Ultra large installations of anything will not be immediately replaced. Small and Midsize organizations are where the replacement cycle starts, as well as in new divisions of existing entities. What we are seeing in these cases is a a large migration away from $soft - not 100% but not 10% either. This is exactly the formula seen in at least two other product/vendor cycles I have seen in the computer industry and a big difference here is the cost structure, which is free, and thus that argument does not need to be surmounted. Also, your argument - same as $soft - on support really is not true: y ou can take an existing group and tune it towards Linux. I have done this and spent very very little in the migration. Existing techs easily made the transition. Ourr costs are dramatically lower after the transistion - not higher. Its just fear that people are using for these argument.s FYI: The only way the transition works is if people are given no choice - actually this is the only way it works and the person at the top must lay down the law. I gave an order and guess what ? Nobody left over a switch in technology since their needs were met. Nobody in their right mind would quit a 300K job because they have to learn a new equivalent tool. From my perspective if they did then i probably did not really lose anyone i wanted or needed to keep.
I have refused a job at 200K$ fixed salary just because I had to use windows. You should have seen the face of the guys when I told them the reason why I declined their offer.
I still can remember the muted answer I got from candidates for a job at my previous company when I told them I was unable to open their ".doc" document and asked them to send me a CV in a more standard format like postscript... It's not worth mentioning this was a very good first filter for the candidates.
My recollection is that on the on old Crays this was more important than on other parallel architectures, and at the end of the day if your problem can run in parallel across several nodes on an appropriate architecture then the inherent differences in the code generation become small issues. C is easier and faster - subjective I know - than Fortran, especially for large projects with many modules/functions. At least it sure seemed that I could get more done .... but not if the libraies were already written in fortran.... These days most things have been ported ...
Yes. I am sure that people do this .... But I will bet it wansn't a job you wanted. If you really wanted it you would have taken it windows or linux - no difference.
May I play devil's advocate too? Yes, pretty, pretty please.... I can remember the very, very same argument many years ago. It was Lotus and WordPerfect (what? who? where are they now?) versus this little upstart with Word and Excel (which products were just a joke in comparison). Windows? What's that? Who needs that? (DOS works fine.) A mouse? What do you need a mouse for? No, all our staff is trained in WordPerfect. And trying to change those who were using Lotus...Still can hear the arguments.... And there was this other company, Netware, dominating the networks in the companies. You could not get a job without knowing Netware. They had how much of the networks, how much? 98%? They are called Novell these days and they too tried to "monopolize" the markets. No, no, we have too much vested in Netware, we cannot change. Same arguments, different times.... Change is inevitable, those who resist change will go under. That is a "natural" law. Maria
A very good book on this topic is "In search of stupidity" http://www.amazon.com/gp/product/15...7?v=glance&n=283155&n=507846&s=books&v=glance The typical reason for all these high-tech disasters is pretty simple : failure to provide a new product without bug and a will to sell a 'professional' and a 'personal' edition of the same software. Taken from one comment on amazon : Long before the dot-bomb collapse around 2000, companies in the PC world had been shooting themselves in the foot, making (and repeating) insanely bad decisions, and doing everything they could to drive themselves into the ground. Many succeeded in killing themselves off, others (like IBM and Apple) did not. The recurring themes sound simply ridiculous, unless you live in this high-tech world. They they sound ridiculously familiar. They include: * Expensive acquisitions of companies with nothing to offer, * Demolition ("rewriting") of bread-and-butter products, * Selling two, three, or more products that all do the same thing, * Annoying and ignoring the customers until they all wander away, and * Whatever it was, doing it again and again.
Yes, I know. I had a brief stint at Computer Associates (CA) and that company really was into this. It looked good on the books and you could put money on it that the founder would do a play on the sharemarket because of the acquisition. Pretty smart way of milking a company without being held accountable for it. Maria
I guess Micro$oft's death has been postponed for another quarter. http://biz.yahoo.com/ap/051027/earns_microsoft.html?.v=4 Microsoft First-Quarter Earnings Up 24 Percent on Growth in Personal Computer and Server Market SEATTLE (AP) -- Microsoft Corp. said Thursday that earnings for its fiscal first quarter rose 24 percent, helped by strong growth in the personal computer and server market. For the quarter ended Sept. 30, the Redmond-based software maker reported earnings of $3.14 billion, or 29 cents per share, up from $2.53 billion, or 23 cents per share, in the same period last year.