CELESTIA IN THE FUTURE

Description: The place to discuss creating, porting and modifying Celestia's source code.

K.A.V M
Topic author
Avatar
K.A.V M
Topic author
Age: 27
Posts: 15
With us: 1 year
Location: Marseille

#41by K.A.V » 08.08.2017, 11:10

Please stop offtop.

Janus
Janus
Posts: 61
With us: 1 year

#42by Janus » 08.08.2017, 15:40

K.A.V wrote:#41by K.A.V » 2017-08-08 03:10
Please stop offtop.

Agreed.

The question remains though, where is celestia going?
Close your eyes and picture this.

What do you see celestia doing in five years?

Collect enough of those and you can start developing a picture of the path it needs to follow.


Janus.

Cham M
Avatar
Cham M
Age: 52
Posts: 4313
With us: 13 years 7 months
Location: Montreal

#43by Cham » 08.08.2017, 16:11

I'll tell the truth that nobody here wants to hear : :fie:

Celestia is going nowhere. It's a dead app, pure and simple.

Nobody will be competent enough to understand its code, modify it, and make it to evolve. (I don't include personal compilations without modifications, and I don't count that "fake" project called Celestia.Sci, since it's a farce designed to glorify Fridger's cosmic scale ego !).

If your computer's OS don't change too much in the future, you'll still be able to run Celestia for years to come, though. But you'll have to accept that one day, because of some major changes to the computers and/or their OS, Celestia will simply cease to work. I'm personaly expecting this day with sorrow and even "terror" ! (because of all the work I've done on the addons front) :ill:

What appears to be sad, is that there is still no other "modern" astronomy app comparable to Celestia, even today in 2017 ! What a pity !

So talking about "Celestia's future" is like beating a dead horse !

The only "future" I could foresee is on the addons front, while Celestia is still running on current hardware and OS.


...

Despite that truth, I'll be very happy to be proven wrong ! (who knows, maybe one of these days, in a very far away future... :pray:)
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"

Janus
Janus
Posts: 61
With us: 1 year

#44by Janus » 08.08.2017, 17:41

@Cham

I personally disagree with you.

I see much potential in the program.
What I have not seen so far is coordination, but that is to be expected at this stage in the resurrection of such a large and complex project.

In no small part, what is holding it back is that no one can tell what it takes to do anything with it.
This is open source, so both the source and its requirements need to be documented.

The biggest source of holding things back I see is that no one can tell what doing anything takes.
I like the windows x64 release very much.
However, it required VS2015/17 to compile it, which came a a shock.
I expected it to be much as it has been, because no one said anything about any changes on that scale.

I suspect the same issues apply in Linux, though I am not enough of a Linux user to know.

In my own private branch I use for my own stuff.
I have it compiling with VS2008/10/12 & QT 4.8/5.4/5.6 using QT creator with the fore mentioned VS versions.
Each version detailing which combination of compilers that was used to make it.
I like things neat and organized.

The change in requirements came as a shock, especially with no warning.
I have managed to get the code to compile back to VS2013 so far, a real limitation imposed by eigen.
The artificial limitation imposed by the release is the precompiled libs that can not be used on previous VS versions.
Which is why requirements need to be detailed, I got the code to compile, then got a whole new fresh round of errors with no warning.
I am still working on finding a way to use the newer eigen on older VS versions, but it is slow going.

In order for people to tinker with Celestia for learning or customization, it needs to be easy to tinker with.
Not having a list of requirement makes that harder.

For my own part, I am using Celestia to study how VS and C/C++ work.
Among my personal projects with are.

1-> A new indexing system and database to fix what I see as shortcomings. This one using a celestia index, then Gaia, Hipparchos, Tycho, Polar Coordinates, Rectangular, Comment index 1, Comment index 2
{ CelestiaDex, Gaia#, Hip#, Tycho#, Polar[RaDecDist[record RA,Dec,Dist]], Rect[Record X,Y,Z], Comment Index1[3], Comment Index2[5], Comment Index3[7]}
With comment index being the same as footnotes in articles.

2-> Load time extension of the database with a nearest neighbor index. Generated after addons etc.

3-> Pathing using a version of the astersims system. One that lets me turn them on/off in a script, and build them as desired in memory, then save them as a json formatted file based on the above database, using which ever index works best. Both in system, and out of system.

4-> A dual display output, using partial screen offsets to generate a VR style display using flexible lcds in a custom headset.

5-> Further expanding the ability of Celx to lookup stars by coordinates in retangular or polar space.

6-> Making time easy to set in script, then add script controlled screen shots, or add frames to video recording.
This allows literal frame by frame recording of a video under script control without worrying about having to do real time recording.
Aiding it being used as a presentation system.

7-> Make it easier to control rotation of view independent of follow focus or orbit.
Such as keep focused or centered on Venus, while following Earth, so you can see how venus moves in the sky, and why.

8-> Further integration of alternate control inputs. I already built a kinect hand motion control system, I am working on adding mapping from a tablet (Which has gps/tilt sensors) so I can move the tablet and have that control the display in celestia.

The list here can go on as far as peoples imaginations take them.

It starts however, with a solid foundation.
Being able to use older compilers is a plus in windows.
I have copies of three, and have borrowed a fourth.
I do not and will not ever use VS2015/17 because they have usage and other telemetry built into them.
I tried to use VS2015 in a VM for a customer, but it kept trying to phone home.
I also do not use community versions because I avoid accounts with places like M$ period.

So for me, my personal priorities are back porting the new code, or learning from it, and bringing the older code up to 64 bit.
If others want any of them, then I will post them here so they can be grafted in by real C/C++ programmers.

Celestia is not dead, it has a future.


Janus.

Cham M
Avatar
Cham M
Age: 52
Posts: 4313
With us: 13 years 7 months
Location: Montreal

#45by Cham » 08.08.2017, 17:50

Janus,

then prove me wrong... and good luck ! You'll have to get a HELL LOT of luck with your project !
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"

Janus
Janus
Posts: 61
With us: 1 year

#46by Janus » 08.08.2017, 19:12

@Cham
For my own uses it is not luck, it is just time.
My main intent in naming them here is hoping that others are working on similar things, and maybe we can share notes.
If nothing else, it shows that this project is not dead.

My hopes are not high, but I am patient, and this is far from my only project.

Goodluck to us all who believe int he program though.


Janus.

Edited because I no better a speller when I am in a hurry, than I am a C++ programmer.

John Van Vliet
Avatar
John Van Vliet
Posts: 2620
With us: 14 years 11 months

#47by John Van Vliet » 09.08.2017, 00:01

I suspect the same issues apply in Linux, though I am not enough of a Linux user to know.

not so much

the current testing version of gcc is gcc7
gcc6 is the current with c++14 support
gcc5 has c++14 support ( long term stable _)

and building MOST software can still be built with gcc 4.8
4.8 is used in the long term support kernel and operating systems - even the current opensuse 42.3

now a fast passed operating system like Fedora or SUSE-rolling will be using gcc 5 and or 6

even my qt5 fork builds with the older gcc 4.8

now for building with 5 abd 6 i looks like just minor changes will be needed

the gcc developers and forcing people to write SECURE code by default in newer versions
this started with gcc 4.1 and likely will not stop any time soon

Janus
Janus
Posts: 61
With us: 1 year

#48by Janus » 09.08.2017, 00:28

John Van Vliet wrote:the current testing version of gcc is gcc7
gcc6 is the current with c++14 support
gcc5 has c++14 support ( long term stable _)

and building MOST software can still be built with gcc 4.8
4.8 is used in the long term support kernel and operating systems - even the current opensuse 42.3

now a fast passed operating system like Fedora or SUSE-rolling will be using gcc 5 and or 6

even my qt5 fork builds with the older gcc 4.8

now for building with 5 abd 6 i looks like just minor changes will be needed

the gcc developers and forcing people to write SECURE code by default in newer versions
this started with gcc 4.1 and likely will not stop any time soon

Nice to know that Linux is not having as much trouble as Windows with going 64-bit.
Do you know if QT with Mingw on windows will compile Celestia?
It would be great if it did, I am not fond of VS.

I also freely admit to not having the slightest clue what you mean by c++14, while the gcc5,6,7 seem apparent enough.
As I have said many times, I am not a real C/C++ programmer.
I am simply following patterns and mimicking what I see elsewhere, and I nearly always choose the simplest implementation.
Speaking strictly as an outsider, C and most especially C++, seem to be designed for the express purposes of obscuration and obstufication.
There are however, even worse languages out there, and they are badly outnumbered by those that are worse.
Mixed language binaries in embedded environments are so much fun to figure out, and so much easier than C/C++ programming.

Admittedly, most of my work involving them is providing example code or working programs that do just one step of a process.
Then one of the customers people takes that code and integrates the functionality into whatever they are doing.
I have so far managed to keep from having to dive into C/C++ object/class dung heap, though I know it will not stay that way forever.

I have no idea what your segue about secure code means.


Janus.

John Van Vliet
Avatar
John Van Vliet
Posts: 2620
With us: 14 years 11 months

#49by John Van Vliet » 09.08.2017, 03:00

the last time i built Celestia on windows was using MinGW on XP sp3

it built very easily , now i had been building Gimp and it's dependencies for a few years

now this WAS 32 bit -- it was xp after all

there is a 64 bit fork of mingw
http://mingw-w64.org/doku.php

the mingw/msys set up is normally the best to use

HOWEVER there are very few 64 bit ports
almost all are 32 bit

So you will need to build almost all the toolchain and prerequisites

-- use qt ( that looks to be the main new GUI )
-- build the prerequisites in mingw-64
-- off hand right now i do not have a full list of all the prerequisites

Janus
Janus
Posts: 61
With us: 1 year

#50by Janus » 09.08.2017, 03:43

That's ok.
I have no clue how to build the prerequisites yet either.

As little as I know about VS, I know gcc/mingw even less.
Even msys is confusing beyond the idea that it is much like the linux commandline.

What is the difference between mingw, mingw64, mingw-w64 etc anyway?
I see lots of them, and none of them have anything like the cygwin ease of install.
Though being able to automatically select source for everything you want in setup would be great.
As it is, I am going to have to just mirror an ftp of gcc since it contains a mingw so I can study sources.


Janus.

t00fri
Avatar
t00fri
Age: 15
Posts: 8809
With us: 15 years 4 months
Location: Hamburg, Germany

#51by t00fri » 09.08.2017, 13:39

Cham wrote:and I don't count that "fake" project called Celestia.Sci, since it's a farce designed to glorify Fridger's cosmic scale ego !).
..and here my old "friend" Cham released another fine demonstration of his mean character.

Cham, please behave yourself! Recall that I didn't provide any kind of reason to you for this sort of rude talking...

Fridger
Image

FarGetaNik M
Avatar
FarGetaNik M
Posts: 307
With us: 5 years 2 months
Location: Germany

#52by FarGetaNik » 09.08.2017, 18:51

t00fri, I don't wan't to defend him for his expression, but I find it interesting that you only participate on this forum anymore when it concerns such things. I don't get a good impression on this behaviour either, that might explain where this is coming from.

t00fri
Avatar
t00fri
Age: 15
Posts: 8809
With us: 15 years 4 months
Location: Hamburg, Germany

#53by t00fri » 09.08.2017, 19:40

Well, perhaps you understand better, if you have a look at the involved time scales along with a condensed "history" of celestia.Sci development:

With Chris Laurel, I was working at the forefront of Celestia development since spring 2002 (!) and thus I am an official early co-author of Celestia. Unlike Cham's claim, I do know the code very well ... ;-)

When Celestia development sunk into some sort of a Coma, I started the celestia.Sci "adventure" in 2010, i.e. 7 years ago. Until 2013, I worked on it alone and owned the Copyright of the new code, which was shared later with the other team members who had joined in meanwhile. All along I posted notes about celestia.Sci news/ achievements in the Celestia forum( shatters.net). You may easily retrieve these ...

Besides being IT experts with long C++-openGL-Qt coding experience, all our coding team members are professional researchers in (astro) Physics and Space research, respectively. Since a long time, celestia.Sci runs at 64bit level with Qt 5.9.x, MSVC 2017 (Windows 7..10), gcc 6.x (Linux), MAC OSX, Eigen 3.3.x etc. The celestia.Sci code is perfectly stable,

Of course, we follow what you guys are doing and vice versa...

But since there are only very few people in your CelestiaProject that seem to have remarkable coding abilities, we didn't feel too much attracted yet.

Of course there is also plenty of competition that is to some extent problematic and should be dealt with somehow!

So much for now,

Fridger
Image

FarGetaNik M
Avatar
FarGetaNik M
Posts: 307
With us: 5 years 2 months
Location: Germany

#54by FarGetaNik » 09.08.2017, 20:57

You don't have to emphasize your archievements that much, I'm well aware of that and respect your work. It's just a shame that the original Celestia, despite gaining an optimistic community again, still lacks the experts and resources to progress.

Actually I was excited a lot when I heard about Celestia.Sci and I am following the Celestial Matters forum regulary as well. You probably heard it often enough, but Celestia.Sci also doesn't seem to progress a lot, or you don't really share any progress. I know you have a different sort of policy in your mind, but it's a shame we can't really cooperate for that reason. Both projects might benefit from each other. But with the current situation we can't expect any significant progress.

john71
john71
Posts: 247
With us: 1 year

#55by john71 » 12.08.2017, 18:46

t00fri, we are not saying that you should give us freely your work or code, but I think you (your team) can help out the new Celestia team in some small ways. Celestia in my opinion in it's current form is almost perfect for 3D astronomy content creation, it only needs improvement considering bugs, performance issues and future compatibility. I think you should help the Celestia community with these tasks, because it is not that hard for you. We don't expect any new functions but a minimal help from it's creators to keep Celestia usable.

t00fri
Avatar
t00fri
Age: 15
Posts: 8809
With us: 15 years 4 months
Location: Hamburg, Germany

#56by t00fri » 12.08.2017, 20:08

John71 and others interested,

first of all, my collaborators are busy for a while, so I am currently the only active celestia.Sci team member. Clearly, what you request, would only have a chance of success, if we'd manage to set up a formal agreement about mutual collaboration (rather than competition). At the same time it must be made clear who has the saying at the development front! Moreover, it must be made clear who is entitled to speak on behalf of your community.

Our development team consists of IT experts with longtime C++-openGL-Qt coding and also much scripting experience. Two members (myself and Andrew Tribick) are professional (Astro-) physicists. Dawoon Jung (aka dirkpitt) has a Master degree in Space research since 2014. His thesis work was about Gravitational Lensing in celestia.Sci. All team members were longtime Celestia core developers and correspondingly are Celestia co-authors.!

It cannot be that someone from your community tells us what kind of bugs we should fix for you ;-) or what I should implement or do next. Don't forget, I have not looked at the Celestia code anymore since 2010, and I am convinced that celestia.Sci is much superior. With many new features, our code is very stable at 64bit for Windows, MAC OSX and Linux, based on Qt 5.9.x , Eigen 3.3.x and compilers MSVC 2017, gcc6.x ..

In 2010, I have separated from Celestia core development (to which I belonged since 2002), since I saw many limitations of the code and of the organization of the development team. So, whenever I would get back at the Celestia code, something quite similar to celestia.Sci would result ;-)

Honestly, I cannot see a practical route at this time without an intrinsically novel approach. But I can see many potential problems..

Fridger
Last edited by t00fri on 12.08.2017, 21:03, edited 3 times in total.
Image

john71
john71
Posts: 247
With us: 1 year

#57by john71 » 12.08.2017, 20:34

t00fri wrote:It cannot be that someone from your community tells us what kind of bugs we should fix for you

I see that, but I suspect that you have a much deeper knowledge of programming Celestia (even in it's pre-2010 state) than any of the new team members.

I don't really see how guiding or helping the coding work on Celestia would be an insurmountable task for you.

Or releasing the first stable (older) 64 bit version of celestia.Sci without your full-fledged scientific improvements would hurt your project.

As a "Starter Community Edition" it would be well received even without any revolutionary improvements.

Janus
Janus
Posts: 61
With us: 1 year

#58by Janus » 13.08.2017, 06:17

I would like to make a suggestion.

How about a thread dedicated to compilation issues.
That would be based on people asking for help with specific issues.
Start with OS, Compiler, Code Version, and error.
Relevant details beyond that would be case by case.
If someone from Celestia.Sci happens to recognize the problem, and are willing, they can answer a specific issue.
This is not a commitment to do anything from anyone.
This is not one person trying to tell another where to put their efforts.
It is simply, hey has some else already solved this, lets not reinvent the wheel unless we need to.

If development is going to happen, people compiling Celestia is needed.
The more people who try, is more people who can.
Once people get used to being able to 'roll their own', some of them will experiment.

Those experiments are what drives movement, and movement drives development.
No one can say what they do or do not like, or want, if they are not using it.

The questions about compiling will give an idea of what people want, based on what they are trying to add, modify, or fix.

I believe I will have the current github code compiling on VS2012 in the next couple of weeks. {VS2013 already works, but needs clean up.}
Once I do that, I am going to try to backport the current Egen to VS2010, then bring the whole project back as well.
I doubt I will be able to take it back to VS2008, but I will try when I have time.

Before anyone asks, no I do not run VS2015/17 community, they have telemetry, and require an M$ account, which I do not have, nor shall be getting.

If there is interest, I will offer the code here.
If there is not, then I will put the code up on my own site.

If people like what I do, they are welcome to use it.
If they do not, then they are welcome to ignore it.

I hope everyone has a nice day, but I am going back to arguing with VS2012 now.


Janus.

John Van Vliet
Avatar
John Van Vliet
Posts: 2620
With us: 14 years 11 months

#59by John Van Vliet » 13.08.2017, 06:44

Start with OS, Compiler, Code Version, and error.

like setting up autotools "makefile.ac" with a qt5 macro test
( the current one is grabbing the qt4 headers during the build )
i have a feeling that i am that person to give that help
you could write a doctoral thesis on autotools and M4 - and people have


but

Qmake works like a charm

so i am striping out non QT code and using qmake as default

K.A.V M
Topic author
Avatar
K.A.V M
Topic author
Age: 27
Posts: 15
With us: 1 year
Location: Marseille

#60by K.A.V » 17.08.2017, 12:35

t00fri, I absolutely endorse the words of john71, and I think you could have a little bit to help the project in the current issues (eg issues on GitHub), which we can not do due to lack of knowledge.


Return to “Development”

Who is online (over the past 5 minutes)

Users browsing this forum: 1 guest