Quoting Alexander Maryanovsky, from the post of Tue, 03 Dec: > > >VMware's product is simply a little too complex and too mission critical > >to be developped without intensive and costly QA, and get things done at > >a speed the big clients expect. > > Hmm, isn't that exactly what many people (ESR comes to mind) thought about > operating systems until recently? > Almost always, the answer to the question "why is there no good (open > source)/free alternative?" is that there aren't enough talented people who > need it, and so they haven't implemented it yet. Perhaps VMWare does > everything they want, or perhaps they simply don't need something like > VMWare, because Linux is enough for them.
first, thanks for overquating. for various reasons my mutt-through SSH session is a bit slow today, so it's been a lot of fun. Second thanks for ignoring my headers and sending me a CC:. to the point: no, Linux is not good enough. fact is it needs patches that IBM, Red Hat and SuSE (to name a few) are paying extra in time and other resources to please thir clients. you don't see many big companies trusting their servers to Gentoo or even Debian (very sadly). mission critical requirements often mean testing and certification, which often introduces many problems to the clients and the developpers. consider the problem I am faced with this week at my place of work - we are building an appliance that runs Oracle on Linux. we are supposed to run it on a kernel+libs environment that were approved by Oracle, but at the same time use a storage solution, which was certified by Hitachi/EMC/NetApp/whoever on completely different kernels. What's a man to do? Linux development (as well as the surrounding libs) is non centralistic, and the large number of lib versions that work there in a combo doesn't help either. so which kernel version, and lib version and distribution to certify? and then for the client, which one to use? SuSE are better at stable file systems, but Red Hat are better with large memory machines (just a random example). many products have forks, sub projects and such. when they grow to the size of the kernel, or wine or XFree or Apache, you end up ith 3rd party source trees and such, and that gives clients a headache. IBM are forced to maintain their own fork of Apache for websphere the same way that Red Hat are keeping their own fork of the kernel. the bottom line is that a project as big and trusted as VMWare cannot be usefull unless produced centralisticly, and probably can't deliver on time unless backed up by solid investments and income, which means (at the moment) per-seat license at a healthy price, a license which does not allow open-sourceness. and all that is without starting a talk about the problems of licensing Phoenix-BIOS and other components they don't own into the open. so like I said, it's not that noone has ever tried - try bochs or plex86 and see what happens when the same effort is attempted in the OSS world, insisting on a license that is not compatable with the non-free components of VMWare means a lot of extra reverse-engineering efforts and lands you at the end with a machine that is not compatible with the clients' requirements of industry standard certified (simulated) hardware. it may be good enough for hackers, but not for the guest OS in the box or the client who wants to run a win2K advanced server with MSSQL on it. I don't happen to have a need for that, so I'm not paying for VMWare at the moment. conclusion - OSS is the way, but it hasn't proven economical yet, nor able to give the best technical solution. the few projects that do succeed are too few to prove otherwise and rely on exceptional individuals with unlikely free time or generous employers. It's fine to preach for the promotion of open source, but it's economicly impossible (for now) to have it come instead of closed source, because at the end of the day it's closed source and services that put food on the table of the hungry OSS programmer. the day this changes will be a happy day for all of us indeed. At the moment, you can help promote OSS, but not at the expense of closed source. the market SHOULD wake up and support lex86 with knowledge and code, but for now they settle for paying VMWare. give them time, they are only recently starting to switch big systems from MainFrames and Minis to linuces.. Para Para. > > > Alexander Maryanovsky. > > At 13:13 03.12.2002 +0200, Ira Abramov wrote: > >Quoting Boris Gorelik, from the post of Tue, 03 Dec: > >> > >> Just thinking: How come none asks for why isn't VMware an open source > >program > > > >1. license for the industry standard BIOS and firmware of the various > >PCI "cards" etc. > > > >2. tens of thousands of man hours in development and testing, QA and > >certifications. > > > >if you want the open alternative, look at bochs. dare to compare > >features and performance? > > > >it's nice to be a zealot, but the truth is (and the current economical > >truth will show) the world is not ready for complete Open Source > >economy. the vast majority of OSS developers have to make do getting > >payed on doing support for their work or write closed source as well. > >think of it as poets - most of them have to keep day jobs since only a > >small handfull of them can actually make a living on revenues on their > >work... > >VMware's product is simply a little too complex and too mission critical > >to be developped without intensive and costly QA, and get things done at > >a speed the big clients expect. > > > >I don't see that as bad, and if I had the need for this product I would > >have gladly payed the price. in the big picture it encourages the use of > >linux in a way, so that can't be bad either. the big buzzword in their > >business now is "server consolidation" and everyone wants a linux server > >to run IIS or other crazy combos, and this is where their big money is. > >I tip my hat to any company I saw on the show floor at Linuxworld 1999 > >that is still keeping its head above the water, and they are definitely > >one of the brightest examples. > > > >so Boris Darling, OSS and Free Software are truths. they are solid > >truths, important truths, but they are not absolute truths. the moment > >you get off the ideals tree and see how you can promote your ideals > >realisticly, we'll be able to subvert the "Closed Source economy" better > >:-) > > > >Happy Hanukkah, > >Ira. > > > >-- > >Stay tuned for the next installment of > >Ira Abramov > > > >http://ira.abramov.org/email/ This post is encrypted twice with ROT-13. > >Documenting or attempting to crack this encryption is illegal. > > > ================================================================= > 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] > -- Dr. Feelgood Ira Abramov http://ira.abramov.org/email/ This post is encrypted twice with ROT-13. Documenting or attempting to crack this encryption is illegal.
msg23706/pgp00000.pgp
Description: PGP signature
