Correcting Bad Data about Open Solaris (Updated)
By ColmSmyth on Nov 30, 2004
O'Reilly's OSDir.com usually aggregates fairly well-informed content but occasionally it get's it badly wrong, like when it picked up this article on WebProNews.
The author, A.P. Lawrence, offers "SCO UNIX and Linux consulting services" which doesn't bode too well for an objective evaluation of Solaris, and sadly the article seems to reflect a certain bias.
There's a lot there that I could take issue with, but I'll focus on the highlights without dignifying the author's opinions by quoting again them here:
- Only specifically die-hard Linux pundits seem to want the "one world, one operating system" scenario; everyone else acknowledges that Linux has an assured place (until the need evolves for a really next-generation OS), and that choice is good
- Most of "Linux" is actually not Linux; it's UNIX-compatible open-source software (see this for more analysis)
- Linux has an advantage over Solaris in just one particular space - drivers, which is of primary importance only to the sundry and diverse configurations typically found on desktop-class machines, not to the embedded devices or the real servers; in all areas outside of the desktop, it's quality, reliability and performance that drive demand (and I would make the case that these qualities are also important on desktops), and drivers are less of an issue in enterprises where there is greater homogeneity of hardware configurations; in these areas and with these qualities, Solaris is a compelling choice that becomes even more compelling when Solaris becomes open-source; I also think Open Solaris will support more devices as it's community grows.
- The author seems to think that Solaris would only compete with Linux; although some pundits seem to forget, oddly enough that's not the only operating system out there. In a world of choice, interoperability and integration is important.
It seems that some folks (even some at Sun today) underestimate how important it is when a major operating system becomes open-source. Choice today drives some users towards variants of Linux (which often defaults to Red Hat). Sun today offers customers a choice of Solaris or Linux. So what would happen if there was a standard application platform that supported sufficient portability between Linux and Solaris? I think that would make for a fairly friction-free choice between Solaris and the Linux variants (and possibly also \*BSD) that supported such a platform. Java, GNOME and KDE already offer a large part of such a platform; what will tomorrow bring? I think the answer will be an open choice.
Update: I was planning on commenting on an OSNews article, but I found James Vastbinder over at Microsoft already sets the record straight. Come on guys, if a Microsoft engineer knows (and blogs) that Solaris x86 is a great OS, how come some Linux pundits can't be a bit more objective?