[OpenIndiana-discuss] Developer funding model musings

[Udo Grabowski (IMK)]

On 01/29/13 06:47 PM, Reginald Beardsley wrote:
>
> --- On Tue, 1/29/13, Udo Grabowski (IMK) <udo.grabowski at kit.edu>
> wrote:
>
>> From: Udo Grabowski (IMK) <udo.grabowski at kit.edu> Subject: Re:
>> [OpenIndiana-discuss] Developer funding model musings To:
>> "Discussion list for OpenIndiana"
>> <openindiana-discuss at openindiana.org> Date: Tuesday, January 29,
>> 2013, 10:50 AM On 29/01/2013 17:36, Reginald Beardsley wrote:
>>>
>>> I'd like to suggest as a social convention that the
>> initial "license fee" be 10% of system cost
>>
>> UAHHH ! Then we should have stayed at Oracle......
>>
>> Really, if you solve a problem for one system, why should someone
>> pay for the other 99 identical systems (think of big clusters) ?
>> Just because he has to money to buy them ? Think of all the EDU/R&D
>> institutions who simply don't have the money for such exorbitant
>> service fees.
>>
>> Remember, OI was started as a community effort, not to make money.
>> If "commercial" support is offered, it's a good thing, but we
>> should not try to mimick the unacceptable behaviour that companies
>> like Oracle show.
>>
>> Have a look at the old service catalog that Sun was offering, that
>> was something really reasonable and affordable, for both sides. It
>> had hardware-only services, cheap per workstation support, bulk
>> service fees, site service fees for an extended period, etc. Also
>> consider per incidence fees, based on real effort needed to solve
>> the problem.
>>
>> Avoid copying business models made for banks and assurance
>> companies....these are not the clients of OI. --
>
> Where I live we have an 8.25% sales tax.  It's also NOT voluntary.
> I'm spending $800US to build an OI based file server.  I certainly
> don't find contributing $80 for OI development exorbitant.  Oracle
> would want $1000 which is more than the hardware cost and I don't
> think support for the N40L is even an option.
>
> If you're prepared to spend $100k on hardware for an institution and
> not prepared to spend $10k to support the software, I think you're
> being unrealistic.  You'll spend many times that for in house staff
> and likely not do a good job of sharing the results w/ the community
> because of the extra work of distributing it.
>
> But if you have a better model for providing financial support for
> OI, let's hear it.  Not arm waving about what's fair, but practical
> details such as how to raise enough money to employ even one person
> full time for a year working on OI.
>
> Oracle has abandoned the small business and technical workstation
> markets because they can't make money at it.  If OI goes away for
> lack of support, I'm going to be forced to either FreeBSD or Linux.
> I happen to like Solaris enough to be willing to pay for it
> voluntarily.
>
> Please remember, Sun went out of business.  That's not a compelling
> argument in your favor for copying their business practices.
>
> FWIW I abandoned an open source seismic processing code I'd supported
> for ~15 years because I couldn't get any support funding. For 10
> years it was part of my day job, but completely unpaid for the last
> 5.  Lot's of people use it, but no one wanted to pay anything.  Once
> it sank in I wasn't doing anything w/ it myself, I decided I wasn't
> working on it anymore unless I needed something for myself.  I'd
> spent hundreds of hours fixing other people's problems w/o even
> "thank you" in far too many instances.
>
> You're getting paid and complaining that 10% is too expensive.
> Martin isn't getting paid anything except a few small donations and
> he's making things work.  You don't like 10%. Then decide what you do
> like and send it to Martin.  Better yet, get him a job at your
> institution.
>
> Have Fun!

To bring you back on your feet: We (a German government research
institution connected to a university) with a occasionally (~ every
5 to 8 years) get the opportunity to renew our machine park by
exceptional funding through the German state or state institutions.
This is usually an exceptional amount (could be a million €) dedicated
for HARDWARE ONLY. Since in science, no amount of money can really get
what we need to do our work, we usually put that into the fastest stuff
affordable at that time, and then keep it running until the next
opportunity (or until it breaks). So after that exceptional funding,
we are thrown back to our yearly budget, which is less than 10% or
even 5% of that amount, for our TOTAL science activities ! So we have
to pay personell and services from that, and 10% of system costs/year
is simply MONEY WE DO NOT HAVE. That was exactly the reason we had to
leave Oracle.
And, we are called "the rich" here, most universities have to live
from much less money than we have. It's just the broken financial
model that causes us these calamities (and no, that cannot be changed,
outside our limits).
And no, sorry for Martin, we do not need SPARC, it's simply not
up to date in performance for our purposes, and that field really
should be left to Oracle and the big money, global players.
But we would spend a fair amount of money for dedicated services
if they are AFFORDABLE for us, like the old Sun services model was
(and we had a lot of money in there). And, not to forget, we are
contributing to OI and illumos as far as our time and abilities
suffice.

So please do not judge about other institutions finances from
the limited insights of a private home server operator.