Sure it was, but how much additional effort did Apple put into the code base by way of WebKit? Apple didn't need to open source that work, even if the original was open source, and I'm sure the leg up made a big time-to-market difference for Chrome.
Edit: Actually, KHTML is licensed LGPL, and that's a viral license if you make a derivative work. So I suppose Apple was forced to open source their changes when it finally went to market.
> Edit: Actually, KHTML is licensed LGPL, and that's a viral license if you make a derivative work. So I suppose Apple was forced to open source their changes when it finally went to market.
When LGPL code is linked with non-free code and the resulting work is distributed LGPL only requires that the changes to the LGPL code be open and that the object files for the non-free code are provided so that users can exercise their LGPL rights with regard to the LGPL portion of the work.
Apple could have separated out much of their work into separate files that contained no non-Apple code and that dynamically linked with the KHTML LGPL code and they could have kept those files closed. They would only have had to open the changes to the LGPL code.
If we accept the terminology that GPL is a viral license then LGPL is more of an immune system than a virus.
Nowhere did I say or imply that Apple was a saint. All I said was that LGPL allows linking with closed code and distributing the resulting work without having to make the closed code open, and that Apple could have structured WebKit in such a way as to keep most of their code from having to be made open.
This is the main difference between GPL and LGPL, and a big reason why the FSF now discourages the use of LGPL.
> The reason webkit is known today, and not KHTML, is because Apple ignored their LGPL requirements.
> First, they didn't publish anything for years.
They published their KJS changes in June 2002, shortly before they released products using it. They published their KHTML changes shortly after Safari was announced.
> Then only obfuscated source dumps with all comments removed for many years.
Citation needed. My recollection is that they published their code in the form that they used it internally. The difficulty for the KHTML developers in incorporating Apple's code was (1) WebKit was designed to do what Apple needed, which was not necessarily what KDE needed, and so the two were diverging, and (2) Apple was making a lot of changes that touched a lot of code making it hard to break it down to individual improvements that could be incorporated independently.
Also (3) Apple had poured way more resources into this than the khtml developers had available. Since Apple kept its work secret for business reasons, that meant that, once Apple published their fork, the sheer amount of changes overwhelmed the KHTML developers (WebKit wasn’t just a port of KHTML; it broadened support for html features)
Yes, but they didn’t even attempt to get it merged, nor did they provide the changes on a per-commit level, but only as a single sourcedump with all comments etc removed.
Apple's JavaScript implementation is based on KJS as well. Arguably, Microsoft was hostile towards open source in general at the time, while Apple simply stole open source software.
Legally they might have done the minimal amount necessary to not violate the license, but morally they've been a huge dick. "We have to open source our modifications to the code base, so here you get exactly one tarball of uncommented stuff."
To be practical, it would have been significantly painful for Apple to adapt and improve KHTML by keeping their code and changes outside it and only dynamically linking to KHTML.
If it were just as easy (or even comparably easy) to carry on (and maintain) development via external linkage, LGPL would have been long useless.
Edit: Actually, KHTML is licensed LGPL, and that's a viral license if you make a derivative work. So I suppose Apple was forced to open source their changes when it finally went to market.