Less than one week after Sun Microsystems’ chief operating officer Jonathan Schwartz pledged to open source its Solaris operating system, The SCO Group has stated that licence restrictions prevent Sun from contributing its work to the GPL (General Public Licence).

Solaris is based on Unix System V, the source code to which has since been acquired by The SCO Group.

The SCO Group’s marketing manager Marc Modersitzki said although the company can't discuss specific details of its licence agreements, it is confident that Sun will be very rigorous in complying with its Unix System V licence as the company defines its plans for open-sourcing any part of Solaris.

“While the details of Sun's plan to open-source Solaris are not clear at this time, Sun has broader rights than any other Unix licensee,” Modersitzki told Computerworld. “However, they still have licence restrictions that would prevent them from contributing our licensed works wholesale to the GPL.”

Modersitzki said by spending more than $US100 million in Unix licence fees Sun has the broadest rights of any of SCO's Unix licensees and has been a licensee in good standing for many years.

Although Sun has not publicly stated under which licence it intends to release an open source Solaris, Schwartz said: “Make no mistake, we will open-source Solaris.”

In an interview with IDG News Service last month, Schwartz said "maybe we'll GPL it" and "We're still looking at that".

AUUG president and FreeBSD developer Greg Lehey said he can think of no evidence that would invalidate SCO’s statement.

“Unless Sun has an exclusive licence for the Unix code, which it obviously doesn't, it would find it difficult to change that licence,” Lehey said. “I suppose the interesting question is if Sun releases those parts of Solaris which it has developed in-house, what does that do to the code it has licensed?”

Lehey said the GPL has the so-called "viral" effect, and that would theoretically cause the remainder (of code) to fall under the GPL as well, “but that's so preposterous that I can't think of any way it could happen”.

“If there's any difference of opinion here, it would be more likely to be related to whether SCO agrees that Sun has written the code in question,” he said. “We've already seen that SCO is very hazy about the origins of the individual components of its code base – witness its display of the Berkeley Packet Filter as ‘its own work’ at http://www.lemis.com/grog/SCO/code-comparison.html#BPF.”

Regarding SCO’s lawsuit against IBM for allegedly improperly contributing code to Linux, Modersitzki released a statement by The SCO Group overnight requesting the trial be put back until September next year and to bifurcate the trials for IBM’s counterclaims which include GPL violation and four patent infringements.

“June 8, 2004 SCO versus IBM motion to amend scheduling order; motion to bifurcate. Update: SCO is making a motion to move the scheduled trial date to September 2005 and split IBM's counterclaims into a separate case,” the statement read.

Modersitzki also referred to a recent SCO memorandum regarding discovery dated May 28 stating that IBM be requested by the judge to provide specific evidence that files produced so far show that IBM improperly contributed code to Linux.

“In the absence of the requested discovery items, SCO has had to rely upon some alternative sources for proof,” the memorandum read. “IBM has so far only produced selected pieces of AIX and Dynix.”

On the last point, Lehey said without knowing the background, it's difficult to assess this statement.

“It rather looks like IBM has supplied individual source trees rather than the source code repository. If so, this would have been sub-optimal: it's one of the things that we consistently referred to last year,” he said. “I think it's valid to ask for the revision control information. I'm a little surprised that IBM didn't do so, but I could imagine that it hadn't been asked to do so. Now it has.”

According to SCO, “The burden on IBM is negligible for gathering the requested versions of AIX and Dynix” and that “gathering these materials is, for a competent engineer, a rather trivial task”.

To this, Lehey said depending on what they're looking for, they could be right.

“I would suspect that the best thing for IBM to do would be to print out every single version as requested and send the resultant 20 tonnes or so of paper to SCO. That would keep them quiet for a while,” he said.