Re: real LSB compliance
>>>>> "Manoj" == Manoj Srivastava <srivasta@debian.org> writes:
>>> "Sam" == Sam Hartman <hartmans@debian.org> writes:
Sam> First I suspect that it is only for rc.d links. Secondly,
Sam> why are we supporting LSB if not for our users?
Manoj> If the product is free software, we shall already have
Manoj> decent, conforming packaging for it. Now, if you are
Manoj> talking about non-free software, that can never actually be
Manoj> part od Debian, well, then. How much effort do we really
Manoj> want to spend here? (Were we not already discussing
Manoj> throwing out non-free from our archive sites?)
The assertion that we will have all free software packaged in Debian
is ludicrous; we may have much of it packaged, and we should work to
package any that we do not have , but there may exist LSB free
software not in Debian. Particularly for fast-moving targets it may
be better for our users to use LSB packages than native debs.
More over, when I agreed to be a Debian developer I agreed to follow
the social contract as written while working on Debian projects. That
includes the part of the contract that acknowledges that non-free
software is useful to our users. Until the social contract is
changed, I will conduct myself accordingly and will expect other
developers to do the same. I do agree that sacrificing implementation
quality is a valid tradeoff and that we should not do so to support
non-free software.
Sam> If LSB is actually useful to our users, and I believe that if
Sam> we participate in future revisions of the spec it may well be
Sam> very good for our users, then they will likely want LSB
Sam> packages to fit well into the Debian architecture. Having
Sam> init.d script names mangled and placed in a special directory
Sam> would annoy me as a user; I would rather see Debian
Sam> participate in the registry.
Manoj> While I agree with the sentiments in general, I have a
Manoj> major caveat: that we do not sacrifice quality of
Manoj> implementation in order to facilitate non free software.
Manoj> So any registry would only be a recommendation for
Manoj> developers, anyway, and we should conform, where feasible.
I do not see how requiring people to register usage of a shared
namespace decreases implementation quality; I think it is a good idea
regardless of the LSB. It may be a good idea that has not been high
enough on our priority list so far, but since someone has done the
work of setting up the registry for us, we should use it.
Sam> Also, consider that we do not want to make LSB packages be
Sam> forced into being a bag on the side of Debian or any other
Sam> distribution. If a software author can get sufficiently
Sam> better integration by packaging directly for Redhat than for
Sam> LSB, then they may very well do so. If this happens a lot,
Sam> we will have done our users a disservice by helping to make
Sam> LSB a sufficiently unattractive option that software is not
Sam> available for LSB.
Manoj> I see you are against the removal of non free from
Manoj> Debian. However, the issue is yet undecided (i'll need to
Manoj> look and see of the people devicing the new voting
Manoj> protocols have done so, and see where the vote on that
Manoj> actually goes)
I believe that regardless of whether we have a non-free archive, our
users will want to be able to install third party (both free and
non-free) software on their machines. I believe very much that
providing a high quality operating system is our primary mission and I
believe we will have failed in that mission if we do not provide
facilities that our users want. It would be one thing if the software
needed to implement LSB were non-free, but that is not the case. We
can comply with LSB while still providing a free operating system. As
such, I believe that if LSB ends up being useful to our users and if
there are Debian developers who wish to do the work of implementing
LSB for Debian, then we would be violating our social contract by
standing in the way of those developers.
--Sam
Reply to: