From: Arkadiy Vertleyb (vertleyb_at_[hidden])
Date: 2005-12-07 18:02:45
"Eric Niebler" <eric_at_[hidden]> wrote
>
> Arkadiy Vertleyb wrote:
> > "Eric Niebler" <eric_at_[hidden]> wrote
> >
> >>It doesn't make sense to have download counters at all, really, if they
> >>randomly drop to zero every now and then.
> >
> > Frustrating though it is (I had to drop typeof download counter in the
past
> > from 99 to zero, because I had to re-upload the file initialy uploaded
by
> > Peder), I think it's a very useful tool, even though imperfect,
providing a
> > good indication of amount of interest combined with author(s) activity.
> > FWIW, if not download counter, there would be no BOOST_TYPEOF now.
>
> You misunderstand. My comment was not out of frustration. It's a
> matter-of-fact observation that if the download counters randomly and
> periodically drop to zero, then the numbers have NO meaning. They didn't
> go to zero because someone set them to zero, or because someone deleted
> and/or re-uploaded a file. They just mysteriously went to zero. Knowing
> this can happen, what can you infer about a file by looking at its
> download counter?
I did understand. What I was trying to say is that the error needs to be
corrected at some point, but meanwhile, even as is, the counter provides
useful information _for library authors_. I used it to determine interest
in typeof, which was otherwise undetectable. I can also see some interest
in RTL, and this also is pretty much the only indicator of this interest.
Even if tomorrow the download counter for RTL drops, I can still monitor the
reaction to any particular change, and this is quite useful, IMO.
Regards,
Arkadiy