[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Please don't remove libc5 - old non-free software might need it!

Here's an idea: somehow setup a way to track packages proposed for
removal, and track downloads of said package for a period of, say, 6
months.  If demand is low during said six months, it gets removed.  Kind
of like Popularity contest (perhaps even inserting a Depends: on it...),
but either server side or a popup telling the user the importance of
registering the package due to the imminent death of it.  This might also
work in the case of entire distribution sets like non-free, so we don't
have people misrepresenting the wishes of the Debian user base.  I see the
protocol as: interested party brings the possibility of deletion up,
maintainer or other responsible party discusses for a short time, package
goes into "tracking mode", once tracking expires (a time period constant
throughout Debian), the decision is made with a reasonable certainty that
the user base concurs. 

On Fri, 14 Jul 2000, Ben Collins wrote:

> > > Your fear is based on the FUD that removing it from woody makes it
> > > completely unavailable for users, and mysteriously removes it from existing
> > > systems. 
> > 
> > ? This is a strawman. I can't think of anyone who believes this. 
> If the software is still availble, then there is no argument, and also no
> reason to leave it in. libc5 will not change. No new archs can use it (or
> would even want to, since their existence is post-libc5, so they don't need
> the backward compatibility). Why keep it around, when it is obsolete, and
> way past it's required backward compatible timeframe?
> IOW, how would removing it from woody actually hurt anything? And don't
> argue about why it wouldn't hurt to keep it, because even libc5-alt package
> maintainers want it removed. Keeping in woody the small group of
> "bleeding-edge, legacy application using" folks don't have to worry about
> it not showing up in dselect's obsolete list (which doesn't hurt anything),
> is completely useless to anyone.
> > Also, from the Jargon File "FUD /fuhd/ n. ... has become generalized to
> > refer to any kind of disinformation used as a _competitive_ weapon." 
> > (Emphasis mine) Unless you're claiming that Harald Schmid is the secret
> > tool of a competetor (and even that would be streaching the original
> > definition), it's not correct to call it FUD.
> > 
> > > Even upgrading to post-potato distributions will not cause libc5 to be
> > > uninstalled, or your libc5 applications to be broken.
> > 
> > Are you promsing that? If so, then why not leave it in? If it's
> > going to take too much maintainance to leave it in, then it's probably going 
> > to break without that maintaince. i.e. upgrading to post-potato distributions
> > is probably going to cause you libc5 applications to be broken.
> Again, fear, uncertainty, and doubt. Regardless of the intention, your fear
> is not based on any sound reasoning, and thus, has no value in this
> discussion.
> Ben

There is an old saying that if a million monkeys typed on a million 
keyboards for a million years, eventually all the works of Shakespeare
would be produced.   Now, thanks to Usenet, we know this is not true.

Who is John Galt?  galt@inconnu.isu.edu, that's who!

Reply to: