Subject: Re: e450 as a modern server
To: Raymond Meyer <raymond.meyer@rambler.ru>
From: Dave McGuire <mcguire@neurotica.com>
List: port-sparc64
Date: 10/30/2006 23:06:27
On Oct 30, 2006, at 9:52 PM, Raymond Meyer wrote:
>> because i really would have rather ran netbsd than gentoo linux on it
>> (i try to avoid solaris if i can help it)
>
> And why would that be?? Major governments, commercial and academic
> organisations run Solaris, it's good enough for them, why wouldn't  
> it be good
> enough for you? Sure some aspects of Solaris could be better, e.g. the
> installation is slow and a real pain, but once it's up and running  
> it's rock
> solid.

   Solaris installation is indeed a real pain...but the stability is  
such that you only do it when you're building a new machine.  And  
that's not an exaggeration.

> In its current state, NetBSD is far behind Solaris on many important
> features. I like NetBSD and would like to see it improve at a much  
> faster pace,
> however I think there is a lack of interest in NetBSD on SPARC  
> hardware.

   I know plenty of people who would love to run NetBSD on their  
UltraSPARC machines, myself included, but it won't replace Solaris  
for me anytime soon for a variety of reasons.

   As far as SPARC (not UltraSPARC) hardware...I personally have  
overseen the development of a great many SPARC machines (we're  
talking hundreds) running NetBSD.  I'd not call that a lack of interest.

> The thing that motivates people a lot is money. Sun employ many  
> skilled
> software engineers who work on Solaris full time. If you're a software
> developer and work on a free software project in your spare time,  
> you can't
> expect serious results.

   ...and here's where I really stop agreeing.  That's a pretty  
sweeping statement, and fortunately for those of us who depend on  
good noncommercial software, it's dead wrong.

   The whole "employing many skilled programmers full time" argument  
breaks down when you factor in clock-watching and motivations for  
work.  This is the same old "free software is good"/"free software is  
bad" argument.

   The only place commercial software still has a foothold is  
operating systems and some vertical market applications...and the  
latter is evaporating more and more every day.  When people care  
enough about working on something to spend their free time doing it  
for free, that basically *guarantees* you better-quality software.   
And some (many?) people who work on free software in their free time  
aren't exactly amateurs, either.

> I wouldn't mind paying for NetBSD that has a rock
> solid, fine-grained multithreaded kernel, that takes advantage of  
> multicore
> CPUs like UltraSPARC T1 and has support for the latest SPARC  
> hardware. The
> problem is it doesn't exists in a free or commercial form. The same  
> goes for
> other software, i.e. desktop applications, multimedia, graphics,  
> etc. I
> wouldn't mind paying for good quality software, but there is no  
> such software.

   UltraSPARC T1 machines haven't been shipping for very long.   
Support will come along in due time.  It takes Sun just as long to  
write hardware support for Solaris as it does the NetBSD developers  
for NetBSD.  Sun's advantage is that they get to delay the  
announcement of the machines until the OS is ready.  Free OS  
developers don't have that advantage...and this has nothing to do  
with those OSs being "free" either...it has to do with the fact that  
they're "not Sun".

               -Dave