Or they only had programmers who understood how to work on a Microsoft platform and this is their attempt at making it cross platform...
Lets face it, not everyone understands the platforms they are programming for to make software cross platform.
Just because people learnt to program in the school of Microsoft Development Environments surely shouldn't mean that we shouldn't support their efforts to look at other platforms and look at moving in other directions.
What are you trying to do - lock people into a Microsoft platform?
Despite my earlier comment I agree with this - Microsoft make it easy to evaluate by having long trial periods where you can get hold of the software easily and use it for a reasonable length of time to evaluate (plus they don't seem to care if you stop dealing with them and go elsewhere).
It is usually smaller suppliers or service companies who build on top of MS products (notably SQL Server, Dynamics CRM, Sharepoint though also other vendors products like Oracle Databases) and work to niche markets that unfortunately don't. I would guess their margins are tighter and that they have to fight more for the market.
And some turn around and imply you are doing something wrong by even wanting to evaluate a product...
Either way the pitfalls in pre-studies and pre-purchase evaluations are huge.
"Purchasing?" With closed source companies evaluation of a product is also an intent to buy and many will insist on a purchase in advance with a "money back" if it "fails to meet requirements". Try proving the points when a product fails to meet expectations to a company who knows it has your money but that you want it back - it's rarely free.
Evaluation Agreements... What about the process of getting the agreements on what the evaluations entails, the details on how to measure performance against the expectation, and understanding the expectations themselves? These are all things that be very difficult to prior to evaluation. Post evaluation its usually far easier to define as the software has been seen and tested and users can say, as part of the testing, the issues that they had with the software.
Holding the Baby... The analyst role in this is crucial - are you analysing the need and putting a solution in place or are you acting as a operational, deployment and maintenance service? If you are the latter and are having to contract so that you can evaluate software its an uphill struggle.
I've seen so many occasions where servicing those who have also undertaken their own analysis is a nightmare. Most of the time you have no information in the analysis on which to fail a piece of software from your evaluation process. The information is simply not there as the end user was not clear about the initial requirement. If you aren't careful you are the last person holding the baby.
In conclusion... The contractual side in closed source is a very big pain compared to open...
In all honesty, compared to the ESX server products with all the trimmings it is a little nasty to get running - and even when running you wonder where certain functionality is - until you remember its on Linux and script your own.
Eventually though, you start wondering why people pay so much more when you just keep getting more functionality that works.
As for the solution - Heartbeat and Xen both take some tweaking to get right but are fine and stable when they are right and the other parts of the solution are mostly bomb proof. We use 2 heartbeat clusters - one for storage (running Heartbeat, the iSCSI target, LVM and a remote drdb connection) with fencing. We had some old but rather functionless Fibre Channel kit sat around - hence the storage cluster, we needed to be able to split the disks down in a granular way for the VMs. The other cluster for the xen servers (Heartbeat, Xen, ocfs2 on iSCSI initiator and raw disk to the LVM partitions depending on what is required). I've had no issues recently since stablising the system (which took a little digging - mostly xen and heartbeat) and some early performance ones with ocfs2 which were annoying but they seem to be rectified in later patches.
Some of the other bits you'd normally get are a bit manual (expanding vm's, LVM snapshots, quiescing the VM's before snapping to name a few) to get right but they are largely things that you make procedures or scripts for once you've got the theory cracked.
The only shame about it for me is currently the lack of decent configuration on the network side of the solution (just a bridge) though there are ways we are looking into expanding this.
Elegant? Try using Heartbeat, OCFS2 and clustered Xen solutions (on top of the Linux iSCSI target/initiator and LVM and DRDB for snapshot and replication). Virtual machine failure with automatic restart and 300ms migration with remote site backup? Nice!
I completely know where you are on this and wish people would wake up to the situation many businesses are in regards Open Source.
Even in those Enterprises that aren't 100% MS, like the environment I work in, have a big reliance on Microsoft and there is fear that you won't get support from vendors if the network isn't 100% Microsoft. I've had people say they won't support their proprietary software on our system because there is a Novell Netware server or a Linux server doing some file serving on it. Even when the software doesn't even touch those systems except for file access you can get this kind of reaction. This same argument is used against Linux and other Open Source software sitting along side it time and time again unless compatibility is proven time and time again.
There is also the old saying, "Noone got fired for buying IBM" - except that its now Microsoft not IBM. Try to stop your management from feeling safe in their hands - to them Vista is still the best thing since sliced bread because that is whom the systems are marketed to. You find lots of execs and senior managers sat in Microsoft's Reading (UK) conference centres being sold the glory of Microsoft systems...
But back to the article, there is really nothing stopping Microsoft making marketable products from Open Source. It is a paradigm shift for them but instead, and I think you are right that the article is pointing to new tactics. Perhaps a competitive "open" or subscription license for developers?
From a design and architecture perspective I'd definitely agree. The only ways I wouldn't would be in that I don't think it had the same impact as the change from DOS/Win3.1 to Win95. Keeping compatibility whilst changing the system like that was a big deal - without that step 2000 would have never taken off (and without ME would probably have never survived the competition with what was a faster, slimmer, though less secure OS).
Actually a lot of OS2 had been developed with similar features to W95 before W95 was released: similar memory model, the backwards DOS compatibility. Microsoft developed that with IBM - its where the 3.1 interface came from before they parted company!
Microsoft did the same thing they always try to do - copy the good ideas and expand them. By this I mean that they did not copy the product but take the idea and expand it in their own (proprietary) direction.
You're right about the 3rd party requirements though - they also leave deliberate holes to be filled with other products. It meant that you had to buy the extra if you wanted it to work well (Anti-virus, Anti-spam, Defrag, Partitioning Tools, Performance tuning utilities) and that's not to mention the real applications - no wonder they thought it was good, they must have been as happy as pigs in mud.
Servers in general seem to conform to this. On the Dell Poweredge range they seem to support Windows, Red Hat and SLES - including having the Dell Server Assistant and Openmanage software suites support them (and run well).
This is great for us - it allows us to monitor the and control the hardware, even aspects not available to the OS (like the Raid config, power and fan sensors) and actively predict failure before it happens.
Unless there is a specific piece of hardware that you need or want that is in a Windows box and not a Linux box, I really don't see the need to buy Windows when you want Linux if there are Linux machines available. One things that is missing from this argument is...
Application Support
If he needs any specialist application its likely to primarily be Windows based (especially if its a bespoke product). Despite all the fine work that the WINE team does and the application developers of the many applications on Linux do, people still require Windows only applications on a regular basis and for them only too often to live on a Windows platform for best compatibility (and support).
Unfortunately, that is the reason I'd see as being the most compelling for buying machines licensed for Windows. At least it gives a choice, if any applications do require the Redmond curse then people can run Windows virtualised on top of Linux for any software legally (Vmware server, Xen, etc).
The issue with patents is not necessarily the idea behind patents. Patents come down to the concept, which someone has already stated, of "it being my hill" and protecting that hill which in theory help the small guy from having *new* ideas from being exploited.
I'd guess that this was something many people would agree with - surely someone who advances the knowledge of the human race should be entitled to some reward, The inventions shouldn't be swiftly gobbled up by someone taking credit for inventions that are not theirs.
No, the issue is based on the definition of that hill and the length of time people are allowed to keep the hill before it reverts to public ownership. If a hill is left alone, whether untouched or not and someone else occupies it, well, "finders keepers". If you desert your ideas you deserve them to be used by others.
In support of this we could look at common inventions - do you want to have manufacturer's to still pay licenses for light bulbs to the families of the many inventors whom history books have claimed were the creators? What are they doing to deserve it? All the technology is understood, some electricity and a filament will build it, just what is so deserving for them to reap reward from it? However, when the inventions/discoveries made a difference to the sum of knowledge there is reason to reward those who made the discovery.
I'd guess the point I'm trying to make is that, in reality, nobody owns knowledge or information - its simply out there to be discovered. Person X and person Y can discover the knowledge at the same time in completely different ways, however, there is usually something to be said in being first...
This runs into the idea of Intellectual Property. Intellectual property? I don't see how it exists - either its knowledge you know or its knowledge you don't. Lets face it, IP is a legal term and is a pure fiction to give value to someone's secrets (hidden knowledge) and hence prevent "stealing". Its a misnomer - you can't own a thought or stop others from thinking that thought or plant your tree on that thought. You can however, patent that thought and make money from taking the credit for that "discovery" - if you know it and other people would want it then it should be patented so that you can get the rewards. You are expanding the sum of knowledge and should be rewarded...
This would make knowledge open and more useful to mankind in general.
But, yes, I also think that the connection between Novell and SuSE contributes to SuSE's problems Novell have got a bad press because of the Microsoft angle and that won't help SuSe in any way (lets face it - SuSe the organisation is Novell now - mud striking Novell hits SuSE).
However, without Novell's deals with Microsoft, I'm sorry to say, Linux would have been out of the door where I'm based and Novell with it - not by choice of technical staff but by "Strategic" decisions based on Microsoft marketing and statements regarding lack of support and antagonism towards the Open Source communities. This shows what a monopoly and power Microsoft have when none-technical managers will not use perfectly good solutions because of Microsoft's statements - and we are an educational establishment.
Why is this? I personally believe that, although people may view Novell as a pawn to Microsoft, it is in fact what Senior Managers in large organisations want. They know they cannot get by without Microsoft and they'd be hamstrung if they had to push out an app, a standard boxed product, in many cases without it - why bother with Wine if the software can already deployed on Windows, very quickly to many machines at once and in a supportable manner. The "compatibility" and agreements give Linux a chance in the enterprise it doesn't have in many other situations.
What compatibility aspects you may say? But because of the mutual support agreements regarding compatibility aspects of Windows I can put Linux clusters of Xen virtualised servers around and not pay out to EMC for VMware or buy proprietary clustering solutions. I could not have done this without a Microsoft statement of compatibility as it would have been seen as an unsupportable risk. I can provide LAMP solutions instead of Sharepoint. I can do things that keep Linux and expand it on our servers and Open Source into our enterprise.
Much as I may like Ubuntu on my machines at home I simply wouldn't have been able to get them on the servers at work. It wouldn't have been diligent...
I'm not so sure. They can only lose share to quite capable (and free) alternatives. Its not the empty marketplace it was 5 years ago.
If Microsoft's options are:
Plan A) Lose market share and goodwill by making things as awkward as possible, making a hash of the implementation or making their products look poor.
Plan B) Lose less market share by having a crack at making the ODF implementation they can one top of the base of their current line.
Its probably a good bet that someone not interested in playing law suit marketing can see the benefit of having the best, most extended (with proprietary extensions) ODF format they can.
Slashdot Mirror
Just seen this review:
opensuse 11.0 review at downloadsquad
Tis a shame tho' - ZFS and dtrace are nice...
Or they only had programmers who understood how to work on a Microsoft platform and this is their attempt at making it cross platform... Lets face it, not everyone understands the platforms they are programming for to make software cross platform.
Just because people learnt to program in the school of Microsoft Development Environments surely shouldn't mean that we shouldn't support their efforts to look at other platforms and look at moving in other directions. What are you trying to do - lock people into a Microsoft platform?
I'd mod you up for talking sense but your too much of a coward...
Apple made a similar deal? Last I heard they were part owned by MS but my info is probably a while out of date...
Despite my earlier comment I agree with this - Microsoft make it easy to evaluate by having long trial periods where you can get hold of the software easily and use it for a reasonable length of time to evaluate (plus they don't seem to care if you stop dealing with them and go elsewhere). It is usually smaller suppliers or service companies who build on top of MS products (notably SQL Server, Dynamics CRM, Sharepoint though also other vendors products like Oracle Databases) and work to niche markets that unfortunately don't. I would guess their margins are tighter and that they have to fight more for the market.
Guess you missed the point. They can just rebrand it like they did with OS X.
Already out and already old
Mac,PC,Linux
2nd ad
And another
Is top of the range something most can afford...?
And some turn around and imply you are doing something wrong by even wanting to evaluate a product...
Either way the pitfalls in pre-studies and pre-purchase evaluations are huge.
"Purchasing?"
With closed source companies evaluation of a product is also an intent to buy and many will insist on a purchase in advance with a "money back" if it "fails to meet requirements". Try proving the points when a product fails to meet expectations to a company who knows it has your money but that you want it back - it's rarely free.
Evaluation Agreements...
What about the process of getting the agreements on what the evaluations entails, the details on how to measure performance against the expectation, and understanding the expectations themselves? These are all things that be very difficult to prior to evaluation. Post evaluation its usually far easier to define as the software has been seen and tested and users can say, as part of the testing, the issues that they had with the software.
Holding the Baby...
The analyst role in this is crucial - are you analysing the need and putting a solution in place or are you acting as a operational, deployment and maintenance service? If you are the latter and are having to contract so that you can evaluate software its an uphill struggle.
I've seen so many occasions where servicing those who have also undertaken their own analysis is a nightmare. Most of the time you have no information in the analysis on which to fail a piece of software from your evaluation process. The information is simply not there as the end user was not clear about the initial requirement. If you aren't careful you are the last person holding the baby.
In conclusion...
The contractual side in closed source is a very big pain compared to open...
In all honesty, compared to the ESX server products with all the trimmings it is a little nasty to get running - and even when running you wonder where certain functionality is - until you remember its on Linux and script your own.
Eventually though, you start wondering why people pay so much more when you just keep getting more functionality that works.
As for the solution - Heartbeat and Xen both take some tweaking to get right but are fine and stable when they are right and the other parts of the solution are mostly bomb proof. We use 2 heartbeat clusters - one for storage (running Heartbeat, the iSCSI target, LVM and a remote drdb connection) with fencing. We had some old but rather functionless Fibre Channel kit sat around - hence the storage cluster, we needed to be able to split the disks down in a granular way for the VMs. The other cluster for the xen servers (Heartbeat, Xen, ocfs2 on iSCSI initiator and raw disk to the LVM partitions depending on what is required). I've had no issues recently since stablising the system (which took a little digging - mostly xen and heartbeat) and some early performance ones with ocfs2 which were annoying but they seem to be rectified in later patches.
Some of the other bits you'd normally get are a bit manual (expanding vm's, LVM snapshots, quiescing the VM's before snapping to name a few) to get right but they are largely things that you make procedures or scripts for once you've got the theory cracked.
The only shame about it for me is currently the lack of decent configuration on the network side of the solution (just a bridge) though there are ways we are looking into expanding this.
Elegant? Try using Heartbeat, OCFS2 and clustered Xen solutions (on top of the Linux iSCSI target/initiator and LVM and DRDB for snapshot and replication). Virtual machine failure with automatic restart and 300ms migration with remote site backup? Nice!
Novell do a really good implementation guide
here
I completely know where you are on this and wish people would wake up to the situation many businesses are in regards Open Source.
Even in those Enterprises that aren't 100% MS, like the environment I work in, have a big reliance on Microsoft and there is fear that you won't get support from vendors if the network isn't 100% Microsoft. I've had people say they won't support their proprietary software on our system because there is a Novell Netware server or a Linux server doing some file serving on it. Even when the software doesn't even touch those systems except for file access you can get this kind of reaction. This same argument is used against Linux and other Open Source software sitting along side it time and time again unless compatibility is proven time and time again.
There is also the old saying, "Noone got fired for buying IBM" - except that its now Microsoft not IBM. Try to stop your management from feeling safe in their hands - to them Vista is still the best thing since sliced bread because that is whom the systems are marketed to. You find lots of execs and senior managers sat in Microsoft's Reading (UK) conference centres being sold the glory of Microsoft systems...
But back to the article, there is really nothing stopping Microsoft making marketable products from Open Source. It is a paradigm shift for them but instead, and I think you are right that the article is pointing to new tactics. Perhaps a competitive "open" or subscription license for developers?
640Kb? You'll never need more than 640Kb!!!!
I think it was introduced in the 286.
From a design and architecture perspective I'd definitely agree. The only ways I wouldn't would be in that I don't think it had the same impact as the change from DOS/Win3.1 to Win95. Keeping compatibility whilst changing the system like that was a big deal - without that step 2000 would have never taken off (and without ME would probably have never survived the competition with what was a faster, slimmer, though less secure OS).
Actually a lot of OS2 had been developed with similar features to W95 before W95 was released: similar memory model, the backwards DOS compatibility. Microsoft developed that with IBM - its where the 3.1 interface came from before they parted company!
Microsoft did the same thing they always try to do - copy the good ideas and expand them. By this I mean that they did not copy the product but take the idea and expand it in their own (proprietary) direction.
You're right about the 3rd party requirements though - they also leave deliberate holes to be filled with other products. It meant that you had to buy the extra if you wanted it to work well (Anti-virus, Anti-spam, Defrag, Partitioning Tools, Performance tuning utilities) and that's not to mention the real applications - no wonder they thought it was good, they must have been as happy as pigs in mud.
Servers in general seem to conform to this. On the Dell Poweredge range they seem to support Windows, Red Hat and SLES - including having the Dell Server Assistant and Openmanage software suites support them (and run well).
This is great for us - it allows us to monitor the and control the hardware, even aspects not available to the OS (like the Raid config, power and fan sensors) and actively predict failure before it happens.
Application Support
If he needs any specialist application its likely to primarily be Windows based (especially if its a bespoke product). Despite all the fine work that the WINE team does and the application developers of the many applications on Linux do, people still require Windows only applications on a regular basis and for them only too often to live on a Windows platform for best compatibility (and support).
Unfortunately, that is the reason I'd see as being the most compelling for buying machines licensed for Windows. At least it gives a choice, if any applications do require the Redmond curse then people can run Windows virtualised on top of Linux for any software legally (Vmware server, Xen, etc).
That is just sooooo wrong......
The issue with patents is not necessarily the idea behind patents. Patents come down to the concept, which someone has already stated, of "it being my hill" and protecting that hill which in theory help the small guy from having *new* ideas from being exploited.
I'd guess that this was something many people would agree with - surely someone who advances the knowledge of the human race should be entitled to some reward, The inventions shouldn't be swiftly gobbled up by someone taking credit for inventions that are not theirs.
No, the issue is based on the definition of that hill and the length of time people are allowed to keep the hill before it reverts to public ownership. If a hill is left alone, whether untouched or not and someone else occupies it, well, "finders keepers". If you desert your ideas you deserve them to be used by others.
In support of this we could look at common inventions - do you want to have manufacturer's to still pay licenses for light bulbs to the families of the many inventors whom history books have claimed were the creators? What are they doing to deserve it? All the technology is understood, some electricity and a filament will build it, just what is so deserving for them to reap reward from it? However, when the inventions/discoveries made a difference to the sum of knowledge there is reason to reward those who made the discovery.
I'd guess the point I'm trying to make is that, in reality, nobody owns knowledge or information - its simply out there to be discovered. Person X and person Y can discover the knowledge at the same time in completely different ways, however, there is usually something to be said in being first...
This runs into the idea of Intellectual Property. Intellectual property? I don't see how it exists - either its knowledge you know or its knowledge you don't. Lets face it, IP is a legal term and is a pure fiction to give value to someone's secrets (hidden knowledge) and hence prevent "stealing". Its a misnomer - you can't own a thought or stop others from thinking that thought or plant your tree on that thought. You can however, patent that thought and make money from taking the credit for that "discovery" - if you know it and other people would want it then it should be patented so that you can get the rewards. You are expanding the sum of knowledge and should be rewarded...
This would make knowledge open and more useful to mankind in general.
Seconded! The bigger the bang the better... It means greater propulsion when it does get put right.
In that environment I'd probably agree...
However, without Novell's deals with Microsoft, I'm sorry to say, Linux would have been out of the door where I'm based and Novell with it - not by choice of technical staff but by "Strategic" decisions based on Microsoft marketing and statements regarding lack of support and antagonism towards the Open Source communities. This shows what a monopoly and power Microsoft have when none-technical managers will not use perfectly good solutions because of Microsoft's statements - and we are an educational establishment.
Why is this? I personally believe that, although people may view Novell as a pawn to Microsoft, it is in fact what Senior Managers in large organisations want. They know they cannot get by without Microsoft and they'd be hamstrung if they had to push out an app, a standard boxed product, in many cases without it - why bother with Wine if the software can already deployed on Windows, very quickly to many machines at once and in a supportable manner. The "compatibility" and agreements give Linux a chance in the enterprise it doesn't have in many other situations.
What compatibility aspects you may say? But because of the mutual support agreements regarding compatibility aspects of Windows I can put Linux clusters of Xen virtualised servers around and not pay out to EMC for VMware or buy proprietary clustering solutions. I could not have done this without a Microsoft statement of compatibility as it would have been seen as an unsupportable risk. I can provide LAMP solutions instead of Sharepoint. I can do things that keep Linux and expand it on our servers and Open Source into our enterprise.
Much as I may like Ubuntu on my machines at home I simply wouldn't have been able to get them on the servers at work. It wouldn't have been diligent...
I'm not so sure. They can only lose share to quite capable (and free) alternatives. Its not the empty marketplace it was 5 years ago.
If Microsoft's options are:
Plan A) Lose market share and goodwill by making things as awkward as possible, making a hash of the implementation or making their products look poor.
Plan B) Lose less market share by having a crack at making the ODF implementation they can one top of the base of their current line.
Its probably a good bet that someone not interested in playing law suit marketing can see the benefit of having the best, most extended (with proprietary extensions) ODF format they can.