Sure, it's not only the size of the population making use of a feature
that needs to be taken into account, but also the feature's
criticality to that population (and also to some degree, the features
alignment with our values, as you suggest). I don't think image
blocking is as critical to users as our accessibility support, though.
> It didn't and still doesn't make sense to me that we acknowledged that
> there's a real and sensible use case for disabling images (slow connections,
> data plans etc.), which may be essential to how some people access the
> internet, but that this use case would be better supported by a dedicated
> bandwidth saving mode, and that the option to disable images could therefore
> go away -- without the hypothetical replacement being available.
https://addons.mozilla.org/en-us/firefox/addon/image-block/ seems like
a much better option for these users than the checkbox we used to have
in the preferences pane.
Gavin
1) is a pretty subjective argument, so I'm not sure we can
conclusively reach agreement. Re: 2), "It breaks Google" was not the
reason to remove the checkbox - it was just an example of how the
effects of the preference are not useful to the vast majority of
people browsing the Web. The reason we removed the checkbox was that
the preference was deemed not useful enough to merit a spot in the
preference pane.
> For users who unexpectedly have to deal with a borderline-unusably-slow
> connection on the road (fwiw, this actually happened to me a few years ago
> in an airport and my solution was to disable images), searching the internet
> for an add-on is exactly what you don't want to do in such a situation.
If we care to address this use case - I'm not sure that we do - there
are far more discoverable and useful options that we could implement
(low-bandwidth mode, etc.). Of course, I recognize that that's not
useful to the people who had gotten used to using this feature to
address this use case (poorly), but I assert that that loss is not so
great that it outweighs the benefits.
Gavin
It's based on pretty simple ideas we should be able to agree upon:
- some users have to deal with connections that are limited enough that
our default behavior is barely workable
- enabling users to surf the web is a web browser's central objective
> Re: 2), "It breaks Google" was not the
> reason to remove the checkbox - it was just an example of how the
> effects of the preference are not useful to the vast majority of
> people browsing the Web. The reason we removed the checkbox was that
> the preference was deemed not useful enough to merit a spot in the
> preference pane.
That's not how I read http://limi.net/checkboxes-kill nor
https://bugzilla.mozilla.org/show_bug.cgi?id=851701#c0.
>> For users who unexpectedly have to deal with a borderline-unusably-slow
>> connection on the road (fwiw, this actually happened to me a few years ago
>> in an airport and my solution was to disable images), searching the internet
>> for an add-on is exactly what you don't want to do in such a situation.
>
> If we care to address this use case - I'm not sure that we do -
Why not?
Dao
Re-reading that, the blog post (not the bug report) did go on to say
that the checkbox was probably only useful to less than 2% of people (an
argument that I countered elsewhere in this thread), but it started with
the assertion that it was harmful based on the fact that Google was broken.
I don't think this use case is common enough, and I don't think our
checkboxes are useful/discoverable enough to address it properly.
There are plenty of other edge use cases we could enumerate that we
don't currently address, and whose cases would be similar to the one
you make here. The only difference here is that this represents a
removal rather than a proposed addition. There should be a higher
burden for removing something than for not adding it, given the user
impact, but we can't have that burden be so high that we can never
remove things.
> That's not how I read http://limi.net/checkboxes-kill nor
> https://bugzilla.mozilla.org/show_bug.cgi?id=851701#c0.
I'm telling you my reasoning for supporting the change, which might be
different than people's reasons for suggesting it.
Gavin
No one said users are idiots; it's just important to know that they
often have things to worry about beyond taking the time to understand
your software's UI in great depth.
I don't think further discussion on this topic is likely to be
fruitful. I think your position is quite well understood, and we just
fundamentally disagree.
Gavin