[plug] Looking for a quote

[Craig Ringer]

On Sat, 2004-06-19 at 00:35, Simon Scott wrote:

> Avoid the low-cost aspects of linux as a selling point, there is no need 
> anymore. As soon as people hear things like that, they want you to do 
> everything for free.
> 
> After all, you may want to make a living from it one day :D

The fact is, though, that that's the driving point behind why many
businesses are using Linux and the software available on top of it. The
main advantage the POST has had from Linux has been minimal licensing
costs, reasonable (but not great) administration costs, and the ability
to run quickly and reliably on low cost hardware.

I'm focusing on business use here, because individuals often have very
different motivations and priorities.

Open source software, including Linux, has let us do a lot that we
couldn't have otherwise have done or couldn't have done as well. There's
no way our mail services would be this fast and reliable (thanks
Cyrus!), for example, or that we'd have the sort of on-line quick access
archival storage we do. There's zero chance we'd have the editorial
story database (the most useful one day's worth of programming I've ever
done) if we couldn't build it in house and on top of open source tools.
There's no chance we'd be using a network of thin-client computers that
take almost no admin work and have a tiny per-client cost if the
software stack wasn't free and lightweight enough to let us run it on
ancient hardware. Mostly that's been because we couldn't have _afforded_
to do things, rather than them not being otherwise technically possible.

Anything we're doing on Linux we could almost certainly do on Solaris or
(with more work) 2k server with other software to provide for our
higher-level needs. It'd just cost us a lot more in hardware, software
licensing, etc. We'd be buying a small mountain of software, including
mail servers, compilers, development apps, web servers, databases,
operating system licenses (client and server), client access licenses
(!!!), and all the other things I just forget most people have to worry
about. It'd be prohibitively expensive to run our network, frankly, as
we use a LOT of diverse software across a mixed client base, connected
to a number of legacy systems.

Even the admin time advantages (which very a _lot_ depending on what
exactly you're trying to do - I'm sure Win2k wins hands down lots of the
time) etc all come down to cost in the end, don't they?

In the longer term, we do pay attention to the fact that we have broad
rights to use the software in flexible ways. It's nice to know that your
vendor will find it a little harder to declare that "you WILL upgrade".
You're rarely forced to get a maintenance contract just to get minor bug
fixes. You can test software to see if it'll be useful to you without
having to buy copies just to get ones that aren't too crippled to test.
It's great to know that if the PSF goes out of business and everybody
gets bored of Python, we still have the code and a reasonable chance of
being able to get it ported to another platform down the track. Having
been bitten - twice - by this first hand, I really _know_ the importance
of it. Knowing that the business(s) supplying your software going out of
business doesn't mean the end of the software is wonderful beyond all
description.

Even all that, though, really comes down to money. After all, we're
currently replacing one of those legacy systems that we can't port and
can no longer maintain. It's just a painful and expensive process that
it'd be much nicer not to have to go through.

Security is the same. In the end, better security - to a business - is
just one less cost. Cost in sales, when your customers hear your SSL web
server with credit card processing was cracked. For the third time. Cost
in staff time - both directly, for tech staff, and indirectly in lost
work - in clean-up after the latest worm, just in time for the next one.

Linux, and the open source software built on top of it, has given us a
very powerful, nice operating system and software suite at an unbeatable
cost, and enabled us to use hardware that provides similar advantages.
It's let us do lots of things there's no way we could've possibly done
otherwise.

For a consultant, that should translate into the ability to beat the
competition more reliably, and into higher margins for the services you
offer. As the business end user, that should mean lower costs and the
ability to afford to do more of the things you want to do with your IT
infrastructure.

(Uggh... I'm beginning to sound like a right PHB)

--
Craig Ringer