Re: [bug #37411] Diagonal matrix attribution error and diag function bugs

[Jordi Gutiérrez Hermoso]

On 21 September 2012 16:52, Paul Soderlind <address@hidden> wrote:
> I can well see that developers could ask for payment to fix issues that are
> (a) not important and (b) tricky to fix. But, the developer has to explain
> that in a polite manner to the user - and give an offer (before the work is
> done). This is not the same as asking for payment for work that is already
> done, at least not in terms of attitude.

Can you explain this a bit more? I sincerely do not understand the
argument here. This is my attempt to understand:

When you hear "free software" or "open source", the first thing you
hear is "no price". This is the most important thing to you, to get
Matlab without a price tag. The parts of "free software" that are
valuable to me, open bug tracker, no black boxes, no license manager,
community work; are all secondary to you. Thus, when I asked for money
for a bugfix you expected to get for no pay, this no-price promise
would feel broken to you, and I would appear like an extortionist,
holding a bugfix ransom.

Does that approximate your thoughts?

This was what happened in my head. Initially, I thought the fix would
be simple after I tracked it down. So I was expecting to spend like 5
minutes for it. 5 minutes of my time initially did not seem like
something worth asking for money, so I didn't ask for money initially.
I started doing the work, thinking it was no trouble for me to just
fix this bug. Then I looked at the clock, wow, 6 hours past. I also
thought of all the frustration I went through, and I thought, why am I
even doing this? I'm not getting anything out of it. I don't care
about this bug. If I am doing this for someone who cares, then the
person who cares should have some involvement in this. So I thought,
the person who cares should be the one doing this work or otherwise
paying for it. Which is why I then asked for money or a patch.

My reasoning was that if the person didn't want to pay nor do work,
then they wouldn't have lost anything, and the only thing I would have
lost was the time I spent doing this. I figured they had something to
lose if they didn't pay, the bugfix. Furthermore, by not setting a
price, I thought the other person could set their own price on what
this bugfix was worth to them. I would accept this price at a loss,
since I was stupid enough to start working on something I didn't care
about for someone who might not even value the result. So it seemed to
me that by setting a price on it, we could both win something: I would
get some money to offset my work, and they would get a bugfix they
presumably cared about. It seemed like a net positive gain for both of
us. It seemed like good business.

Any person involved in software development knows how difficult it is
to estimate how long a project will take. While the overall shape of
the bugfix was clear in my mind, the particulars of how to fit it in
with the rest of the Octave codebase were not. So there was no way for
me to know ahead of time if it really would take me 5 minutes or 6
hours. I could not have known that I was breaking the "no price"
promise, if that is the promise you thought you had.

- Jordi G. H.