> > So given all that, it's up to you to decide whether > OpenSolaris is suitable > > for you. The best approach with anything is > usually in this order: > > * first, make a list of everything you know or can > anticipate you might want > > to do, and put it in priority order > > * second, find applications that do that, keeping > in mind that there may > > be multiple applications for any given task, not > all of which run on every > > platform > > * _last_, pick the OS that will support the most of > the top priority applications > > You know, I started giving advice pretty much > identical to that nearly > 30 years ago. Seeing it again, it hit me that - well, > I didn't do that > any more. Thinking about it, I realized that that's > because it's no > longer correct. [...] > So you want to get your list of applications more > like so: > > *) List the people you're going to be communicating > with, and what > you're going to be communicating with them. > Talk to them, and find out what applications they for > what you're > going to send back and forth, and what other > applications can read > those things. Find out if they feel willing and > able to help you > should you have problems, and who they talk to when > they need > professional help. > Talk to the pros about the things you're going to be > doing, and see > what applications they recommend. > f course, for most of the last couple of decades, > what falls out is > almost always Windows. I'm happy with that - I've > always used Unix, so > I can tell them "... but I don't use Windows, so > don't call me for > help."
That's just a specific expansion of what you know or anticipate wanting to do, but one that depends also on those with whom you want to collaborate or exchange information. Indeed, MS (and perhaps somewhat Apple and others too, to a lesser degree due to their lesser market share or scope) have long played the game of trying to make the combo of their OS plus their apps (or even just apps alone sometimes) be never less that "first among equals", if not for any other value, then by providing the path of least resistance for exchange of data. Standards, provided they're truly open, can help a lot, but the big guys have to be dragged kicking and screaming to the table, because they've gotten too darn complacent about competing on lock-in rather than more positive forms of value. > These days, virtualization technology is changing > that answer again, > but I haven't figured out exactly how yet. I think > Sun sees this, and > hope they make sure VirtualBox make Solaris a viable > platform in a > world where most of your applications run on virtual > systems. I think Solaris helps itself both as host and as guest; the former as a scalable stable platform, the latter as something for developers and students to use when a full-size development system paralleling a production system isn't needed. The question in my mind is whether Solaris catches on as a guest OS for production use (probably under xVM or VMware rather than VirtualBox, the latter being mostly for local desktop use). On the one hand, that increases the use of Solaris for production use, and penetration at the low end. On the other, as long as it's just consolidation, it's parasitic, although better perhaps than losing ground to e.g. Linux doing the same thing. And maybe as newer generations of hardware give more MIPS and bandwidth per unit cost, consolidation is a perfectly reasonable thing to do to keep relatively static apps running on maintainable hardware. So all the roles play their part, and even the less immediately profitable ones may lead to repeat business later on. But the price pressure is intense; and anything other than commodity hardware (current x86 on current bus using current storage, give or take RAS features) for anything that doesn't require very large scale SMP (or sort of SMP, like NUMA) will IMO always be struggling. For those, it's single thread (where PowerPC is presumably well ahead) performance vs multi thread performance, possibly taking into account power and cooling. So I wonder about the survival of SPARC, which is why, although I don't like x86 much, I'm glad Solaris on x86 is no longer abandoned or 2nd class (give or take large system features that simply don't apply there). (The one good point of x86 with virtualization is that there's the widest range of guest OSs for that architecture.) -- This message posted from opensolaris.org
