With Windows CE, though, derivative works bear a fee responsibility to Microsoft for every copy sold, something which doesn't apply to FreeBSD or Apache, the base for which is free as in cost. Other differences emerge if you compare Windows CE licensing terms to the GPL terms under which Linux is licensed. GPL licenses require derivates to be licensed under the GPL, which means that all source code must be released to all derivative works. In other words, derivative works cannot keep their changes secret, something which wouldn't appeal to companies hoping to derive revenue from sale of software as such.
Microsoft's shared source efforts aren't automatically a different planet from the rest of the open source world. Fees are often involved, but as the Windows CE example shows, Microsoft can offer access to source code on terms comparable to other open source licenses.
What I'm hoping to see
I'd like everyone with a valid licence to SQL Server to get access to SQL Server source code. Perhaps that's self serving, but I'm not 1,500 people (though my friends under the pillow say I am), and I want to see the SQL Server code, anyway.
Why would I find that useful? It's not because I want to make a derivative variant of SQL Server. I could care less about database development, not because there isn't anything interesting in it, but because I have 80,000 other programming responsibilities which take precedence over worrying about the proper way to write a database. Division of labour is a good thing, and I've made the decision that my time is not best spent worrying about writing a database.
Rather, I want to debug my own code with access to SQL Server source code. That's extremely useful. If something unexpected happens, whether catastrophic or just weird, I can dive into the code to figure out what instructions were executed. This is better than trying to untangle assembly language, which is God's way of saying you have no life.
Access to source code also makes things more predictable. When I make a call into SQL Server, I'll know exactly how it will handle the data I'm passing because I can see how it is being handled.
Again, none of those things require that I have the ability to make derivative works. 99.99999% of the programmers in this world do not need to make a derivative version of SQL Server, and many among that last 0.00001% probably shouldn't, anyway. To put it simply, programmers who don't work for database companies and mess around with the inner workings of a database are like auto mechanics that waste time trying to fix wonky plumbing. Their time is better spent concentrating on the things they do best rather than obsessing over areas they are not experts.
Consuming red herrings
Some characterise the proposed release of SQL Server source code as Microsoft placing the burden for finding bugs in their system on others, while keeping the benefits for themselves.
Imagine I owned a beach. Beaches are big. I try to comb the beach for pieces of glass and other nastiness which might make life unpleasant for visitors, but god knows I'm human, and it's a big beach. So, I ask visitors to tell me if they find anything unpleasant. They're all over the beach, and a thousand pairs of eyes are better than one pair. This enables me to fix the problem by knocking them out with a club (no, no, removing the offending item).
Continued ...
7%
3%
