FreeLists / openbeos / [openbeos] Re: Windows Vista Performance Kludges (that Haiku does not need)

[Stephan Assmus <superstippi@xxxxxx>]

On 2006-12-12 at 20:01:18 [+1300], Mikael Jansson <lists@xxxxxxxxxxxxxxxxx> 
wrote:
> Stephan Assmus skrev:
> > On 2006-12-12 at 07:44:44 [+1300], Ryan Leavengood <leavengood@xxxxxxxxx>
> > wrote:
> >> Well indeed Vista, XP and OS X certainly have more built-in features
> >> than BeOS. I don't think the order of magnitude difference is quite
> >> ten million, but I see your point. I would argue that in a few years
> >> Haiku could add many of these features and still not see the
> >> performance issues that Windows (or for that matter even OS X) sees.
> > 
> > No one knows for sure. If we ever have the resources to add any feature we
> > want, we would need to address the problem by strict philosophy that 
> > "simple
> > is good". In a way, this means to me that we should design things so that 
> > the
> > user is empowered to understand what is going on and by this understanding
> > make the computer do what he wants and compensate "missing features".
> >
> There is, however, a problem with that thinking. See a recent
> JoelOnSoftwareArticle:
> 
>      http://joelonsoftware.com/items/2006/12/09.html
> 
> Sure, simple is good. But "which" simple?

Well, I read that article, but I also read the previous article in which he 
*simplified* 14 ways to shutdown Windows into a single option. Both articles 
seem to contradict each other. Suppose you are a company selling a software 
that shuts down Windows. So the absolutely best way to make more revenue, and 
also do what your users want most, is to release new versions with ever more 
features to shutdown the computer? Do I just follow the logic of his latest 
article? Joel would probably disaprove of that. The truth is of course that 
"it depends". And this is exactly what I really don't like about Joel's 
latest article - it takes a subject so complex and barely scratches the 
surface. He doesn't even give any hard prove to back up his assertions. So 
his revenue undeniably increases whenever he releases a new version? Where is 
the prove that this has to do with the new features? Maybe it does, maybe the 
new version has features specifically requested by many of his users? A 
different story already.

Staying on the subject of "shutting down the computer", I'll try to give you 
an example of the "simple but powerful building blocks" concept: Look at the 
screenshots I attached (sorry in German, but you'll get the idea). They show 
such a shutdown application for Windows, which scared the crap out of me when 
I first saw it. It has a ton of options how exactly to shutdown your (or even 
a remote) computer. I've also displayed the timing options, because that is 
mostly what my girlfriend is using it for: "shutdown in half an hour".

Now suppose you had a "timer" application for BeOS (indeed there are some 
like that) which would let you execute any program at a certain time. 
Together with the "shutdown" command (or some graphical analog), this would 
enable you to combine these two tools to do what PowerOff for Windows does. 
But of course, the timer tool is still available to be used in combination 
with all sorts of other programs! So it is actually a more powerful solution. 
And the user would have to learn the timer program only once. Of course, you 
can't just throw these things together, you have to think about how they will 
be used, and you might need special features to make it easy on the user to 
build persistent workflows and anticipate some workflows that many users will 
likely want (inbuilt templates, like the default Tracker queries for email). 
But in the end, I believe this is the right way to go and one should look for 
a solution in this direction always first before settling with something else.

Best regards,
-Stephan

Attachment: PowerOffShot-2.png
Description: PNG image

Attachment: PowerOffShot-1.png
Description: PNG image