[oclug] Sat evening rant

Sandy Harris sandy at storm.ca
Sat Feb 17 14:37:03 EST 2001


I'm not at all sure a challenge based on Linux instead of Windows 2000 would succeed.

No doubt one of the requirements is that whatever they buy run their MS Office and
whatever else they have. Yes, there are products for Linux with similar functionality,
but that's an uphill fight, and a long hill.

Also, there's Wine, but I'm not certain it is entirely functional (especially for
running MS code, which may have various undocumented tricks) and there are overheads.

The gov't contract I'd like to see someone bid on with open source software is the
office suite contract. A few years back DND did one for several * 10,000 seats, MS
won and Corel screamed. I'm hazy on details.

So assume once they have their 50,000 seats of Win 2000, they want a 50,000 seat
license for an office package. Expect MS and Corel to bid. Even at $100 a seat,
that is $5 million. Plus training costs on the new features, support costs, ...

What would it take to construct a winning bid against that for StarOffice? Sun
sell CDs of it for $10 US. Say we provide 500 CDs for the admins, one per 100
users. That's under $8,000. Not bad if the competition is at $5 million. Of course
an actual bid ought to cost more than that, and do more for the client.

For one thing, you'd provide a channel for upgrades, some method of making new
releases available for evaluation and, if they passed that, for use withing DND.

Ideally, include a mechanism to help get the product modified to DND specs.
What happens when they discover a bug? When they need some strange feature,
say sorting on NATO part numbers in their weird format? Methinks a bid might
include budget for a small team -- project manager, C hacker, support staff
-- who could look at such things. You could probably recruit everyone you'd
need off this list :-)

Small things they might do themselves, passing patches to both Sun and DND.
For most problems, they'd function as co-ordinators, checking whether there
was a work-around, finding out if the problem was general across DND or just
a small group, narrowing down the problem description, letting Sun know about
it, ... If a problem came up that Sun was not prepared to fix in a timely manner,
they'd be in a good position to spec the project for a DND fix or write the RFP
for a contractor fix.

Spend half a million a year on such a team for 5 years, and you're at
$2.5 million, still way below the $5 million guess at the competing bids.



More information about the OCLUG mailing list