Slashdot Mirror
A New Look at Linux vs. Windows TCO
An anonymous reader writes "Laura DiDio, research fellow at the Yankee Group, published a column this morning in which she discusses key findings from a new survey on the total cost of ownership of Windows vs. Linux. DiDio often is written off by the Linux camp as being pro-Microsoft, but she offers excellent, neutral advice for any IT department considering a fundamental systems switch: 'If you do not know what is on your network, if you cannot at least estimate the hourly, monthly or yearly cost of downtime, if you do not know how long it takes to recover from a security outage, if you cannot answer questions about the extent of your company's license compliance, then you cannot truly evaluate whether Linux, Windows or Unix is right for your business. Chances are, if you cannot answer most or all of those questions, it does not matter what operating system you have because ignorance of the core TCO tenets means that your business is not getting the most out of its networks.' "
You know, I wonder how much of the DLL version hell in the Windows world might be a hold-over side effect of building a mentality around short 8.3 style filenames. I know the restriction isn't there anymore, but it is kinda a hack outside of the NT world (and everyone wants to maintain some sort of Win9X compatability)
Likely the reason that problem isn't as bad as in the *NIX world, is because we've never had much of a filename-length option, nor a limited extention issue, and have tacked version numbers onto library filenames for a long time.
Of course the way many Windows program installers dump crap into the WINDOWS directory is also quite bad, but an area I'm not sure *NIX is much better. (sure, more organized, but still not much better fundamentally) Of course Apple seems to do a much nicer job here, making programs quite "self-contained".