Re: Forge Software Evaluation by FSF

[Han-Wen Nienhuys]

On Wed, Feb 26, 2020 at 11:50 PM David Kastrup <address@hidden> wrote:
> > It's also useful for tracking downloads.
>
> But for that it is not necessary to know the IP addresses.  And we
> likely don't.

We certainly do. They are stored in the Apache server logs on the
lilypond.org site.

>  But employing Google's services here means that they do.
> And Google did retire its "Do no evil" motto a few years ago without
> much of a fanfare.
>
> Now in connection with us still using Rietveld and having used Google
> Code, this looks like at best a minor problem.  On the other hand, they
> tend to work a whole lot less for profiling a wide range of people than
> our main website would.
>
> If we contribute to tracking down and jailing Chinese users who looked
> at LilyPond for creating protest songs, that would be nothing to be
> particularly proud of.

If we decide that this is a genuine concern of ours, then we can make
changes here. However, rather than looking at the FSF's guidelines for
Forge sites, we should look at our own project, decide what *we* want,
and then take actions to address our concerns.

For example, if we decide that we are concerned with Chinese
government, we could stop serving unencrypted http traffic.

> > We recently discussed the PDF manual, and if it makes sense to add the
> > download size. If we'd annotate the link, we could easily see how
> > often people download the manual, and whether it's worth time
> > tinkering with it.
>
> We don't offer downloads just for the PDF notation manual.  And we don't
> see how an advertised download size influences download decisions.  If
> we want to maximise downloads, we should probably remove all size
> indications and have users hit their limits.  I think of the objective
> more to provide a courtesy rather than maximize web traffic.

I mean: I think we should balance the effort we invest (eg. putting
size on the download page) against the number of people this benefits.
Having easy to access statistics helps inform this tradeoff.

-- 
Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen