[daisy] regarding the DB2 issues -- @Aaron

[Steven Noels]

On 22 Jan 2007, at 14:01, Aaron.Digulla at Globus.ch wrote:

> Unfortunately, I cannot submit patches against trunk because I have  
> to use
> the current production version of Daisy. I'm not paid to support  
> OSS, I'm
> paid to build production ready systems. Being the economy as it  
> currently
> is, this is something we both can't do something about. So while your
> argument makes sense from an OSS point of view, it doesn't make  
> sense from
> my point of view as a patch submitter.

Looking from where I sit, we're all properly fed, have an income and  
pay taxes, so there's obviously some sort of economy involved on our  
end of the line as well. We could discuss POV at length but that will  
scare people away- one can however not selectively ignore parts of  
the economical equation IMNSHO.

I'm extremely happy with any sorts of observations and remarks, but I  
don't like them oversimplified. It's easy to do "genuine" open source  
development if one doesn't have to worry about income. Our quest of  
the past three years has been two-fold: (a) showing serious open  
source product development is possible without external funding, and  
(b) educating people about this business model as we go along, hoping  
they will get it and jump aboard.

Because of that, I'm actually pleased with contextual discussions  
like this. It hopefully makes people realize things aren't quire  
simple as they have come to expect. I do realize for instance that  
business competitors of my company are reading this list. And I do  
realize that my customers also need production ready systems, and  
that it's because of them that Daisy is availably to anyone for free.

> My argument is this: It took me about thirty minutes to write these
> patches (the bigs ones are Search&Replace, the smaller ones are  
> just a few
> lines where most of the time was spent finding the line).
>
> I figure it will take *you* five minutes to look at the patch,  
> understand
> what it fixes and port it to trunk. Therefore, your constant arguing
> against it is only achieving to infuriate me especially since it's a
> non-issue really. You've already spent more time arguing that it would
> take to fix the issue, goddamnit :-(

True, and we realize that as well. Now, that was then, and now is  
now. No need to start yelling at each other.

Let's simply say that we didn't have time to take a thorough look at  
your patch then, that we were happy with your remarks in general, but  
now we have planning issues to spend time on your contributions  
because they need to be changed to reflect changes in trunk, and the  
whole issue definitely needs more time anyhow. And that your  
insistence makes it seem as if we don't welcome contributions as a  
whole, which simply isn't true.

The issue as such isn't very exciting nor hard to solve, it's what we  
as a project choose to support. Do we simply say "Daisy is tested on  
MySQL, but might work on DB2 and Oracle and PostgreSQL as well.", and  
expect users to debug? Or do we say "Daisy runs on MySQL, DB2,  
Oracle, PostgreSQL and some more RDBMSes, and there's actually  
someone who's tracking trunk development to guarantee so, and there's  
serious testing against each database and JDBC drivers before every  
release." What about database upgrade scripts between releases? Who  
will maintain those for non-MySQL databases? Who will provide end- 
user support when people have particular issues with a particular  
database?

I would be very happy to be able to say that Daisy runs and is  
supported on all major RDBMSes out there. However, I also know how  
many of them we can currently support ourselves headcount-wise. If  
anyone wants to change that, he or she will need to look at the  
bigger picture. In my mind, I don't want end-users - perhaps during  
their initial trial installation - to debug database support by  
seeing Daisy fail at them: that's bad publicity. If we provide  
support for other databases, it should be done as part of the entire  
development and release process.

> Okay, since it's now four months since I originally submitted the  
> patches,
> it will take you more than five minutes because of all the changes you
> made since but that's something I can't change. I did all I could to
> prevent this from happening.

There's a lesson about communication in there. We should (be able to)  
communicate about our whereabouts and what we are working on more  
frequently and persistently (in the sense that it *is* clear what we  
are working on, if you read this list on a daily basis, however maybe  
we shouldn't expect that from contributors). One more thing we should  
add to a revamped Daisy website.

Writing this mail indeed took me more time than economically  
justifiable. I also see now that Bruno has reached his point of  
closure about this matter. I hope some of what we wrote helps you (or  
your management!) to see the bigger picture, and not infuriates you  
even more.

</Steven>
-- 
Steven Noels                            http://outerthought.org/
Outerthought                              Open Source Java & XML
stevenn at outerthought.org                stevenn at apache.org