From: UNIX admin <[EMAIL PROTECTED]> Subject: [osol-discuss] Re: RE: And that would break... what, exactly? > Those places are > Solaris heavy, in my experience. What does that tell you?
It tells me that Solaris is a more robust, more reliable, more engineering-appropriate system that Linux is. I already knew that, which is why I use Solaris where I do, and why it makes it into the places in my environment where it is. To dismiss anyone who deploys Linux as a hacker who doesn't understand real systems, though, is short sighted. Different organizations have different needs and different capabilities.
Can one really call oneself a developer, if she/he doesn't know the underlying OS he/she is developing on/for?
I certainly don't call myself a developer, but I think this reinforces my point. They do know they underlying system they're developing for. It's just not Solaris, and that's precisely _why_ it's not Solaris.
A nice new turbo diesel may be a lot better than an > older gasoline engine, > but so help me, I don't have any idea how to change > the oil in one. Like this: you drain the oil, then unscrew the oil filter, and then screw the new one on. Ee, wait! That's *exactly* how one does it on a gasoline engine. Whaddya know?
I think this also reinforces my point, in a somewhat oblique way. I've never been under the hood in a Diesel engine. It's unfamiliar to me. If I decided to start poking around, I'd probably figure that out fairly quickly. There would also be limits to what I'd figure out without training, guidance, and more experience than I might have time to gain. There is cost associated with all of that. If I already have a developer (or a bunch of developers) who can give me the app I want on Linux, I need some pretty compelling reasons to move them to an unfamiliar system. It's likely the case that Solaris has those compelling reasons, but it's only apparent if you already know Solaris. If the familiarity barrier is lower, it makes the transition easier, reducing that cost. More importantly, by making Solaris more accessible, it's more likely that the Linux savvy admins/developers/users will or have made themselves familiar with Solaris, allowing them to discover it's benefits on their own.
Funny isn't it? I with my Solaris/UNIX knowledge can easily figure out and use any Linux system, but the same does not hold true the other way around. At least not in most cases, as we can clearly see from comments and questions here on opensolaris.org. Hmmm, now why would that be the case? Why doesn't the opposite hold true in most cases?
Well there are two likely answers. Linux people are dopes or Solaris is harder to use. I know which one I think the answer is.
Hacking something up in the PoC (Proof of Concept) phase is fine -- that's how some of the most ingenious things came to see the light of day. What's not fine, and definitely not OK, is when a hack is released as a finished product, and when quantity comes before quality. >From what I can tell, both from experience and reading the responses here, Solaris folks want quality over anything else. It's most likely one of the biggest reasons they are just that - Solaris users, and not Linux ones. There's of course another lesson to be learned from that, only if people that this concerns were willing to learn.
Again, this is all about customer expectation. There's a lot of market at the middle and low end of the spectrum. There's a lot of good people doing real business in a less sophisticated way than, say, big finance does things. What is categorically not fine for some organizations is perfectly acceptable to others. I'm not saying it's better to reduce the quality that Solaris provides. If that happened, it would eliminate the benefits that often make Solaris a better choice. I think Ian said it perfectly a couple of messages ago (and I paraphrase): don't break the functionality that the large enterprise market needs, don't abandon that market. But there is a ton of development that happens at the grass-roots level -- you point that out in your PoC reference. Capturing that end of the market so that it's viable for the work to _start_ on Solaris instead of eventually only having a chance to land on Solaris once it's mature is preferable. In any case where Solaris has compelling advantages over Linux, such a situation would be better for the developers, better for the orgs building the new ingenious systems, better for their users, and better for Solaris. Rich
_______________________________________________ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
