1/1

|
How to sell an open core?
by Unknown on May 8, 2008 |
Not available! | ||
|
I was wondering that with the LGPL license I am letting anybody make
profit out of my cores as long as they leave my name in the author's field... would that permit the obscene situation in which an enterprise sells my core to other enterprise without even making any changes to it? Or am I wrong? I mean, acting as a devil's advocate, what if I take all the LGPL cores at OpenCores and sell them thru Xilinx's coregen program? I do know I can get them for free at opencores.org... but who else does? I'd like to know anybody's thoughts about this. Personally, I don't mind about other people making money by integrating my opencore in a bigger system (JPEG encoder core as a part of a digital camera, for instance), but I'd like to know who is using it, just for the sake of it. What license would make sense for that? Regards, VÃctor López |
|||
|
How to sell an open core?
by marcus.erlandsson on May 9, 2008 |
marcus.erlandsson
Posts: 38 Joined: Nov 22, 2007 Last seen: Mar 7, 2013 |
||
|
Providing OpenSource IP (OSIP) is necessary in order to take the next
step up in design methodology, we canÂ’t re-invent the wheel for every
design.
Another beauty with OSIP is that if the core is frequently used, then
this core is more verified then commercial cores, since its being
tested by so many different engineers world-wide in different ways.
The benefits from a developer standpoint are to get help with the
extensive work in verifying the core. We all know that the design
proportion of an IP is 10-20% and the rest is simulation/verification.
Another benefit is the credit which is given to the developer by the
header-text in all the design files. To have a track record working
towards OpenCores is very good when looking for job opportunities for
commercial companies.
So, my comments to some of your questions are:
Regardless who uses your OSIP, your header-text is always included,
providing credit too the engineer, presenting the license form and
that itÂ’s from OpenCores.
To make commercial products in the end is vital for the OpenSource
growth, basically there are allot of commercial companies that put
allot of time/resources in verification which brings the OCIP up to
another level. It also creates a bigger demand for commercial support
which can benefit the developing engineer.
Regards,
Marcus Erlandsson
----- Original Message -----
From: VÃctor Lópezvictor.lopez at o...>
To:
Date: Thu May 8 20:40:55 CEST 2008
Subject: [oc] How to sell an open core?
I was wondering that with the LGPL license I am letting anybody
make profit out of my cores as long as they leave my name in the author's field... would that permit the obscene situation in which an enterprise sells my core to other enterprise without even making any changes to it? Or am I wrong? I mean, acting as a devil's advocate, what if I take all the LGPL cores at OpenCores and sell them thru Xilinx's coregen program? I do know I can get them for free at opencores.org... but who else does? I'd like to know anybody's thoughts about this. Personally, I don't mind about other people making money by integrating my opencore in a bigger system (JPEG encoder core as a part of a digital camera, for instance), but I'd like to know who is using it, just for the sake of it. What license would make sense for that? Regards, VÃctor López |
|||
|
How to sell an open core?
by Unknown on May 9, 2008 |
Not available! | ||
|
VÃctor López wrote:
I was wondering that with the LGPL license I am letting anybody make
profit out of my cores as long as they leave my name in the author's field... would that permit the obscene situation in which an enterprise sells my core to other enterprise without even making any changes to it? Or am I wrong? I mean, acting as a devil's advocate, what if I take all the LGPL cores at OpenCores and sell them thru Xilinx's coregen program? I do know I can get them for free at opencores.org... but who else does? I'd like to know anybody's thoughts about this. Personally, I don't mind about other people making money by integrating my opencore in a bigger system (JPEG encoder core as a part of a digital camera, for instance), but I'd like to know who is using it, just for the sake of it. What license would make sense for that? Regards, I am not a lawyer, but the way I understand GPL is that there are two core requirements: - if you extend a design that is covered by GPL, you need to set the new design also under GPL. - if you pass a GPL-based design to someone else, you need to provide the source code as well. For LGPL there is a separation between libraries and application code, but I don't really know how that applies to HDL design, as here you cannot really create a library in the same sense as with software design. But maybe I am wrong here and e.g. a VHDL library can be seen that way? So to come back to those two core issues, they will allow you to sell a GPL design, but you need to provide the source code to the Customer as well. If I understand it right, that does not hinder you to make a special contract with that Customer, forbidding him to pass that source code on to any 3rd party. Cheers, Guenter |
|||
|
How to sell an open core?
by Unknown on May 9, 2008 |
Not available! | ||
|
marcus.erlandsson at opencores.org wrote:
Providing OpenSource IP (OSIP) is necessary in order to take the next
step up in design methodology, we can’t re-invent the wheel for every design. Another beauty with OSIP is that if the core is frequently used, then this core is more verified then commercial cores, since its being tested by so many different engineers world-wide in different ways. I fully agree. But opencores.org lacks a feed back here. At the moment it is a one way street. There are developers that put up code on the web page. Then there are users that use the code. The first think I don't like is, that already the developer can not get any feed back from the download form. What I mean is, if someone wants to download the code, s/he has to fill out a form. The developer does not get any feed back from that. The next lack is that there is no encouragement for a user to provide feed back. It would be great if a user could provide feed back about how simple or complicated it was to implement a core. Then the developer could get some ranking about how good his/her cores are. In addition there could be a statistic how often a core got implemented in what technology.
The benefits from a developer standpoint are to get help with the
Now, I think that opencores should also allow some more distinction
here. You are saying that an implemented core improves the core through
feed back by the user. But that feed back is not encouraged. Some people
may even not give it, as they don't want to show that they are using
that core.
On the other side, a developer on opencores is considered a developer of
a core. That is somewhat natural, as that is at the end all a user
wants. But actually, what a user wants is a verified core. So why not
also have the possibility to add a verification project?
For example, Teal/Truss http://trusster.com is a C++ based verification
framework and the developers used the opencores UART in one of the
examples. It would be great if they could have set up a new project for
their code and add a link to the UART project.
Then a user going to the UART project page, would see that someone else
did some verification effort to that core, not related to the developer.
I mean part of the verification problem is, that the developer tends to
be blind for own development errors, so having someone else do the
verification will increase quality.
Just some thought and sorry for the long post.
Cheers,
Guenter
extensive work in verifying the core. We all know that the design proportion of an IP is 10-20% and the rest is simulation/verification. |
|||
|
How to sell an open core?
by webmaster on May 12, 2008 |
webmaster
Posts: 4 Joined: Sep 12, 2011 Last seen: Oct 31, 2026 |
||
|
Hi Günter,
We appreciate your input and we completely agree with your
statements. ItÂ’s good that this topic has come up, so that we can start
a discussion about improvements.
We encourage more people to let us know what functions/improvements
that you would like to see on OpenCores and try to prioritize them.
We have many ideas/plans on how to make OpenCores better, both
from a developer standpoint and also from an “IP-users” perspective.
It’s important that there is a “win-win” partnership between the
developers and users.
We also would like to invite OpenCores users with web-development
expertise to help us improve OpenCores website. Send an email to
webmaster at opencores.org with a short description of your expertise if
you want to contribute.
Best regards,
The OC-team
----- Original Message -----
From: Günter Dannoritzerdannoritzer at w...>
To:
Date: Fri May 9 12:31:02 CEST 2008
Subject: [oc] How to sell an open core?
marcus.erlandsson at opencores.org wrote:
> Providing OpenSource IP (OSIP) is necessary in order to take
the next
> step up in design methodology, we canÂ’t re-invent the
wheel for every
> design.
> Another beauty with OSIP is that if the core is frequently used, then
> this core is more verified then commercial cores, since its
being
> tested by so many different engineers world-wide in different
ways. I fully agree. But opencores.org lacks a feed back here. At the moment it is a one way street. There are developers that put up code on the web page. Then there are users that use the code. The first think I don't like is, that already the developer can not get any feed back from the download form. What I mean is, if someone wants to download the code, s/he has to fill out a form. The developer does not get any feed back from that. The next lack is that there is no encouragement for a user to provide feed back. It would be great if a user could provide feed back about how simple or complicated it was to implement a core. Then the developer could get some ranking about how good his/her cores are. In addition there could be a statistic how often a core got implemented in what technology.
> The benefits from a developer standpoint are to get help with
the
> extensive work in verifying the core. We all know that the
design
> proportion of an IP is 10-20% and the rest is
simulation/verification.
Now, I think that opencores should also allow some more distinction
here. You are saying that an implemented core improves the core
through
feed back by the user. But that feed back is not encouraged. Some
people
may even not give it, as they don't want to show that they are
using
that core.
On the other side, a developer on opencores is considered a
developer of
a core. That is somewhat natural, as that is at the end all a user
wants. But actually, what a user wants is a verified core. So why
not
also have the possibility to add a verification project?
For example, Teal/Truss http://trusster.com is a C++ based
verification
framework and the developers used the opencores UART in one of the
examples. It would be great if they could have set up a new project
for
their code and add a link to the UART project.
Then a user going to the UART project page, would see that someone
else
did some verification effort to that core, not related to the
developer.
I mean part of the verification problem is, that the developer
tends to
be blind for own development errors, so having someone else do the
verification will increase quality.
Just some thought and sorry for the long post.
Cheers,
Guenter
|
|||
|
How to sell an open core?
by Unknown on May 14, 2008 |
Not available! | ||
|
webmaster at opencores.org wrote:
[...]
We also would like to invite OpenCores users with web-development
expertise to help us improve OpenCores website. Send an email to
webmaster at opencores.org with a short description of your expertise if
you want to contribute.
Would it make sense to create actually a project on OC for that? One idea would be to implement some kind of formal improvement development cycle. First step would be to specify what new feature should be implemented. In the next step a specification is created that provides details about the interfaces to the current OC web page source code that is needed to implement the feature. The advantage would be that in the first stage nearly everybody could contribute. In the second stage people with insight to the current OC web source would need to provide input about interfaces. If that specification is published under an OC project, anyone can read it and determine whether they want to contribute to the task. Having done some development myself I know how complex such a task can become. Especially if it has not been specified thoroughly up front. By working out first a specification and publishing it, it may help to show the needed effort and also encourage some more developers to contribute. Cheers, Guenter |
|||
1/1

