Responsibility for orphaned libraries
The mysterious disappearance of Jonathan Turkanis (not heard from here since last October; home page says he's at UCB, but a directory search at UCB does not find him) set me wondering about the risk of using Boost libraries. Like iostreams, they are generally one-man works, which may be beyond an end user's ability or remit to maintain. Even if the author does not disappear, he may eventually lose interest or simply be too busy to maintain it forever, so do Boost have any procedures in place for such eventualities?
Keith MacDonald wrote:
Even if the author does not disappear, he may eventually lose interest or simply be too busy to maintain it forever, so do Boost have any procedures in place for such eventualities?
I am not absolutely sure about this, but isn't it the case that the sentence: "and to prepare derivative works of the Software" will allow anyone to continue on the work once the original developer is not available anymore, for whatever reason? E.g we are currently trying to rewrite the thread lib to get it under the new license to avoid further stall of maintainance. Roland
On 3/25/06, Keith MacDonald <boost@mailclan.net> wrote:
The mysterious disappearance of Jonathan Turkanis (not heard from here since last October
I was still in contact with him last nov (we were planning to get ready a new version of the rational library for 1.34), and he was very busy working on a book. I can send you his e-mail address in a personal mail if you want to try to get in touch with him. br, Andras
I posted patches for 3 bugs in boost/rational.hpp a month ago at http://sourceforge.net/tracker/?group_id=7586&atid=307586 They've all been assigned to turkanis, but it's unclear if they will ever be applied. Responding quickly to bug reports and patch contributions is THE best way for a free software project to encourage new contributions. Conversely, allowing patches to languish in limbo is a good way to scare off new contributors. On 3/25/06, Keith MacDonald <boost@mailclan.net> wrote:
The mysterious disappearance of Jonathan Turkanis (not heard from here since last October
Andras Erdei wrote:
On 3/25/06, Keith MacDonald <boost@mailclan.net> wrote:
The mysterious disappearance of Jonathan Turkanis (not heard from here since last October
I was still in contact with him last nov (we were planning to get ready a new version of the rational library for 1.34), and he was very busy working on a book. I can send you his e-mail address in a personal mail if you want to try to get in touch with him.
I was last in touch with him on 16th Feb by e-mail. Jim
Sorry, I should have made it clearer that I was only using the lack of response to iostreams questions as an example of what could happen. The real question I was trying to ask was is there a risk of using Boost in an application, perhaps written under contract for a company, and then finding that you have to maintain the library, as well as your application, because the library developer has moved on? To paraphrase my original question, what would Boost do with the library under such circumstances? "Andras Erdei" <aerdei@gmail.com> wrote in message news:7fe28deb0603251024q5af7f47eu76b02a157b99a289@mail.gmail.com...
On 3/25/06, Keith MacDonald <boost@mailclan.net> wrote:
The mysterious disappearance of Jonathan Turkanis (not heard from here since last October
I was still in contact with him last nov (we were planning to get ready a new version of the rational library for 1.34), and he was very busy working on a book. I can send you his e-mail address in a personal mail if you want to try to get in touch with him. Andras
"Keith MacDonald" <boost@mailclan.net> writes:
Sorry, I should have made it clearer that I was only using the lack of response to iostreams questions as an example of what could happen. The real question I was trying to ask was is there a risk of using Boost in an application, perhaps written under contract for a company, and then finding that you have to maintain the library, as well as your application, because the library developer has moved on?
There is always a risk. On the other hand, you run the same risk with libraries from any vendor... except that you may not have the right to make changes to the source of those libraries. That said, enough people have an interest in the viability of Boost that we hope to be able to transfer maintenance responsibility to someone else when a developer moves on.
To paraphrase my original question, what would Boost do with the library under such circumstances?
Most likely we'd continue to ship it and fix any regressions at least. And if it was brought to our attention that the library was languishing, we'd look for a new maintainer. -- Dave Abrahams Boost Consulting www.boost-consulting.com
David Abrahams wrote:
"Keith MacDonald" <boost@mailclan.net> writes:
To paraphrase my original question, what would Boost do with the library under such circumstances?
Most likely we'd continue to ship it and fix any regressions at least. And if it was brought to our attention that the library was languishing, we'd look for a new maintainer.
This is not entirely true, I think. Iostreams is a good example for this. (See http://comments.gmane.org/gmane.comp.lib.boost.devel/139335) There are still merges from 1.33.1 to MAIN pending and several issues and patches have been posted to this list but with mostly no response. IIRC, it took more than a year for the thread library to get a new maintainer team. As for the reasons for this, I have no idea. Probably we would need some process to officially orphan a library so that it's clear that a new maintainer is needed. Markus
Markus Schöpflin <markus.schoepflin@comsoft.de> writes:
As for the reasons for this, I have no idea. Probably we would need some process to officially orphan a library so that it's clear that a new maintainer is needed.
Please make a proposal. -- Dave Abrahams Boost Consulting www.boost-consulting.com
participants (7)
-
Andras Erdei -
David Abrahams -
David Benbennick -
Jim Douglas -
Keith MacDonald -
Markus Schöpflin -
Roland Schwarz