GNOME - Future ideas

Hyperbolic view: "There would be no need for a window list."

That is, and will be, needed for a long time forward. Having a lot of apps "minimized" on the edge of the screen would just clutter the workspace and take up the much needed space. Having less on the screen is better for the eyes than a lot of unneeded details, and those apps floating on the edge would be this sort of unnecessary and distracting annoyance. Besides, a window list is easy to integrate to menu panels and such, and easy to hide and still be accessible. It also takes up less screen space. And besides, the circular popup menus seem cool and neat, but they also eat alot of real estate. Yes, one doesn't need to move the mouse around so much, but isn't that also the point of having a window list?

File management:
"Automagical emblems for dirs guessed from contents. Search in selected files"

Atleast I like to have a responsive desktop environment, and having the file manager to constantly rummage around my hard drive would pretty much slow everything down. It would eat up some CPU time, and also the time for the hard disk to retrieve all the data added to that, would it really be so good an idea? And as a note to this: laptops and such usually have really slow hard drives and should be accessed as little as possible, also to save battery power.

Full spatiality:
No. Absolutely no. Perhaps an option for this, but no removing of the "regular way".

-WareKala

> Hyperbolic view: "There would be no need for a window list."
> That is, and...
Point taken (partially). I am purposedly putting things a bit on the edge here. I am still convinced that things could work, but would also be reluctant to leave out pagers and taskbars. But both things could coexist. A boiled down version of the window-managment ideas would be feasible to some degree at least. Project Looking Glass as actually already implementing some sort of hyperbolic desktop.

> File managment:
> "Automagical emblems for dirs guessed from contents. Search in selected files"
> Atleast I like to have a responsive desktop...
Well. Everything in my idea-list should be preceeded by the words: "A proper implementation of...". I wouldn't like the filemanager to scan the directories as I browse around. A beagle-like daemon (or just weekly cron-job) could sort this out. And it wouldn't have to run intensively very often, since your basic filesystem structure isn't altered that often. On the hardware side of things... Note that Gnome 3 is at the very least two years from now.
PS: I don't mean that the fm should search dirs/files as you selected them. That would be too cpu-itensive. I give you that.

> Full spatiallity:
> No...
Full spatiallity would be a major part of the core system. It would probably not be possible to have this optional. I would very much like it, some wouldn't... I use my Tomboy notes more and more just because of this feature.

I sort of like the idea of hyperbolic view, but I can't really say anything definite before trying it out. But, would it be too strange/distracting if the window being pushed would minimize on taskbar if the other window squeezed it all the way to the edge? Also, I need to say that it would be good if the windows resized themselves back to the original size when they have enough space. I would just hate to have to position and resize them all the time.

I don't think the "automagical emblems for dirs guessed from contents" would be such a usable idea. Either it would have to guess filetypes purely by suffix, if any, or it would need to categorize everything by contents. In the latter case the process would take quite a long time on any normal cheap system and if the results were stored in some sort of a database, it would need to be very fast and compact. Of course, if also the modification times were safed, then it would speed up any subsequent scans.
-WareKala

Apart from the nice stuff i'm not sure about circular menus.
The eye is trained to scan from right to left (or left to right) so that may outweigh the advantage of shorter distance to each item.

Yes. It is hard to judge these issues. One would really need a proof of concept implementation to verify this. Though I would tend to agree with you. Having a sane and _consistent_ way of placing entries throughout apps, would aleviate this a bit...

Great idea about the extended hyperbolic view! I will definitely add it!

I love PSLOP.
The circular menus are great, only that icons without caption are not a very accessible solution. But the idea is so good that hopefully this could be solved somehow.

The only thing that I do very much not agree with is

"Better Space Utilisation: Apps utilizing entire window/screen to show documents."

With ever growing screens/screen resolutions, that would IMO be a waste. On my 24" screen, a fullscreen wordprocessor would also be almost unusable - it would simply be too huge.

Screen space is a fast growing resource today - smaller resolutions will cease to exist in the long term. Window managers of the future need to take this into credit.

About circular popups:
======================
Yes. I know there are several issues attached to this one. It just fits the bill so perfectly regarding Fitts law. That was how I came up with it :)

Pros:
-----
- Very fast acces once you learn the placement. If "save" was always in the top, "exit" always in the bottom etc. It would be very fast indeed.
- Eye candy :)
- Easy (and obvious way-) to close when accidently triggered

Cons:
-----
- No text (not in any obvious way atleast)
- Not left-right or up-down oriented
- Few or many entries could yeild weird layout. Workaround needed.

The more advanced tooltips that I hint, would really be a gem also - If you ask me...

Space Utilization
=================
I have a 17'' monitor at 1280x1024 and would really enjoy more screen real estate. I can't afford anything bigger :( Laptops (as also pointed out elsewhere) are likely to hit a limit on screen sizes also.

Currently I have a locationbar as the only toolbar element in Epiphany. If I could make it fade in/out as desired I would definitely be doing it!

This one combined with better memory utilization might be nice on older hardware.

Screenshots of easyGestures here. I might just steal some of their ideas (or even illustrations) :)

We should not discard something that saves space just because (some) people can afford multiple huge screen.

And that wasn't my point. The point is, basing windowmanagement on fullscreen is bad because it doesn't work everywhere. Just like wasting space is bad, because it doesn't work on small resolutions.

Application menus in the System panel bar is not progress, it's confusing.

Apple only did it because their displays were too small and they didn't want to use more space than was absolutely necessary. However, the classic Mac OS they were doing it with and the machines they were doing it on were so simple and basic that they pretty much didn't run more than one application at a time, or at least you couldn't fit two applications on screen at once, so having a single application menu bar that's right at the top above the application makes some sense.

Skip forward 20 years and I'm sitting here looking at a dual monitor display, each of which is at 1600x1200, I have applications that never go near the top of the display, so why does it make sense for their menu to be up there?
Ok you get the mild benefit that they are at the display border so you have to aim slightly less accurately to get to them, but that is a particularly weak argument imo, as every other UI element in the application still requires the accurate pointing one generally assumes of a computer user.
The downsides are that your menus are a bloody long way from the application, they're in a place that doesn't make sense (unless we're going to split out application toolbars and put them on a common dock too?) and you open yourself up to a world of really confusing behaviours, such as when you close the windows of a Mac OS application, but it still lives on as a set of menus at the top of the display.

Great progress.

7. Beagle or Storage inclusion
[...]
Speed and memory footprint reduction.

I think this are conflicting goals, a VM has its resource requirements.

This clearly needs to be in synch with the target audience on which GNOME choses to focus. Strong hardware requirements are likely not to go well in 3rd world countries or on company desktops, so bleeding-edge technologies are more likely to succeed on home desktops in industry nations. I would welcome lighter hardware requirements (though GNOME is lightning fast on my 3GHz home desktop, it is not really usable on my company desktop), only because I love the idea of supporting 3rd world countries.

Has GNOME ever defined a target audience?

Hence when I said I would exchange all of the feature wants for more speed optimization and memory footprint reduction.

But come on folks .....

We all knew gnome-terminal had some nasty mem leaks hidden in its innards and guess what? It did.

The folks at Novell figure its a commendable enough task in fact to offer the bounties for reducing the memory footprint. (Is that a gnome pun? :) ).

KDE is a big project with a feature set so large and cumbersome you could almost say that gnome 2.0 was a direct reaction against its confusing nature.

Still they went on a resources diet and slammed a couple of coding speedballs and a lot of KDE fans are damn happy they did in what I believe they said it was like 3.1 or something.

There are a ton of things we should NOT mimick about KDE but slimming up the memory footprint and optimizing the speed in general is one of them.

People always say that features = slow but then again a fast, small desktop environment like XFCE is X.org and GTK2 based and it has better windows network browsing and it also has an xdg menu editor and its still lightning fast. Fast and small is their schtick. Not saying gnome is ever going to be that fast. I am just saying its a noble goal and the gnome project members themselves recognize the need to decrease the memory footprint.

By the way, notice hard, most of the conflicting goals including memory reduction and the inclusion of Beagle are already on the roadmap for gnome 2.12.

Not sure if it will all get in mind you but at least right now its on the map.