[Bf-committers] exit in blender

Matt Ebb matt at mke3.net
Tue Apr 22 13:20:00 CEST 2008


On Tue, Apr 22, 2008 at 7:06 PM, Ton Roosendaal <ton at blender.org> wrote:

>  So! If users accidentally lose work, let's tackle that. But don't play
>  the popularity card or the standards card. These are hollow and
>  manipulative arguments, only leading to mediocre solutions.



They're not hollow and manipulative arguments at all, often times they
are practical ones. In a world where humans are rational creatures,
who read and memorise manuals before sitting down to evaluate an
application, who are perfectly capable of completely shifting mental
gears to different idiosyncrasies  every time they alt tab to a new
application, perhaps this would be true. However humans aren't like
that. They have habits, they build up experience and expectations that
some things to work in certain ways, and they forget things.

There's a point where the above argument makes sense - it's whether
the benefits of doing something non-standard exceed the costs. If for
example there's a way to do things unconventionally, that saves a
significant amount of time or makes things significantly easier for
people, *and* this exceeds the negative aspects of people not
understanding it, or people having to re-learn things, or the
confusion caused by mixing habits between applications, then it's a
good idea.

In the case of quitting, what are the benefits of the current setup?
Well, it saves a fraction of a second of time (the time it takes to
press the shortcut for "don't save") for developers who quit Blender
often. For most users, at least those like myself, I'll open Blender,
work in it for a few hours, then close it. So there's not really any
groundbreaking benefit there. Out of all the workflow problems,
annoying popups and confirmations in the wrong place in Blender, this
is hardly making people's lives much easier.

And what are the costs? Well, potential lost work. Which is one of the
worst possible costs you can have. And for most people that work is
indeed lost - the vast majority of people don't know about the
quit.blend (which is made quite apparent by the number of people who
request this all the time), so for them, for all intents and purposes
the work is gone. Not only is this a very frustrating experience, but
it's costly. If you're trying to start to use Blender in a commercial
setting, and you lose an hour's work, you've just lost money. Weighing
the costs vs benefits, this doesn't look too good...

The attitude that "we've got it right, it's the users' fault who need
to learn properly" seems almost arrogant and cultish. It's almost like
a desire to be different *for the sake of being different* that would
drive all of these complex and extremely difficult ideas to solve this
like trying to teach everyone who picks up Blender about this feature
that to most people seems obscure. This is just as bad, if not worse
than being standard *for the sake of being standard* since at least
most people are accustomed to the standard and won't lose work!

I think this is all summed up pretty well in a recent chat I had with
my boss, who's gradually learning Blender as the rest of our studio
is. He quit Blender, expecting it to ask to save changes, and angrily
lost a fair bit of work (commercial work, on paid time). Luckily I was
there, I then went over to his desk, and for a few minutes went
through the whole spiel, explaining about this quit.blend, the 3
different places you might find it on Windows, how it doesn't work
sometimes if you're running multiple blender sessions, explaining
about how Blender saves relative vs absolute paths and how using the
recover last function will mess up relative paths, even in ways you
don't immediately notice if they're in textures, and how it's best to
copy the quit.blend into your working folder, etc, etc, etc. To which
he said "that's nice, but why don't they just add the *@&$%#^*
confirmation!?"

Which is now what I think about the issue.

Matt


More information about the Bf-committers mailing list