IANAL, and I’m not sure Oracle lacks standing. But if they lack it, they lack it.

And the deep pockets thing — e.g. through appeals to relitigate whether they have standing — could work on both sides of the case. Not everybody with deep pockets wants Oracle to have monopolist rights on a particularly popular piece of open source database software.

You argue “I don’t know why Boogeyman
Oracle would have standing to sue”.

Oracle just happen to have plenty of
lawyers and enough cash to be able to
sue without standing. By the time the
dust clears there may be nothing left
of the company they sued.

Assuming rational behavior is clearly
unwarranted.

anonymous

The point of the layer is to GPL part of your code, getting the original GPLed program to the point that the rest of your code can have a more APIish connection to the first (now altered) program.

Actually, that could be done either by adding to or subtracting from the original program, whichever made more sense. In the case of MySQL, it might be subtraction, using only the parts that provide connectivity to the rest of the world (BI tools and so on).

On the definition of derivative work, the GPL itself is vague and inconsistent, but there do seem to be a few points of consensus among the legal experts. The first is that dynamic vs. static linkage is probably not a defensible argument in the “one program or two?” test. If the programs communicate across a network protocol, say through web service interface, then the coupling gets looser and confidence is higher. Even in the loosely-coupled configuration, however, there is a prevailing notion that programs exchanging “complex, non-standard internal structures” operate as a single unit.

The “non-standard” part of that argument relates to the point you raise about vendors who deliver proprietary solutions atop the Linux libraries. Since those are considered to be standard system libraries, programs that call them are not exposed to the GPL. In your prior thread, Barry Kiawans touches on this point, saying that he believes the use of standard interfaces like ODBC and JDBC offer good insulation from the GPL. There are some pretty good lawyers who would agree with Barry and say ODBC/JDBC are the RDBMS analog of Linux system libraries.

With regard to MySQL storage engine technologies, they 1) integrate with MySQL via an API that would probably not be considered a system interface, and 2) they exchange complex, internal structures. It seems pretty clear to me that MySQL storage engines are derivative works of MySQL. And I think the notion of wrapping the storage engine API with non-GPL’d code to create a GPL buffer is flawed – by definition the buffer code will be a derivative work, triggering flow-down of the license.

To your question of legal pursuit against someone who violated the GPL via a MySQL fork, I believe Oracle probably would have standing to sue as the original copyright holder. Again, I’m not a lawyer, but this would make sense to me. Oracle could alternatively ask the Free Software Foundation, which certainly does have standing to litigate open source license violations, to step in, possibly with Oracle underwriting the FSF’s legal expenses.

Looking up, I’ve just typed a lot of words! I hope some of them make sense and are additive.

Fred