Fridger,


Ah, okay. I do not read the texts of the menu bars; I just click intuitively at the right place (at least with programs I often use - that's the only reason why I prefer smaller font sizes...).


No, it isn't
I avoid to work with too many different line- and point sizes. As with fonts => too many (wonderful)
fonts destroy the harmony. Furthermore, proportions are important. But, what I see is very well done;

the "CAS"-icon - in its unproportionality (don't know if this word exists, but I hope you understand what I try to explain) - will add a certain touch... (By the way, I think it is a good idea to leave the very tiny dots.)


The colors are well choosen. Too many colors seem rather infantile.

What I intended to address (regarding the grid icon) - at least I would give it a try - e.g.

horizontal lines: red, - vertical lines: blue... and a clear edge.

...because similar design (like the grid can be seen now) isn't seen within the other icons...
but that's just eye-cosmetics.

Great work, Fridger!

Michael

Michael,



There is a well-defined startegy behind my design about when to use which one of the two colors!

The red color displays precisely those features that the clicking of a certain icon controls! That's why the orbit display in the orbits icon is red and why the entire grid in the grids icon has to be red! A click on these icons swich orbits AND grids ON and OFF.

Fridger

Question not related with the UI at first sight but... : do you plan to support planetary simulation with special care on atmosphere simulation, rings etc? In another words, do you plan to improve the existing code from Celestia at this level?

Yes. But this still requires substantial experiments: The general framework (Mie theory) is quite resource intensive. So one has to implement a number of successive approximations that still give a nice rendering and yet don't lower the fps too much.

I forsee such rather extensive projects during the public development.

Fridger

What would be interesting for me is to discuss some ways to achieve a UI capable to target both neophytes and professionals; the purpose would be to give an experience or a knowledge by "layers", beginning with some basics to end with the most sophisticated features...

Celestia's minimal UI was a real pain for the neophytes (or for a younger public) it's why I imagine it would be fantastic to achieve a UI that could guide (and educate!) the user all along the complex experience such software may propose...

So basically my interrogation here is if such thing may be possible or if the reboot of the project (let's consider Celestia cold dead) will target only professionals and gifted amateurs?

Honestly, I have not spent so far any thoughts to the possibility of a unified layer-based UI that could address BOTH astrophysical / astronomical professionals and neophytes alike. Instead, I always put the 3D visualization of crucial astronomical / astrophysical / cosmological concepts ahead, complemented by a UI that hopefully incorporates a clearcut underlying logics along with an intuitive operation.

But in principle, this could be realized, based on a scripted configuration for the two UI layers.

It may well happen that the classical Celestia application will satisfy those people more that have little background in astronomy / astrophysics or want to focus on creating their own worlds etc. We shall see how it goes...

As the name of celestia.Sci suggests, I am not planning to make significant compromises in scientific rigor for the sake of making the software LOOK more educational. Throughout, a reliable documentation of both imaging and data sources will play an important role.

In any case it seems unavoidable that an extension beyond Celestia towards a colorful Deep Universe and cosmological visualization, paired with multi-wavelength astronomy and relativity, will require a certain amount of astrophysical background in order to understand what exactly is being visualized.

But also here, the proven sentence "learning by doing" will surely help, once celestia.Sci has been made available for people to judge and to express their preferences. In any case the UI of celestia.Sci (and that of Celestia Qt, pre 1.7.x ) is already now much superior to the classical Celestia 1.6.1 UI.

Fridger

Héhé, just say it, NO BULLSHIT add-ons allowed!

More seriously, the add-on topic is not that simple; on one side keeping the purity of data and work seems interesting but there is no way to control what people will do after a realease unless you encrypt the final package to prevent any modifications of the files.
On the other hand, the add-on system is useful to change data sets, add guides or whatever and the liberty to increase the default available features seems quite interesting... But then again, people can do and will do bullshit... ... so, que faire?



I totally agree here the philosophy and the deep mechanisms need all the rigor possible. Now the UI do not interfere with this basic rule; if documentation (and/or other) may help the user, it could be part of this UI (via info panels, whatever) and it's part of the idea of layers, elements useful to neophytes that could be removed or deactivated by pros...



True, that's why I'm asking about the target, because to me it's a very important point impacting the entire development of the UI, from the first thoughts to a final absolutely exquisite piece of software!

All agreed.

No reason to worry, since the Qt toolkit is so flexible that any later expansions of the UI can easily be incorporated. The Qt Designer tool makes such changes very easy and time-saving.

Therefore it's probably the most pragmatic approach, if I first implement those aspects in the UI that will most probably remain in the end.

BUT a timely discussion about the scope of the UI on a somewhat longer time scale is most valuable, anyway!

Fridger

Hi all,

meanwhile I have incorporated quite a few more features into celestia.Sci's UI.

Here is a typical display with three toolbars activated:
the Time toolbar starting on the left, followed by the Guides toolbar and the new Graphics toolbar. Their separation is barely visible in this screenshot:



Besides the option of turning off some or all of the toolbars, they may also be moved according to gusto: e.g. like in this view (with the Time toolbar deactivated)



They even may be moved outside the main window space, somewhere to the desktop.

Next let me focus on a few new features:

First of all, I now display all toolbar icons by default with an increased size of 32x32 px, which hopefully reduces the guesswork about their meaning. In the preferences dialog of the menu, this may be reset to the Qt default (which is generally smaller).

A main new feature concerns the two icons on the new Graphics toolbar. Each stands for an exclusive sub-menu that is shown upon clicking on the black triangle on the right of the respective icon. The chosen sub-menu options are saved and restored upon restarting.

1) Choice of star color table (left):
=========================
Here one may select the desired association of the stars' surface temperatures with RGB colors. Presently, two color-temp tables are supported: The vivid Blackbody D65 color table and the weak classical colors (that were used in older Celestia versions). I prefer the vivid Blackbody colors. The displayed icon changes with the selection made. Here are the respective screendumps:

• 
• 

2) Choice of DSO color scheme (right)
============================
The options one is offered here for DeepSpaceObjects (DSO) require some explanations. Three different color schemes are available for now. They differ by the value of the Saturation S in the HSV coior space (H=Hue, V=Value):

• Vivid DSO colors <=> S = 1
• Adaptive DSO colors <=> S increases with image zoom
• Grayscale DSOs <=> S = 0

The Saturation S, with 0 <= S <= 1, is the appropriate color-space parameter to modify the Colorfulness of an image!

One easily derives a general formula for a RGB triplet as function of S with everything else kept fixed:

(r', g', b') = (S * r + M, S * g + M, S * b +M)

where M = (1 - S) * max(r, g, b);

At least for the two limiting cases S=1 (Vivid DSO colors) and S = 0 (Grayscale DSOs) you can immediately see that

S=1: r'/g' = r/g; r'/b' = r/b
S=0: r' = g' = b' = max(r, g, b)

In the 2nd adaptive DSO color scheme, I let S depend on the image resolution (zoom) or in other words on the fraction of the screensize that a given DSO occupies. The physics idea behind this option is to interpolate smoothly between naked eye vision (no DSO colors visible) and "telescope" mode (<=> Automag!), with colors getting more perceptive as the size of the telescope increases.

Here are screenshots of the galaxy triplet: NGC 70 (blue, Sc type), NGC 71 (orange, S0 type) and NGC 68] (orange, S0 type) illustrating the three DSO color schemes.

A color photo from the site of Prof. C. Seligman

http://cseligman.com/text/atlas/ngc0a.htm

may serve as a low resolution reference


with the resolution of celestia.Sci's NGC 70 & friends,... being much better

Note that the icons change again, depending on the option chosen...

[Click on images for true size]
A) Vivid DSO colors (S = 1)



B) Adaptive DSO colors (S increases with size scale!)


C) Grayscale DSOs (S=0)



Finally, I hope that at least some of you understand the meaning of the icons that go with the three choices. If not: Colorfulness is displayed for increasing size scale (from left to right). In the first case, colors are vivid throughout. In the second case colorfulness increases towards vivid colors with growing size scale. Last: Grayscale DSOs throughout -- for purists

Actually, in the Preferences menu that also differes much from the original Celestia-1.7.x one, one may set ANY desired value for the Color Gain. This value determines how fast the transition from colorless DSOs to full vivid DSO color takes place when the DSO size on screen increases!

Here is the relevant menu page:


Any comments are welcome!

Fridger

Excellent.

_________________
Guillermo Abramson
Bariloche, Argentina

the stunning images and the new features are as extraordinary as the new toolbar icons are... pretty terrible (sorry ).
But it doesn't matter too much.

Once again - to avoid misunderstandings - the new features, especially those convincing pictures, are
really great

Michael

This is the word...