Re: Changes I've been thinking of...

[Jamie Ramone]

On Sat, Oct 10, 2009 at 6:27 AM, Michael Thaler
<michael.thaler@physik.tu-muenchen.de> wrote:
> Hi,
>
>> The gray and dull has nothing to do with it but, rather, the lack of
>> glitter. I happen to work for a big company and while we use PCs with
>> Windows XP (using it's default theme), the tools made specifically for
>> us use the old Win NT look from the mid 90's. In fact one monitoring
>> tool remotely running on some UNIX system and it's made with Motif.
>> I've never seen any professional use glittery interfaces, they go for
>> more neutral looking ones.
>
> I work for a company that develops software for the German Space Operation
> Center. They use Linux with KDE and Windows XP in all their control rooms.
>
> Why should something like Expose, window shadows or animations or transparancy
> make a user interface less useable? It can certainly make it less useable if
> overused, but it can also make it more useable when used at the right places.
>
>> No argument there, everyone has different taste. So this is a case, in
>> my own opinion, to favour skins in GNUstep as done via Chameleon for
>> instance.
>
> I doubt that this is a solution. I doubt it is possible to make KDE or GNOME
> look like Snow Leopard. There was the Baghira theme which did quite some
> hackish things but even with Baghira the look and feel of KDE was not really
> similar to MacOS X. I once created a style for Chameleon (a KDE3 plastik
> style). At that time it was bitmap based and I doubt that it is possible to
> create a style that even resembles Snow Leopard.
>
> I don't say that gnustep should adopt Snow Leopard's Look & Feel. But I think
> gnustep should adopt a more modern default Look & Feel that is more familiar
> to people coming from Windows, KDE, Gnome or MacOS X.
>
>> > My former institute used KDE as desktop. But people usually worked with
>> > Mathematica or MatLab. If we did coding it was mostly low-level numerical
>> > stuff in Fortran, C or C++. I doubt that it is a good idea to target
>> > researchers with gnustep. What advantage would gnustep give them?
>>
>> No less than the ones you mentioned, and more considering GNUstep is
>> way more advanced in terms of usability and consistency.
>
> NeXTSTEP was used at universities back in the 90's because it was way better
> then other systems. But the world changed, Windows / Linux with KDE or GNOME
> is good enough for people today. Developing applications with ObjC/gnustep
> might be easier / more convenient then developing applications with say
> C++/Qt, but it is not a fundamental improvement.

Except I wasn't talking about code development with objc, I was
talking about making apps that are far more usable without necessarily
doing much more work to achieve this goal. The fact that those other
systems you mentioned are "good enough" is because people don't know
any better. They don't  have as much information on alternative GUIs.
Apple's is known, but not well known. And both Gnome and KDE are
basically Windows clones (interface-wise)

> And back in the 90's Windows was not as dominant as it is today. Today most
> people are familar with the Windows GUI (even scientists  / researchers) and
> it is always hard to get used to something new. gnustep Look & Feel is
> radically different from what people are used and I guess most people prefer
> toolkits that enables them to write applications that has a Look & Feel people
> are used to.

This is bullshit. The fact that the world has changed and new things
are hard to get out the door is just your own point of view. Getting
people to embrace a "new" concept as GNUstep just depends on how you
do it and how much work you're willing to put into it. That being
said, The whole NeXT interface was built to be as usable as possible,
and one of it's well known advantages is that it's easy to learn.

>> behind some other compilers. As for GC, do you REALLY want that? I
>> think it's way overrated. It tends to encourage bad programming
>> practices, and usually kills performance.
>
> Yes. I have to code Java / .Net at work and I think GC is one of the things I
> like about Java and .Net. I know you can still have memory leaks, but GC
> allows me to think about the problem I want to solve without having to think
> about allocating / freeing memory all the time. Newever versions of the JVM
> include stack allocation and the garbage first collector, I don't think you'll
> have any performance problems with that.
>
> Michael
>

I don't use Java or .NET, but the one person I know who uses Java
hates it for lousy performance. And as much as a holy grail as it
seems GC is still wrong for objc (even if it fits well with Java)
because it has no primitive to dynamically allocate blocks of memory
in the first place, it's done through standard library calls. If
you're going to have an automatic deallocation mechanism, then you
should have an automatic allocation one as well.
-- 
Besos, abrazos, confeti y aplausos.
Jamie Ramone
"El Vikingo"