... | @@ -318,7 +318,13 @@ has the right to make them available to the public. |
... | @@ -318,7 +318,13 @@ has the right to make them available to the public. |
|
|
|
|
|
#### Q: Is CERN-OHL-S compatible with GPL?
|
|
#### Q: Is CERN-OHL-S compatible with GPL?
|
|
|
|
|
|
A: No. From the rationale document:
|
|
A: No. To explain why, let's take the case of an FPGA design. If any block is GPL in that design, all other blocks have to be released under GPL. Same for CERN-OHL-S: one core as CERN-OHL-S would force a distribution of all other cores, along with the top-level design, under that licence. So it's easy to see how a GPL core and a CERN-OHL-S core in the same design would create a problem. This is a well-known issue in the software world, and the reason why people try to avoid the proliferation of different, incompatible, strong copyleft licences.
|
|
|
|
|
|
|
|
So, we could add a clause to CERN-OHL-S which says that any design licensed under CERN-OHL-S can be relicensed under GPL by the licensee. This would allow the combination of GPL and CERN-OHL-S in the same design. We decided not to do that for various reasons, in particular:
|
|
|
|
* We think that GPL [is not an optimal license for the case of hardware](https://www.ohwr.org/project/cernohl/wikis/CERN-OHL-v2-draft#q-why-not-use-existing-licences-such-as-gpl-and-any-in-the-family-of-creative-commons-licences). For example, in the case described above, after carefully reading the GPL3 text, one could very well argue that any primitive/macro libraries from the FPGA vendor (Xilinx, Altera/Intel...) should be distributed too, which is explicitly forbidden in the licensing of those libraries.
|
|
|
|
* When drafting the CERN-OHL-S and CERN-OHL-L, we took special care to guarantee, to the biggest possible extent, that the recipients of hardware get access to its design files. There are special provisions in the licence texts to ensure that, which don't exist in GPL. One example is the ability for a licensor to specify that the URL for a design should be physically engraved in the final product. Allowing people to relicense under GPL would provide an easy means to escape such obligations. The lack of an appropriate reciprocal licensing regime for hardware was one of the key reasons to draft the CERN-OHL to begin with.
|
|
|
|
|
|
|
|
The rationale document explains all this with other words:
|
|
|
|
|
|
> We thought hard about compatibility with other licences. An earlier
|
|
> We thought hard about compatibility with other licences. An earlier
|
|
> version of the draft contained a mechanism for options, and a specific
|
|
> version of the draft contained a mechanism for options, and a specific
|
... | | ... | |