OpenRISC Will Be Accepted Into The GCC Compiler

[Weasel]

I guess as a millennial I just don't really get the concept of copyright and ownership of non-tangible things.

Ownership is like being root. Only the owner has the root password/access.

Others are just given rights to use it (licensed) as regular users with respective capabilities.

And of course if there's an exploit only root can solve it.

[lowflyer]

No its not asking for trouble at all. And its a clean room from scratch coding of the the bits & pieces to get OpenRisc support in the upstream GCC codebase which is a good thing. 100% legal and ethical. It's not any different then being pissed off at FreeBSD or Linux for being clones of UNIX.

I'm not really sure you got my point. The term "clean room rewrite" is a term that is used to fend off lawyers. It does not mean much to a developer. For a developer a "rewrite" is a "rewrite". Whether you call it "clean room" or not does not make a big difference. However for the lawyers it clearly sends the signal: "This is completely new work" and thus effectively blocks any attempts to call out the rewriters.

My point is not about that. Even you admit in your answer that there are "pissed off" UNIX guys out there over FreeBSD or Linux. Now, we're talking about the OpenRISC project here. Having "pissed off" one or two developers in a small-ish project like OpenRISC can be a serious issue. Therefore let me ask again:

Was this done with the consent of the original developers?

Legality is not the point here. Other projects got into trouble over less serious issues. Losing one or two bright developers can mean the death of an open-source project.

[jbaxter]

I'm not really sure you got my point. The term "clean room rewrite" is a term that is used to fend off lawyers. It does not mean much to a developer. For a developer a "rewrite" is a "rewrite". Whether you call it "clean room" or not does not make a big difference. However for the lawyers it clearly sends the signal: "This is completely new work" and thus effectively blocks any attempts to call out the rewriters.

My point is not about that. Even you admit in your answer that there are "pissed off" UNIX guys out there over FreeBSD or Linux. Now, we're talking about the OpenRISC project here. Having "pissed off" one or two developers in a small-ish project like OpenRISC can be a serious issue. Therefore let me ask again:

Was this done with the consent of the original developers?

Legality is not the point here. Other projects got into trouble over less serious issues. Losing one or two bright developers can mean the death of an open-source project.

I started my involvement with the OpenRISC project in 2008, working for ORSoC in Stockholm, who then ran OpenCores, which they'd purchased from the remnants of the original OpenRISC team in 2007.

The original OpenRISC team forked all of their open source development into a proprietary spinoff called Beyond Semiconductor in 2005. They still exist. Since then, not a single line of code has been contributed or forum/mailing list post has been made from anyone involved with Beyond. That remains the case to this day.

Now jump forward 8 years, and a new community had emerged around the OpenRISC project, and we got sick of maintaining an out of tree GNU tools suite. Some very dedicated folks tracked down every contributor to the binutils, GCC and GDB ports and asked them to go through the process of assigning their copyright to the FSF, as you're required to do in order to get your tools upstream. This is more of a practical than an ideological matter for most.

It took a couple of years, but every single contributor was contacted and eventually we all signed over our copyright assignment for the port work to the FSF. All except one, Matjaz Breskvar, who still works at Beyond and I believe is quite senior. We invited him to the OpenRISC conference a couple of years ago (actually now a general open source digital design conference called ORConf) to ask him in person to help us get the GCC port upstream. Although, we couldn't actually identify which commits he'd made (the original CVS repos are no longer online, commit history was lost by someone ineptly doing the changeover to SVN) I believe his name was in the headers somewhere, we still needed his permission. He refused. The reason he gave was bizarre but something to do with company politics, and sounded irrational to me. It had nothing to do with OpenRISC, it was something to do with other members of his company not liking the fact that he might sign over copyright to the works done before they did the proprietary fork. Whatever the situation he was in with the other members of this company (who seem like dicks considering the entire company was founded on an open source project), it still seemed unreasonable to us for him to deny us of getting the tools upstream.

Anyway, that's the history, and that's why the port needed to be re-done from scratch. One guy who didn't have a very good reason. Who hadn't been a developer on the project for more than a decade.

To answer your question about the original developers - they had been gone for more than a decade, and hadn't been in touch with anyone in the community since they left, and quite frankly were being dicks about helping the community out. So, no, we didn't seek their consent, for the obvious reason that they didn't care any more.

On a positive note, It's been a tremendous amount of work for one man, but Stafford is clearly very talented and committed and we're very happy he's done so.

[jbaxter]

I'm not really sure you got my point. The term "clean room rewrite" is a term that is used to fend off lawyers. It does not mean much to a developer. For a developer a "rewrite" is a "rewrite". Whether you call it "clean room" or not does not make a big difference. However for the lawyers it clearly sends the signal: "This is completely new work" and thus effectively blocks any attempts to call out the rewriters.

My point is not about that. Even you admit in your answer that there are "pissed off" UNIX guys out there over FreeBSD or Linux. Now, we're talking about the OpenRISC project here. Having "pissed off" one or two developers in a small-ish project like OpenRISC can be a serious issue. Therefore let me ask again:

Was this done with the consent of the original developers?

Legality is not the point here. Other projects got into trouble over less serious issues. Losing one or two bright developers can mean the death of an open-source project.

I started my involvement with the OpenRISC project in 2008, working for ORSoC in Stockholm, who then ran OpenCores, which they'd purchased from the remnants of the original OpenRISC team in 2007.

The original OpenRISC team forked all of their open source development into a proprietary spinoff called Beyond Semiconductor in 2005. They still exist. Since then, not a single line of code has been contributed or forum/mailing list post has been made from anyone involved with Beyond. That remains the case to this day.

Now jump forward 8 years, and a new community had emerged around the OpenRISC project, and we got sick of maintaining an out of tree GNU tools suite. Some very dedicated folks tracked down every contributor to the binutils, GCC and GDB ports and asked them to go through the process of assigning their copyright to the FSF, as you're required to do in order to get your tools upstream. This is more of a practical than an ideological matter for most.

It took a couple of years, but every single contributor was contacted and eventually we all signed over our copyright assignment for the port work to the FSF. All except one, Matjaz Breskvar, who still works at Beyond and I believe is quite senior. We invited him to the OpenRISC conference a couple of years ago (actually now a general open source digital design conference called ORConf) to ask him in person to help us get the GCC port upstream. Although, we couldn't actually identify which commits he'd made (the original CVS repos are no longer online, commit history was lost by someone ineptly doing the changeover to SVN) I believe his name was in the headers somewhere, we still needed his permission. He refused. The reason he gave was bizarre but something to do with company politics, and sounded irrational to me. It had nothing to do with OpenRISC, it was something to do with other members of his company not liking the fact that he might sign over copyright to the works done before they did the proprietary fork. Whatever the situation he was in with the other members of this company (who seem like dicks considering the entire company was founded on an open source project), it still seemed unreasonable to us for him to deny us of getting the tools upstream.

Anyway, that's the history, and that's why the port needed to be re-done from scratch. One guy who didn't have a very good reason. Who hadn't been a developer on the project for more than a decade.

To answer your question about the original developers - they had been gone for more than a decade, and hadn't been in touch with anyone in the community since they left, and quite frankly were being dicks about helping the community out. So, no, we didn't seek their consent, for the obvious reason that they didn't care any more.

On a positive note, It's been a tremendous amount of work for one man, but Stafford is clearly very talented and committed and we're very happy he's done so.