Re: Nautilus 2.6 - We're going all spatial

[Ali Akcaagac <aliakc web de>]

On Tue, 2003-09-16 at 09:27, Eugenia Loli-Queru wrote:
> I agree with that. Embedded viewers on a file manager are
> useless and specialized apps should be used instead. Let's
> not pull a Konqueror on Nautilus.

Ok I'm replying to one mailinglist now to not duplicate the emails I
receive on my own by CC'ing them.

Besides the fact that I think the KDE people made a brilliant work with
Konqueror and the philosophy behind it, we also need to understand that
they are approaching a different way of doing it. Not that bad after
all. I install a new program which also embedds kparts and I'm on my own
deciding whether I use it or not through Konqueror. The final decision
is left to the User.

The entire 'New user', 'Expert user' talk is sowhat matured into a
religion on GNOME that even 'ordinary people' (please excuse this
terminology) claim wisdom in what the 'New user' or 'Expert user' likes.
We will never find a 100% solution for this and I think concentrating in
a good overall solution is the way to go.

How long does Homecomputer exists now ? ~40-50 years ? During all the
years the computer business and developers made a lot of mistakes as
well as a lot of innovative improvements and with every passing year
developers improved their software in the well known 'itterative
process'. ~40-50 years of this is the best use case study ever made. We
never noticed it as use case study because we behaved normally as every
other normal day in our life but in all the years we reached a point
where many Operating Systems came similar solutions. Desktop, Window,
Toolbar, Filemanager, Icons.

Maybe by trying to find a new way of doing things we indirectly cause
more problems and confusions as we aimed in first place. On the one hand
you want to be innovative by doing it differently than other systems did
(which is honourable) but here is exactly the problem 'doing it
differently than others' is exactly the huge mountain where people may
confuse because this system acts differently than others. No matter how
easy this system is. Ok getting new users and teach them 'our' way may
make sense but what when they need to use a different system say they
are working for a large company and need to deal with 'NanoHardware
DoorsXP' (name changed).

These are a few lines going through my head now.

As for the viewers part I do agree totally to Eugenia and the many other
people who raised this point in the past (some battles has been fought
about this on gnomesupport.org forums way back). They should reall
disappear. Conparing this with KDE's Konqueror is a wrong thing because
their techological framework is quite different while it fit's there
seamless and looks professionally solved on the otherhand looks 'hacked'
in Nautilus. Hey this is just my very personal opinion here.

The OO design is a nice one it remembers me of my old Amiga times where
we had our Desktop boot in no time and where we had a few icons on the
Desktop and clicking on them opened new Windows with icons inside and so
on. We can launch applications, click on documents (and have that open
MultiView) or click on a media file and have it open (some Media
Player). We could cleanup the Windows (arrange the icons inside it), we
could arrange the Window around icons (e.g. adjust the Window so it
covers the icons contained in it and save its Window positions), we
could group elements in Window and so on. Now where are we again ?
Nautilus' way decided for going OO and we met again a solution done 20
years ago on the Amiga or Mac plattform. What does this tell us ? It
tells us that old proven concepts are still good one. If these old
concepts are being realized in a modern framework and done nicely then
why not.

I want to click on a text Icon and have it open in Gedit, I want to
click on an application Icon and have it execute it, I want to click on
a mp3 file and have it open a media player (or have that module be added
to my playlist). The views stuff is a nice idea no doubt but not done
nicely imo. Let's be real here and agree that they have no real life
benefits. Nice for showing some eyecandy to some people but real life
benefits ? A lot of the views eat CPU (as someone recently mentioned in
the gnome irc channel, and even others simply segfault when switching
from one view to another or do not really remove controls that got
applied to Nautilus' Toolbar. Well we all know these issues and
shouldn't ignore them.

Benefits would be:

- Removing VIEWS, reduces error reports, less maintainance for
  developers (app developer, nautilus developers). Of course stuff
  like thumbnailing capabilities should be kept.
- clicking on an Icon runs the individual application (why do we
  have a mime system ?)
- Reduce of code inside of Nautilus (removing not necessary
  codeparts will also increase speed and make the app become
  snappier.
- Going the OO way will become a cool thing (in case it's done
  correctly) and I like to point again to the previous email that I
  wrote related to this with the former way Amiga did these things.

  ;--- reference
  MorphOS: http://www.morphos.net/images.php?gallery=13
  AROS...: http://www.aros.org/pictures/screenshots/
  AmigaOS: http://wuarchive.wustl.edu/~aminet/pix/wb/index.html
  ;--- reference

I also think that a few replies raised already e.g. 'it will add another
step of complexity because we deal with Objects and Lists' are not
really valid because we deal with Objects already since day one of
Nautilus' existence. We only switched from one mode to another because
of the reason that we can.

Ok I could write more about it but I think that looking at the reference
screenshots above may give a hint how it can be done correctly. At the
end, no matter what will be decided we one day will met anyways to old
proven concepts. We only shift the ball so far left that it shows up
from right again. This sentence describes that we usually turn in a
circle here.

greetings,

Ali Akcaagac