Comments on Insufficiently Random: The tragedy of Eclipse.org

Thanks for the direct link. I appreciate it. I d...

2010-03-23T17:21:24.000-07:00

Thanks for the direct link. I appreciate it. I don't use eclipse so that's easiest.

Comment 9 by Bill Robertson:
> I was jus...

2010-03-22T15:08:20.000-07:00
Comment 9 by Bill Robertson:
> I was just looking for the latest jgit jar file last night, and I could
> not figure out where to download it from. Is it possible to just
> download it or do I have to go through some weird gyrations with
> eclipse to get it?

You can add an update site in Eclipse, http://download.eclipse.org/egit/updates to download the full EGit/JGit set and install it into the workbench. If you want just the JGit JAR, e.g. for embedding into other applications, the JAR itself is available from the update site too as http://download.eclipse.org/egit/updates/plugins/org.eclipse.jgit_0.7.1.jar.

Interesting post. I was just looking for the late...

2010-03-22T15:02:30.000-07:00
Interesting post. I was just looking for the latest jgit jar file last night, and I could not figure out where to download it from. Is it possible to just download it or do I have to go through some weird gyrations with eclipse to get it?

Comment 5 from Patrick Mézard:
> Being c...

2010-03-21T07:04:51.000-07:00
Comment 5 from Patrick Mézard:
> Being completely ignorant of the “Eclipse Development Process”, could
> you give some examples of what kind of “paperwork” is required?

Read the "Eclipse Development Process" document I linked to above, but in a nutshell the foundation requires documentation proving each change was intended to be contributed to the project. This needs to be documented before the change gets committed to the source code repository *AND* again before the product can ship a release. It also needs to be documented in the source code history, in the header of each file the contributor modified, and in an XML file which is used at the release review.

So, make sure you fill out that contribution data in triplicate.

For most small bug fixes from non-core contributors to a project, the "paperwork" involved can be significantly more than the actual code change. And in most jurisdictions, copyright of the small bug fix might not even be valid.

Comment 6 from Anonymous:
> Dumping a large chunk of IP into an open-source project is a
> Big Deal and its right that the normal legal due diligence process
> is followed?
>
> I understand its tedious (I’ve done it myself) and be strung out
> for weeks but it really is the right thing to do.

I fail to see why its a Big Deal. Maybe its because I've worked for so long around open source. Companies that open source a lot of code don't make a Big Deal out of contributions. They just make the contribution and move on.

I think part of the problem right now is there is too much risk being assumed on the side of the project. If an employee of a company open sources a change that the company doesn't want to be open sourced, the company has no real recourse with the project... they have no business relationship. Their issue must be with that employee who open sourced something that they didn't have the rights to.

In communities like the Linux kernel or the Git C implementation changes get open sourced on a mailing list that goes to thousands of people, and to mailing list archives. Once its public it cannot be taken back. If that has caused damage to the company, it shouldn't be the responsibility of Linus Torvalds or Alan Cox to clean up all of those mailing list archives and user inboxes to "protect the IP". No, the best the company can do is sue the employee for damages, and the kernel team will revert any code from the kernel that may have been applied and rewrite it. But they won't go rewrite their entire history, nor do they need documentation in triplicate saying a change can be included.

So, its pointless. Its lawyers ensuring they have a job in the future to continue to review these documents. They aren't providing any real value into the process.

Comment 7 from Jonas Gorauskas:
> That being said, is hosting your project at the Eclipse.org side
> necessary? Couldn’t you just cut the crap, fork your own stuff,
> and host it somewhere else?

EGit and JGit used to be hosted elsewhere. We moved to Eclipse so we can have EGit bundled as part of the standard distributions that are based on the core Eclipse projects, that way users have GIt support "out of the box", no matter how they obtained Eclipse.

Unfortunately, the only way to get bundled is to host your project at Eclipse.org and run the project according to their rules. So we aren't doing it for us, we're doing it to make things easier for our users by making EGit more widely available.

I'm an eclipse and egit user. The extent of my...

2010-03-20T22:05:38.000-07:00
I'm an eclipse and egit user. The extent of my interaction with the Eclipse foundation is zero and I know where to go download the software and how to make it work for my purposes. So, forgive me if I sound like I don't know what I'm talking about. I most likely don't... That being said, is hosting your project at the Eclipse.org side necessary? Couldn't you just cut the crap, fork your own stuff, and host it somewhere else? Host it at github, perhaps? Is the Eclipse seal of approval a prestige thing? I just want to understand why you have to put up with that crap when there may be alternatives.

The Eclipse Development Process is fine for normal...

2010-03-20T21:38:52.000-07:00
The Eclipse Development Process is fine for normal committers/contributers

Dumping a large chunk of IP into an open-source project is a Big Deal and its right that the normal legal due diligence process is followed?

I understand its tedious (I've done it myself) and be strung out for weeks but it really is the right thing to do

Being completely ignorant of the "Eclipse Dev...

2010-03-20T02:33:19.000-07:00
Being completely ignorant of the "Eclipse Development Process", could you give some examples of what kind of "paperwork" is required?

It is useful to try everything in practice anyway ...

2010-03-10T11:30:23.000-08:00
It is useful to try everything in practice anyway and I like that here it's always possible to find something new. :)

[...] Another day. Another compliant from me about...

2010-02-10T03:47:53.000-08:00
[...] Another day. Another compliant from me about running a project at Eclipse.org. This time it wound up in the jgit-dev mailing list archives, as replies to a thread that I think started from my blog post on the tragedy of Eclipse. [...]

If only all lawyers that work in tech were engilay...

2010-02-09T04:38:09.000-08:00
If only all lawyers that work in tech were engilaywers with a decent grasp of how things work. Then it could be _their_ responsibility to provide everybody else with a decent system (such as that XSL formatting and the script to mine the data they need). Bureaucratics need to not get in the way of software development :(.

Yup, the iplog is a PDF file holding XML. The fou...

2010-02-09T02:40:07.000-08:00
Yup, the iplog is a PDF file holding XML. The foundation is trying to move from an HTML based iplog to an XML based format. So we generated our iplog in XML, using code I wrote in JGit to mine the Git revision history. Someone told the legal team at the Eclipse Foundation that you can't edit a PDF, so they "freeze" the iplog by putting its contents in PDF. But they don't have an XSL to format it in human readable text, so we get this instead.

Apparently someone needs to take some time to write an XSL:FO stylesheet for the iplog XSD so we can render it through Apache FOP to get the magically immutable PDF in human readable format.

Wow, that sounds pretty horrible indeed. It's ...

2010-02-08T17:02:00.000-08:00
Wow, that sounds pretty horrible indeed. It's understandable that the eclipse foundation's priorities are with keeping their sponsors happy, but surely there's a way to do that without screwing over non-corporate contributions? Best of luck, hope things miraculously start picking up soon.

Also, the iplog is a... pdf file with xml?

Comments on Insufficiently Random: The tragedy of Eclipse.org

Thanks for the direct link. I appreciate it. I d...

Comment 9 by Bill Robertson:> I was jus...

Interesting post. I was just looking for the late...

Comment 5 from Patrick Mézard:> Being c...

I'm an eclipse and egit user. The extent of my...

The Eclipse Development Process is fine for normal...

Being completely ignorant of the "Eclipse Dev...

It is useful to try everything in practice anyway ...

[...] Another day. Another compliant from me about...

If only all lawyers that work in tech were engilay...

Yup, the iplog is a PDF file holding XML. The fou...

Wow, that sounds pretty horrible indeed. It's ...

Comment 9 by Bill Robertson:
> I was jus...

Comment 5 from Patrick Mézard:
> Being c...