Subject: Re: improving project maintainership
From: Michael Tiemann <tiemann@redhat.com>
Date: Fri, 08 Feb 2002 08:20:19 -0500

Simon: I disagree with your point #1.  Having a bunch of maintainers provides 
tremendous competitive advantages:

a) when we bid contract word, we have a damn good idea what all is required to 
deliver the goods.  Many of our competitors (both pre-Red Hat and current Red 
Hat) grossly underbid projects because they don't understand things from a 
maintainer's perspective.  They may win the short-term business, but they lose 
customer satisfaction, and we've won more than a few $1M+ deals because of bad 
judgement by our competition.

b) maintainers have a much better view of overall product quality, hence we can 
be more intelligent about when to branch, what needs the hardest testing, etc. 
  Red Hat was last to market with a 2.4-based Linux distribution, but were the 
first to ship such a product with no known data corruption bugs.

c) even if maintainers can be impartial about the source of an inquiry or their 
response to a patch, it is human nature to suspect that if one is competing 
with a company that has a maintainer, one wants to keep at arms length from the 
maintainer.  That disadvantages one's technical position, giving an advantage 
to the company who has the maintainer.

Of course, Great Bridge had 3 of 6 core maintainers for postgresql and that 
didn't help them, but that's because there wasn't *any* significant postgresql 
business.  I'm sure that if there were $200M/year spend on the postgresql 
database, GB would have done very well with their position.

M

Simon Cozens wrote:

> Tom Lord:
> 
>>"More resources" means that there should be a greater number of more
>>senior people being well paid to be full time maintainers for shared
>>projects.  There should be a self-imposed "tax" on FS-based businesses
>>for that purpose.
>>
> 
> In some cases, this already does happen; VA Linux, for instance, picked
> up a lot of good guys and paid them to maintain their projects full-time.
> On the other hand, when the economic downturn happened, these were the
> first against the wall. 
> 
> But in a lot of other cases, it isn't going to happen, for two major
> reasons, as seen from the FSB's point of view:
> 
>     1) shelling out to make someone a full-time maintainer doesn't give
> the company any competitive advantage other than the dubious quality of
> "prestige". There's no business case for doing it; it might be nice,
> public-spirited, and generally a great idea, but, like it or not,
> the purpose of a business is not to be a good citizen, it's to make money
> and keep the shareholders happy.
> 
>     2) Developers cannot hold companies to ransom; the whole point
> of free software is that nobody can hold anyone else to ransom over it! 
> And so there's no way established maintainers can *make* it worthwhile 
> for a business to pay their wages. If I suddenly declare that I'm not going
> to do any more maintainance work on a project because I want to be paid
> for it, someone else will take over, and I end up looking foolish. And
> people could be quite justified in questioning my motives for maintaining
> the software in the first place.
> 
>