[SunRescue] Re: Museum idea (was Offerings)

[Reagen Ward]

On Fri, Aug 20, 1999 at 08:40:37AM -0400, Kent Rankin wrote:

>     There have been a few mentioned attempts at this, but they usually come
> from students that can't come up with another excuse to use the gear for
> themselves.  =)  God... I guess my old "Well, I plan to port Quake to it
> one day" excuse just doesn't work as well anymore.  =)  Generally the gear
> gets wasted heavily in such cases(we're talking about larger machines,
> here).

I always used it as an excuse to my wife.  My plan is to have one of every major
production OS and architecture seen in enterprise environments.  Why?  Well, funny
you should ask...

I live in Houston, a city with a rather large IT community, and many OpenSource-
type programmers.  Wouldn't it be nice to join a club (with dues, unfortunately)
where you had access to one of everything for porting and testing?  I plan to get
a nice office space in the warehouse district (about 750 sqft; they don't limit
power utilization (included in the rent) and each space has its own A/C system)
and set it up.  Most machines will remain off unless needed, we'll settle for low-
end examples of an architecture (for instance, an HP H is an example of an HP 800
class system) for most work, and try to get specific examples when needed.  The
cost of membership would be put into the rent/acquisition pool, and donations 
of needed equipment would, of course, be credited towards dues.  Connectivity will
be a bit harder ; probably just a BRI for starters (I only have a Pipeline 75),
and maybe RoadRunner / SDSL / whatever later.  Storage would be a big cost, too,
so that every host had as many OS choices as possible.  No raised floor, though, 
and no three-phase power.  Alas.

Will it make enough money to stay afloat?  Of course not!  I'll be the sugar daddy
of the project.  Why?  I want to do it.  That, and my wife would like to have the
house cool off again.  Is it worth it?  Well, it sure sounds like fun, and it's a
lot cheaper than a gambling or crack habit.  Probably cheaper than a nice sailboat
or the like.  If I get tired of it or lose my job, I'll just donate the eq to the
NetBSD team or other such groups.  My wife would like for me to donate the current
19" cabinet to anyone who will make it go away.  It's a very ugly shade of blue.

>     And, really, if anyone will do it, and I can tell that it would be a
> good
> situation, I'd be more than glad to start contributing.  It wouldn't be too
> hard for me to obtain a number of supers for such an institution as long as
> they would be properly managed, and available to the proper sorts.

I really wasn't thinking about supers; my background is in scientific computing,
with Alliant, Convex, SGi, etc.  Mostly air-cooled supers.  I just want a low-
budget version of a software house's porting lab (I work for BMC.. we have some
fantastic testing / porting / QA labs) that is easily within the price range of a
hobbyist, late-night programmer, etc.  Supers are fun, mainframes are fun, but the
support costs would be prohibitive.

Too many cool packages only have binaries for Linux.  Too many folks develop just
for Linux, and completely ignore the rest of the OS world.  Imagine being able to
download binaries of your favorite package (say, mutt or the like) for just about 
every OS/architecture combo you can imagine.  Well, maybe you like compiling (I do
enjoy it, but I'm weird that way), but it really helps the average user try 
software.

The machines would also be good for OS porting (for NetBSD, etc).  I've got a nice
32 port MUX and a 16 port Xyplex for serial consoles and I plan to have  an X-10
switch on each box; you get a remote power switch and a remote console, all you 
need for OS porting.

Reagen