On Sun, Jan 13, 2008 at 06:34:42PM +0200, Nadav Har'El wrote: > I don't understand what you mean. Imagine that Google initially had some > Sun Solaris servers in their university rack. At some point, they decided > that PCs, and Linux, was more effective so they switched. What makes you > think that Google only "expanded", and never switched technology? My guess > (not founded on anything, since Google don't like to publish the technology > they use) is that if you looked at Google today, almost nothing would look > like the Google of 10 years ago. It's like that old philosophy question of, > "if you take a red cloth, and start switching threads one by one with blue > threads, you end up with a blue cloth. At what point did it become a blue > cloth?".
However Solaris and Linux are significantly close to each other that the switch was almost painless. I think you would find that once you got beyond the Kernel, the "userland" programs were very much the same when they were using Solaris as when they switched to Linux. That's one of the nice things about Linux as an operating system and IMHO a very bad thing for running "classic" hardware, is that it makes almost everything the same from a userland point of view. Except for the "word size" and "endian" differences, any computer made today that fits the minimum requirements for Linux can run it and any program that fits. If you can squeeze it into the available RAM or somehow create a swap file, you can run the same webserver on your router as on your mainframe. If that is not scalability, I don't know what is. BTW, Solaris was probably a good choice to start out at, in the late 1990's Linux was not as robust as it is now, and if you wanted something out of the box that worked well and was highly scalable, Solaris was it. To hazard a guess, if Sun had "open sourced" Solaris in 1999, or Google had asked for and obtained a source license (they were available to universities, HUJI had one), they may have gone on to enhance Solaris as opposed to Linux. To use your cloth analogy, yes it would now be blue instead of red, but it would still be a blanket and not a cushion :-) > > Yes, this is obviously true. Which means they should choose a good CTO - > not that a good CTO is useless once the initial decisions were made > (which I thought is what you meant, but if I misunderstood, sorry). I was addressing the issues with a startup hiring someone and saying, someday you will be CTO, but we are hiring you because you can do what we want done, the way we want it. If the technology decisions are made by that time, it is a rare management team that will accept another direction, unless it is fixed by outside forces. A company that makes an add-on to Internet Explorer (all the rage about 5 years ago), is not suddenly going to port their product to to the Playstation, but they might move to similar markets and embrace FireFox, Safari, or Opera. However a CTO or board of directors who knows nothing about those products might resist the change. I once (ca 1989) had the privilege of meeting Pres Eckert, the engineer behind the ENIAC and later the CTO of Univac, long after he retired and UNIVAC was on it's way out. He told me that he made far more money from real estate than from his UNIVAC stock. One reason was that they had refused to listen to him about making PC's in the early 1980's. By the time they got involved with them it was too late. It's especially bad in Israel, where they "my stick is twice as big as your stick" mentality is very common, and once a decision is made it is followed. I don't like the idea of getting political here, but it comes from people whose only large project experience is in the IDF, which has IMHO become an army run by middle management. If you read the Jerusalem Post article I mentioned, and go beyond the obvious commercial for a late stage investment fund, you see that investors are more looking for experienced entrepreneurs, and less for "young" ones. http://www.jpost.com/servlet/Satellite?cid=1198517337859&pagename=JPost%2FJPArticle%2FShowFull Geoff. -- Geoffrey S. Mendelson, Jerusalem, Israel [EMAIL PROTECTED] N3OWJ/4X1GM IL Voice: (07)-7424-1667 U.S. Voice: 1-215-821-1838 Visit my 'blog at http://geoffstechno.livejournal.com/ ================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
