Message: previous - next
Month: January 2012

Re: Re: Re: [trinity-devel] Centered or "Smart" Default Window Placement

From: Martin Gräßlin <mgraesslin@...>
Date: Sun, 15 Jan 2012 23:32:14 +0100
On Sunday 15 January 2012 13:06:09 Darrell Anderson wrote:
> > ok. I cannot give you usability studies, because they don't
> > exist. But you can
> > try it yourself. Just use it! Just try this setting for a
> > day or two. Give it
> > to a non-experienced user.
> I have tried Smart placement. I don't like the placement scheme. :)
so please take a step back and remove your personal opinion. Finding defaults 
is not about what you consider the best option for you but for what is the 
best option for the users. Think about the fact that I argue here for an 
option in a software I don't use. In fact about a software which I could 
consider as a hostile fork of the software I develop.

Personally I have written quite some code which is used as default but which I 
don't use. Quite strange, isn't it? I and other developers are able to step 
outside and think what is good for users to be the default and tailor the 
system in a way that it works good for users who don't want to touch any 

Nevertheless we offer the possibility to change the sane defaults to something 
else for advanced users like yourself so that you can use e.g. a centered 
placement strategy.
> I can say with confidence that quite often what is "cool" to a "geek" is not
> cool to every day users. I have observed this often.
I quite agree. That's why centered is *not* a useful placement strategy. It 
requires users to move their windows to have it useable. This is the worst 
thing which could happen.

Users should not notice that there is something that places windows. But with 
centered they notice, because the system does not act in the way it should 
> That you offer no usability studies related to this specific topic means we
> are limited to opinions. I can live with that.
No we aren't limited to opinions. I can give you the experience of several 
years of developing and maintaining a window manager. I know quite a bit about 
window management and what are good options. You mentioned you watch people 
using their computer. I watch people interacting with the window manager. I 
don't look at anything else except the window management.

Furthermore I have the complete confidence of our bug tracking software. We 
have millions of users reporting hundred of bug reports each year to multiple 
parts of our window manager. But reports to the placement strategy? Feature 
requests? Hardly any. I am involved in multiple KDE related support areas 
watching for window manager related topics: nothing about placement.

I am completely sure that our users are well suited with placment strategy 
smart as the default. Anything else will harm your userbase.

You can think it's just opinion or accept it as the experience of a younger 
developer who is expert in this area. You know you can count the world wide 
maintainers of serious window managers on one hand. There's:
* MacOS X
* Microsoft Windows
* Compiz
* Mutter
* KWin
> > Of course it has to do with technical aspects of the window
> > manager. You can
> > only choose a default if you understand it. If you
> > understand every detail of
> > it. If you know and understand the underlying code and the
> > history behind it.
> The original poll question has nothing to do with the technical aspects of
> each placement option. The poll question is about usability, of which you
> provide no related studies.
The technical limitations have to be considered when changing defaults. This 
is as important or maybe even more important than useability.
> As I mentioned in the bugzilla, I can see from the code (I can read C++ but
> am not a C++ hacker --- I'm teaching myself as we speak) that a lot of
> sweat equity went into the thought and design of Smart placement. Having
> hacked lots of code myself, I appreciate the effort behind the work.
It's not about the complexity of smart, but about the not-developed placment 
strategy centered.
> With that said, for you to tell me I can choose a default only if I
> understand the technical aspects is condescending. I need only to use the
> options to decide what I like or don't like.
no, sorry, this is short sighted. What if an option has not been designed to 
be used as default. If it has known limitiations or even bugs which are just 
not important enough to be fixed. Developers concentrate on the default 
options to make them really good. You would expose untested code to your 
users. And you have nobody to fix it (compare my initiative to use KWin as 
> I never argued that Smart placement should be removed. More than once I
> emphasized that this discussion is not about the merits of each option, but
> only about which option should be the default for the initial installation.
> Users can change the option after the initial installation, but what should
> the default be? As you offer no usability studies about the topic, then as
> I said, we all are limited to opinions only. Fair enough. Just don't "talk
> down" to people about their choices.
see above
> > If you just had looked into the code, if you had tracked
> > down why it is the
> > default, you would not even have come up with the
> > proposal.
> Condescending opinion.
> > Just to give you an idea. The placement policy centered was
> > not always called
> > centered. It was committed as "StupidlyCentered". It did
> > not even had a GUI
> > option, because it is so stupid. It is from 2002 and has
> > not changed at all,
> > the code is still the same (except an internal
> > adjustment).
> I never have been impressed with these types of developers' attitudes.
> I don't care that the Centered option has remain unchanged since 2002. I
> have tools, appliances, and furniture in my house that are --- I'm guessing
> here --- older than you. My pickup truck is 23 years old and runs like a
> top. Does that make any of those objects "stupid"?
sorry this completely misses the point. What the name (and commit) tells us 
* the developer who wrote the code, considers the functionality as "stupid"
* he himself never touched the code again and improved it
* nobody else ever considered to improve the code
* looking at the commit where it was introduced it becomes clearly visible 
that it is example code for the now configurable placement strategy

=> centered has not been developed to be used!
> Tim and I started this discussion in the bugzilla. As much as I want
> Centered, er, StupidlyCentered, I am not so arrogant to think others want
> that option as the default. I wrote a patch to change the default.
I just looked at the patch, I'm sorry to tell you, but it is completely wrong.
> So I
> have sweat equity involved in this discussion. Perhaps a little pride too
> because I am not a full fledged C++ coder.
> I might get out-voted and Smart will remain the default option for new
> installations. Therefore in the bugzilla I proposed an alternate method to
> change the global default for new installations without patching code.
You know that this existed for years? You know that this is the way that 
distributions use to configure their default settings?
> That
> is one of the attractive elements I like about the Trinity project. We
> don't see things here as "my way or the highway" or "either or." We don't
> pretend to "know better." We see all options as being viable and we try to
> work with everybody.
sorry, what a bullsh*** I cannot tell you how arrogant I find this statement. 
It's this complete utter non-sense I have seen here more than once on this 
list and elsewhere with the opinion that KDE developers are all assholes not 
caring about their users. It's such a stupid thing to think that KDE 
developers pretend to "know better". Such nonsense makes me really angry and I 
really have to think about whether I want to continue to offer my help to the 
Trinity project.
> There might be some usability reasons, but as you have not provided any
> related studies,
There are no studies because studies are very expensive. I would love to have 
studies on everything. If you have a spare $100000 I happily accept it to do 
usability studies at the local university for every aspect of the desktop.
> I am left to think the underlying reason is ego and what
> some geeks thought was cool.
*sigh* Ego is never a reason why something is the default. And yes if you 
write that you consider ego as a reason for defaults, I consider this as an 
insult (I have not been involved when the default for this option has been 
set) and are convinced that you consider KDE developers to be stupid as you 
quite nicely highlighted with various side notes in your mail.

I would appreciate if you could use a constructive manner to discuss and I 
think you should appoligize towards the KDE developers you insulted in your 
mail, e.g. the quite nice Kate developers who I know personally.

Kind regards by a rather annoyed KDE developer given the mail he had to read 
before going to bed.

Martin Gr��lin