Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [gmt-dev] VIATRA2 is not operating as an Eclipse open source project

Dear Wayne and All

All previous "releases" of VIATRA were IP clean - at least this
aspect was verified by the Eclipse IP team. For all contributors of the VIATRA project,
we have faxed a corresponding Eclipse Consent signed by the employers to the Eclipse Foundation.

My estimate of 15+ developers included those people who contributed to
experimental features which have not yet reached a certain level of maturity to
get submitted as a contribution to Eclipse.

As for the four developers mentioned in CQ 6703, their share of contribution only relates to
the delta, that is, the novel contributions since the last "release". There are 12 contributors
to the total code base - as you list below.


From:        Wayne Beaton <wayne@xxxxxxxxxxx>
To:        varro@xxxxxxxxxx
Cc:        gmt-dev@xxxxxxxxxxx, "'PMC members mailing list'" <modeling-pmc@xxxxxxxxxxx>
Date:        2012.08.02 21:26
Subject:        Re: VIATRA2 is not operating as an Eclipse open source project

Eclipse projects must follow the Eclipse Development Process. Operating in a "single committer mode" is no excuse. The processes exist for a reason, primarily one of protecting the developers, the community, and adopters from IP exposures.

Eclipse projects are responsible for maintaining a log of all contributions. How are the contributions of 15+ developers accounted for? All of the project commits are attributed to you, and there doesn't seem to be 15 developer's worth of contributions captured in Bugzilla. Based on the three contribution CQs, I have the following list of contributors:

Daniel Varro (GMT committer)
Istvan Rath
Andras Balogh,
Gergely Varro,
Akos Horvath,
Peter Pasztor,
Gabor Bergmann,
Andras Okros,
Andras Schmidt,
David Vago
Abel Hegedus
Zoltan Ujhelyi

Are there other contributors? Were all of these contributions accepted in conformance with the IP Due Diligence process? [1]

The contribution attached to CQ 6703 represents, AFAIK, the entire code base of VIATRA2. It lists four developers. If it's a revision of the previoius contributions that includes all of the code, shouldn't it list all of people who contributed to the previous work? Should it list 15+ developers?

This is a potentially huge IP exposure. Who are these contributors? What did they contribute? Do they have the rights to contribute? Do their employers consent to the IP being distributed under the EPL?

This needs to be sorted out immediately.

Are you in a position to bring the project into conformance with the Eclipse Development Process?



On 08/02/2012 03:02 PM,
varro@xxxxxxxxxx wrote:
Dear Wayne and All,

Let me give some further information about the background of this issue.

The VIATRA2 project has been an incubation phase as part of the GMT Eclipse subproject since its start in 2005/2006.
It started as a purely academic initiative, and slowly evolved into something more mature.

Since GMT was an umbrella project for many different initiatives related to modeling and code generation,
a single committer account was granted for me - as VIATRA project lead. With 6-8 other projects,
GMT already had quite a high number of committers in the very beginning.

After the birth of the Modeling project, several subprojects of GMT (e.g ATL, oAW) gradually
moved to Modeling. VIATRA stayed in GMT up to now - with a single committer - despite
the fact that 15+ people contributed it during the last couple of years.

Due to this "single committer mode", it was impossible for us to properly
follow the Eclipse processes as it caused unrealistic overhead. Furthermore, GMT
has somehow become a "forgotten" project in the last couple of years, so no one was really there to
help us avoid policy violations.

(In fact, our ad hoc release numbers date back to a time when the current Eclipse policy
was not yet in effect - and several other former GMT projects were renumbered)

Of course, I would be grateful for any kind of assistance or mentorship we can get
to sort out non-conformance issues. But it would be also great to get  5-6 additional
committer accounts first for the regular VIATRA contributors to avoid my virtual "one-man-show".

Best regards,

From:         Wayne Beaton
gmt-dev@xxxxxxxxxxx, "'PMC members mailing list'" <modeling-pmc@xxxxxxxxxxx>
Date:         2012.08.02 20:03
Subject:         VIATRA2 is not operating as an Eclipse open source project

Greetings GMT project and Modeling PMC.

Daniel, if you have not already joined the GMT dev mailing list, please do so.

I am concerned that the VIATRA2 project is not operating as an Eclipse open source project

CQ 6703 was submitted today. That CQ is for a code contribution. This appears to be a significant contribution of code. It is the first contribution in more than two years to the project. In comments on that CQ, the project lead indicates that the contribution is "the newest combined release of the VIATRA2 model transformation framework" and that "In each iteration, we compile the contributions from the developers and push the code to"

This runs counter the open source rules of engagement set out in the Eclipse Development Process [1]. Specifically, the project is not operating in an open and transparent manner leaving little opportunity for others to join the project. The project has been established for some time, and should be operating fully at

There are several developers listed in the contribution. Based on the amount of contribution, I believe that each of these developers should become committers and do all future work directly in the code repository.

The project website [2] indicates that there have been numerous releases of the project, including statements like "The VIATRA2 Release 3.1 has been approved by!". AFAICT, the project has not undergone a single release review as required by the Eclipse Development Process. The project must plan to undergo a formal release as soon as possible.

Further, the 3.2 release indicates a third-party update site only on the downloads page [3]. It is okay to provide alternative download sites, but they must be *alternatives* to; the downloads for the project must be obtainable from A third party site cannot be the "official" site.

The project is--according to my records--in incubation. Incubation branding must be provided on the download page and in downloads. I have marked the project as "incubation-non-conforming"; the project cannot benefit from parallel IP until the incubation branding issues is resolved.

I believe that the project would benefit greatly from Architecture Council mentorship (I'm not sure why mentors weren't assigned in the first place). Can the modeling project suggest somebody who might be appropriate in this role?



Wayne Beaton
The Eclipse Foundation
Twitter: @waynebeaton
Eclipse Projects

Wayne Beaton
The Eclipse Foundation
Twitter: @waynebeaton
Eclipse Projects

Back to the top