> Gee thanks. You say the kindest things. Yeah, even I have a moment once every 100 years or so.
> Show me a successfully innovative company that does > not suffer a > degree of NIH. That's not to defend NIH, it just > comes with the > territory. No dispute there, in fact that's what I wrote. But it is a shame that it must be like that *almost* 100% of the time. We're engineers, which means we should have enough intelligence to spot superior technology and assimilate it instead of trying to outsmart it by inventing our own "hot water", or more commonly put "reinventing the wheel", sometimes suboptimally. > Don't confuse conservatism and inertia with NIH. When > Sun adopted > SVR4 packaging, just about the only alternative was > CPIO. Oh, I'm not. Conservativism is not all bad. If a solution is solid, scales well and functions well, I believe that there is no need to change that just because of the latest fashion. That is to say, I'm against following the latest trends just because they're currently fashionable. Fashion comes and goes, solid solutions stay. > And we've always cared about compatibility. And I always loved you for it. > And maybe we don't always agree that the problem has > been solved > successfully by others. Anyway, where are all these > incredibly > successful UNIX companies of which you speak? I was writing of UNIX companies that sucessfully solved certain problems, but that is not to imply the companies themselves were successful. And the companies in question are sgi and hp. sgi had some awesome engineers, and they were numero uno without any doubt, until the management ruined that company. You might recognize them in your Nvidia colleagues, since lots of them defected there (which is why Nvidia implements sgi graphics concepts like pipelines in their accelerators). Anyways, sgi and hp solved some of the problems Solaris is facing today many years ago, in very, very elegant ways, you could say - slick. Solaris would do well to assimilate those solutions, at least conceptually, instead of trying to outsmart everyone and come up with their own stuff. Being intelligent is certainly good, but being intelligent and wise at the same time is smart. > And when it comes to technologies like SVR4, SPARC, > NFS, NIS, > threads, Java, DTrace, Zones, ZFS ... I'm relieved > (even proud) that > NIH won the day over me-too mediocrity. > > Not all NIH is bad. No, I agree with you. If a solution works well and is an elegant one, then it would be unreasonable to "fix something that ain't broken". But my beef with the miniroot fiasco and System V packaging is that of technology that either didn't exist before in Solaris (miniroot) or that's lacking some serious stuff (System V packaging) - I should know, since I crank out several packages on any given day, and I often push the SVR4 packaging subsystem to his limits. > Yes, some kinds of NIH are a real shame, but please > don't think we > don't care. Sure, we're always going to get some > things wrong, but > perhaps some of what is mistaken for NIH is simply a > resourcing issue? Could be, but if you/we are strapped for resources, shouldn't that be all the more reason to look into how others already solved stuff *we* are facing? More people know more, or? > What is OpenSolaris if not (in part) a big admission > that "we can't invent it all"? Perhaps. Or perhaps it is the inherent desire for something so slick (on the whole) as OpenSolaris to dominate the computing world (;-) This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
