Slashdot Mirror
OSDL CEO: Microsoft Has to Accept Linux
PenguinCandidate writes "The OSDL's Stuart Cohen has been in the news lately following a clandestine meeting with Microsoft about a dubiously independent TCO study; a study that has since been rejected by the OSDL. The idea of an independent Windows/Linux TCO comparison may be dead, but did Cohen have an additional card up his sleeve?
In this interview, Cohen states that while he "awaits the reply from MS's Martin Taylor on the results of his internal investigation" into how an off-the-record meeting became public, he will continue to promote his belief that MS will eventually have to accept Linux as customer demand increases."
If I change the oil in my car myself, every ~3000 miles, it costs me about $20. If I have to take it to Quicky Lube it's about $32 (plus they try to sell me a bunch of useless stuff). Obviously TCO of the same car varies depending on the expertise and willingness of the customer to crawl underneath and get dirty.
Similarly, if a customer has to hire someone to edit his inittab then it's probably going to cost more than a Windows jockey clicking on services attributes. Dunno, there's just too many indefinite variable to compare complex systems.
try { do() || do_not(); } catch (JediException err) { yoda(err); }
Recognition vs. Acceptance vs. Surrender
MS has recognized Linux's place in the marketplace. They know of Linux's server market growth and its desktop push. Ms recognizes the choices in Open Source software (OO.o vs. MS Office, for example).
MS has not accepted this. Probably will not either. Thus the TCO fudging and other FUDs. If Microsoft accepts the qualities of Linux (and OSS), they wouldn't FUD it. They'd acknowledge its benefits and make THEIR offerings better.
So once you make the distinction between recognition and acceptance, the statement pretty much pans out ok.
MS accepts that it must compete against Linux, but I think that Mr Cohen is trying to point out some of the dumber MS practices, in regards to interoperability.
In my experience alot of this is related to how MS wants to integrate it's enterprise level products into the OS. The two biggest examples I can think of are:
- SQLServer
- .NET
Basically they both run as services (IIS, I think) on Windows OSs, making them inherently incompatible with Linux or any other OS. It's all part and parcel to this Microsoft mentality that to make the most money, we need to be an end-to-end solution; for everything.So what is my point? Well, if MS was really about making the best product you could run it on a multitude of OSs. Because if SQLServer and the .NET (web apps) were really that good they could be more OS agnostic. The alternatives, Oracle, J2EE, PHP, etc run almost anywhere. It would also be nice to see Active Directory provide full LDAP support.
And it isn't even the end-to-end solution that bothers me. It is also the lock out of everyone else (but, I guess Microsoft can always say, "Look how well we play with ourselves" ;-) ). This also seems to be half of what causes all of the OS security problems and release delays. Instead of having each app provide its own services (to some extent), the OS comes with bits and pieces for other MS apps. Some of these bits don't seem to get used much, but everyone gets them. This all adds to the complexity of the OS. While Bill and Balmer spout that it makes "everything easier to do", I disagree. I would rather set up a cluster of app servers for J2EE than attempt that nonsense for .NET using the built in configuration options (from the control panel).
I think it is possible that Vista/Longhorn will not be that sucessful. Then MS will have to make SQLServer and .NET web apps run on something else... like everything else in the world. That is just conjecture (or wishful thinking, perhaps), but that will be the only way MS can hold its ground in the long run (at least in the business IT world). Ceterus Peribus.