I'd like to reopen the discussion of "free software" style license terms for HDL code. There seems to be a lot of ambiguity about software-oriented licenses (especially the LGPL) mean in this context. In practice, any time I've wanted to use an LGPL-licensed core for work, I've talked with the author personally to clarify what they do and do not intend to allow. This only works when the author is reachable, and when he or she has the authority to choose/change the license.

I'd like to think a bit more about how HDL code is used differently from traditional software, what rights a free software license *ought* to convey in this context, and how well existing licenses match that.

I've started a draft of some thoughts at https://github.com/ewa/free-hdl-license . If you're interested, please take a look at that and give me your thoughts! Feel free to edit the GitHub wiki directly, but *please* only do so if you're OK with me using whatever you put in there. And, of course, you can comment here.

(The main Wiki page is https://github.com/ewa/free-hdl-license/wiki )

We did have this discussion for a potential or2k. Some pros and cons can be found here
http://opencores.org/or2k/Licensing

Is the discussion itself archived anywhere?

Hi ewa

You might like to have a look at the CERN Open Hardware License used by the Open Hardware Repository (www.ohwr.org).

HTH

Jeremy

--
Tel: +44 (1590) 610184
Cell: +44 (7970) 676050
SkypeID: jeremybennett
Email: jeremy.bennett@embecosm.com
Web: www.embecosm.com

Thanks, Jeremy.

The CERN OHL seems a lot like the TAPR OHL, in that they're both really oriented toward physical artifacts and not software/firmware. In fact, they both explicitly exclude code running on programmable devices. If I want to make a circuit board, or a hammer, either of those licenses pretty much cover it. If I want to make a compiler or a driver, the GNU or BSD licenses work pretty well. If I want to write code which might go into an FPGA, or might go into an ASIC, or might be used for simulation and verification on someone's computer, I don't think any existing license really works well.

Here's the crux of the problem for me: If you download a core that's licensed under the LGPL, do you _really_ know how you can use it? I don't. I've e-mailed individual authors to clarify what they intend to allow, but that doesn't scale. My goals here are to (1) identify how well existing licenses apply to what we do, especially in terms of their language, (2) think about what rights and expectations a free/open license for HDL code _should_ contain, and (3) if necessary think about how to write (or revise) a license to meet those needs.

-Eric

Hi Eric,

You hightlight the need for a license suited to FPGA/ASIC.

There has been some discussion of this on the OpenRISC mailing lists (with particular reference to OR2K). You might like to start a thread of discussion there, since I don't think many of them watch this forum (see http://opencores.org/or1k/OR1K:Community_portal#Mailing_lists for details).

HTH

Jeremy

FYI. Here are some more folks discussing this same problem. Just showing that this is indeed a need.
http://www.rhinocerus.net/forum/lang-verilog/168090-using-ip-core-licensed-under-lgpl.html