, but SNMP support should be as simple as "name the host and the SNMP community".
So it will be a while before the US sees something like this:
b an_btcom_con_hub_Fusionpromo_buy
http://www.btfusionorder.bt.com/?s_intcid=con_int
Quite correct, but the effectiveness of that is proportional to the quality of the PSU.
I've seen more desktop hard drives die than servers, but generally speaking I'd expect the PSU in a server to be substantially better quality.
Why can't you merge changes back to BSD-licensed code?
Nothing. But if I merge changes into a piece of BSD-licensed code, there's nothing to stop my competitor 20 miles down the road downloading this piece of code complete with my changes, making a few minor tweaks to it, keeping those tweaks private and selling it - even though I provided a lot of the code which his product is based on.
With the GPL, he can still make tweaks to it but he's got to make those tweaks freely available to anyone he distributes code to. Many choose to simply submit these tweaks as patches upstream rather than maintain their own fork of the software - not really a lot of point in being that anal when the license explicity allows your customer to do that anyway.
What we're seeing happening now - particularly in the embedded space - is that manufacturers are taking the free stuff in Linux, tweaking the kernel and submitting changes where necessary but keeping the majority of their proprietary code logically separate in userland so they don't have to GPL it. Hence why you can have a router based on Linux which is technically open source, but the clever stuff (eg. removing the complication from configuring iptables with a web app, a means of holding firewall rules and some glue to turn these into iptables commands) remains private.
My computer allowed me to run my operating system which allowed me to download this utility which allowed me to circumvent this encryption.
Not to worry. With the advent of Trusted Computing, your next computer won't.
While I'm happy to admit that most of my original examples are now a few years out of date, let me set a few things straight:
Cheapest Network Card: £5.88
Most Expensive: about £200, but it's a 4-port gigabit NIC so not really a fair comparison.
Cheapest USB cable (0.9m): 49p (yes, pence)
Most Expensive USB cable: £146 (no, I don't understand that either)
PCs: starting at £149.99
All at ebuyer.co.uk. As regards quality of service - I think a lot of that depends on your local store. YMMV and all that.
I'm talking about people like the late lamented John Peel
Now that is very true. A legend in his own lifetime. He must have listened to loads of crap, as you say - but the difference was that was his entire livelihood, so he did have the time to do it.
the major labels won't even promote someone who's really ugly
That would really pose a problem for Mick Jagger then.
Here in the UK, there is 1 (yes one) company which pretty much has a stranglehold on about 90% of the bricks & mortar large computer stores. While smaller ones exist, I'm only talking about the large places with a reasonable selection that tend to be located in out of town industrial areas, since that's the fairest comparison with CompUSA and BestBuy.
It's the closest thing we have to Best Buy or CompUSA.
It's called PC World, and it's run by a company called the Dixons Stores Group.
Let's just say: £80 network cards (that was apparently the cheapest), £20 USB cables (again, the cheapest), £1,200 PCs. They stock cheaper ones but I've never met anyone who ever bought one - I suspect those who are savvy enough don't go there in the first place, and those who aren't are generally told "You want to send email? Well, in that case you'll need this...". And extended warranties which cost 70% of the value of the PC, yet are serviced by spotty 16 year olds who wouldn't know a PC if it dropped on their head.
Just like today, when your ISP's stock helpdesk answer is "Disable any firewalls and then try it"?
I think you'll probably find that the reliability issue with laptop hard drives has more to do with them spending most of their working life being thrown around in a laptop than any inherent problem.
2) Use MS Site License to put Windows on them all
You can't do that. The terms for all of Microsoft's site licenses state "this is an UPGRADE, and as such can only be applied to other versions of our OS. The full licensed version must still be purchased at some point for every system in your business, either preinstalled on the PC as an OEM copy or purchased as the boxed retail product".
It's a fine line, and it's an entirely artificial distinction as the CDs you get when you buy a site license aren't anything particularly special - they don't check, for instance, that you already have a valid OS to upgrade from. But breaking that rule is a very good way to have the BSA come down heavily on you, and that will almost certainly cost an awful lot more than just buying the right licenses in the first place.
(For the purposes of simplicity, I'm overlooking the fact that most of those licenses are quite obviously worded in the most awful way in order to set you up for failure).
Just sounds like some young buck made a bad decision without due dilligence and is passing the buck to Dell.
So sounds like someone without a lot of knowledge and poor problem resolution skills.
That or astroturfing FUD. IME, it's practically impossible to find a SCSI controller which can't be made to work like a charm in Linux so the OP is, IMO, either lying or incompetent.
Funny, I'm running a couple of x8xx series servers (which also use an LSI-based RAID controller) and they work just fine.
The only thing that's closed-source is the program which configures, manages and gets diagnostics out of the RAID controller. This, I accept, is a problem because without it you can't usefully monitor things like disk failure. But you can certainly load the OS without anything proprietary.
All very true, but Postgres' defaults are extremely conservative, and as a result tend to produce a system which works beautifully, but somewhat slowly.
In the Windows world, it's relatively unusual to have to do that much tweaking to a piece of software to get the most out of it.
rpm -e --force e2fsprogs-libs will do it nicely. Of course it's trivially easy to fix with a rescue CD, but apparently ESR hasn't noticed the rescue option when installing.
It would also be trivially easy to setup a list of absolutely vital, do-not-remove, no-really-we-mean-it software which cannot be removed even with rpm --force, and hack rpm to read this list every time it's invoked with the -e switch. Then if you try rpm -e --force this-vitally-important-thing, a big red hairy error comes up saying that you really really don't want to do that.
Document where this file is so if you're a real nutcase you can remove stuff from the list of "packages not to be removed under any circumstances" and it's problem solved.
Of course, if all the packages were perfect in the first place, then there would be no need for such a file. The very fact that you find yourself trying --force would serve as a warning. But I've seen packages which you do have to do that for a fair few times now.
You tried Redhat lately? I'd say they basically lost it a few years ago, shortly after spinning off fedora (and the warning signs had been there for some time before).
Mandrake and YellowDog had done a lot to obviate the most glaring deficiences in rpm with urpmi and yum, but things were still a long way from being perfect.
Now as a Linux user, I find that whenever I want to use anything RedHat based (eg. Fedora, CentOS) frustrating because it's missing fundamental packages which I need. That's OK, I can live with that and roll my own. But then dependencies are missing, and it's frequently quite difficult to find out what "cannot link lib-obscure-random-string" is fixed with which package. Then you find that you're downloading the latest version of a given tool, which was developed on the assumption that you've got the latest version of every library. Except you don't, as your OS was put together and tested 6 months ago and hasn't moved since. So it's back to compiling,
The only thing which amazes me is that ESR has either put up with this for so long or has never been seriously hurt by it.
Agreed.
I'd say ESR is taking a very reasonable approach. Software needs to work - or if it's not quite perfect, it shouldn't be so broken as to be awkward/impossible to fix without source code hacking. At most, it should be a few bits of relatively straightforward configuration away. And for desktop software (I'll make an exception for fundamentally server side stuff like Apache, because the target audience there is supposed to know what it's doing), that configuration should be click-click-click, not click, bring up terminal, spend half an hour wading through a manpage, tweak a config file, pray it worked.
In any event, it should be difficult if not impossible to fundamentally break a system as an end user unless you're putting it to some extreme abuse.
ESR has clearly decided that RedHat is unsuitable for him because it doesn't meet these requirements. Good for him. IMO, I could have told him that 5 years ago.
As others have said, it's only news because it's someone well known in the community. If it was almost anyone else, there would be a collective yawn.
Exactly. "Unsupported" is a magic word used by tech support departments so they can wash their hands of the problem.
Your OS keeps crashing? You're running an unsupported application. Go away and don't come back until you've fixed that.
Still crashing? You must be running unsupported hardware. What's the exact make and model number of every single component in your PC? You don't know? Go away and come back when you do.
Hardware vendors are just as bad:
Your hard disk appears to have failed? Sorry, you're running an unsupported operating system. Go away.
Your power supply has exploded? Sorry, we only support people who don't actually ring up requiring support.
Your power supply has caught fire, destroying your house and all your belongings? [click]
I have my doubts that this article can possibly be true. Remember that the music is only being sold because the big labels of this world - the RIAA (do they cover Canada as well?) say that it may be.
As soon as any of the music stores start selling RIAA-covered music without DRM, expect the RIAA to come down on them like a ton of lawyers.