Discouraging contributors (Was: g++.old-deja/g++.other/anon7.C)

[Daniel Berlin]

>>
>> only sue for infringement based on the right they own).
>
> I've read about these legal issues on the FSF site, but it's not the
> point I was trying to make.
>
> Not loosing or upsetting lots of contributors is much more
> important than being able to eventually sue someone in the unlikely
> event that it becomes necessary in a distant future.
This is probably true
However,
1. I've not seen evidence that we lose *lots*, or upset *lots* of 
contributors through this process. It is required of all GNU tools.
2. As time passes, it becomes *more* likely that the need to sue 
someone will need to happen, not less likely.

>
> Most companies know how it works in practice and prefer to take the
> risk of not being paid by a few customers instead of forcing all of
> them to sign contracts ahead of purchasing.
>
No, they just use EULA's and click-through licensing, actually.
Copyright is different, since the contract *must* be in writing.
I can guarantee if contracts outside of writing were not valid for 
purchasing things, you wouldn't be able to purchase stuff online.

> It's especially true for insurances

I've never had an insurance policy i didn't have to sign, even if it 
was an online signature.
> , subscriptions, e-commerce,
> ISPs...

These all still require contract agreements, they just can be done 
electronically. They are therefore, bad examples, because it just 
proves that everyone feels the need for an official contract in order 
to be safe.  You can't do contracts for copyright transfers online, so 
it's not an available option.
>  They try to make it absolutely easy to become a customer
> to order or subscribe by phone or e-mail (and very hard to quit,
> sometimes ;-).
>
> In the case of free-software development, projects don't primarily
> compete for customers. The limited resource is developers: to be
> successful, a project must attract developers instead of letting
> them go elsewhere.

Where else do you think our developers go, exactly?

> Imagine there were two competing compilers: GCC and an imaginary ZCC.
> GCC requires an extremely annoying procedure to become a developer,
> while ZCC does it best for getting newbies involved.

> In a few years, ZCC receives a lot more contributions and
> outperforms GCC. Then everybody looses interest in GCC and
> switch to ZCC, leaving GCC in the dust.

> What's the point in being absolutely sure you can sue someone
> when you have no longer a product worth defending?

If this happens, it happens.  I don't believe it will.   The number of 
developers who are willing to survive the process  has always 
outnumbered the few who go away instead.
Not that i don't wish we could do the process electronically and 
whatnot, but that's unlikely to happen soon.
We are talking about having to wait a few weeks in order to be able to 
have patches accepted.
It's not like it's forever.
If it's taking too long, complain, and someone will poke and prod the 
right people.
Otherwise, use the time to read the style guidelines and become more 
familiar with the code.

> I'm sure Linus would have better arguments than me in explaining
> how much success you get by making it easy for people to get
> involved in a project.