Celestia Forums

Accordingly with the content of the recent threads:

- Some architectural issues
- Nice little soft for Solar System Simulation
- Is it time to fork Celestia?
- Galaxy Colors etc
- others...

It seems that beside:
- the persistence of news shortage from Chris (his last post is of March 30th) and
- the lack of spare time form the historical development team,
there is a new growing community of skilled developers that are going to criticize constructively and with apparent good competence the current Celestia 1.4.0 pre. 6 (or whatever) code.
These people are proposing interesting issues in order to implement new bug fixes and new exciting features.

So here it is my breackdown proposal: a real fork of Celestia. I know it sounds blasphemous! But this long period of stagnation has to terminate! The new project should be called "Experimental Celestia" or "Beyond Celestia" and will have the purpose to merge again with the main Celestia (hopefully in a near future or ASAP).

This new project will be opened to as many developer contributors as possible (a minimum of skill selection will be done of course) in order to produce Experimental extensions and improvements to Celestia.

To avoid any useless complication the Experimental Celestia will be GLUT only, so:
- no specific UI will be supported,
- no end user runtime distribution packaging problems,
- nor different OS versions,
- multiplatform test and debug by definition.

These statements will allow experimental developers to concentrate on the real interesting and coding things. If the new shared experimental code will be:
- useful,
- bug free and
- won't introduce very bad performance issues
the current official Celestia Development team members will be able to extract portions of the Experimental Code and insert them in the official version.

Obviously the Current Celestia Development team members will be able to work on both projects.

For anyone that is interested I can upload the current Celestia Code in my CELUI SourceForge project (its a matter of minutes) and ASAP you'll be able to start playing and sharing experimental code.

In order to allow good communication (that is absolutely essential for efforts coordination) between members I suggest to reactivate the Motherlode phpBB opening the: "Experimental Celestia" forum. This because the Ibiblio server has better performance, and continuity against the shatters.net. Instead I can open a new forum in CELUI project in the Sourceforge.net, but it is not as good as phpBB.

So please Vote! And post your comments.

This sounds like it could work - it's similar to what happens with POV-ray: there are several unofficial builds (e.g. MegaPOV) which extend the features of the main program, usually with experimental and/or specialist features which are still in some kind of testing phase. From time-to-time, features from the unofficial builds make their way into the main program, so those who want or need bleeding edge features can get them, while the main build retains stability.

Paolo wrote:...and will have the purpose to merge again with the main Celestia...

This is the cruxial point for me, but as I'm not a contributor in the code I won't vote right know...
I'm wondering if your proposal (Paolo) will bring a better confort of work for the coders... (actually people seems to work and communicate without too much difficulties...). Also, is a second forum really necessary?

I'm hoping Chris will take 2 minutes of his precision climbing time (:wink:, Fridger) to give us his opinion...

Bye.

ElChristou wrote:
I'm wondering if your proposal (Paolo) will bring a better confort of work for the coders... (actually people seems to work and communicate without too much difficulties...). Also, is a second forum really necessary?

Hi El Christou

I think that you are aware that now the most of the communication between developers happens behind the scenes using private messages and the Sourceforge mailing list (lately not so much indeed).
This doesn't allow to achieve the correct exposition to implementation issues, and moreover prevents for correct coordination of efforts. The less desirable thing is that many coders work on the same thing creating conflicts or without previously agreeing about important structural decisions.
This forum should be of course used, but are few months that responds very slow. Coders hasn't to spend precious time waiting for web pages update, they have to concentrate on real coding problems.

Hi all,

after some thinking, I have decided that I shall NOT join such a venture. Some reasons are these:

--Chris has selected the original developer team both according to required abilities, but also such that there was a quite uniform agreement about Celestia's design philosophy among the developers!

I just have NO intention whatsoever for lengthy discussions with MANY new potential developers about reconsidering Celestia's main goals and purpose. In addition I share Chris' view in preferring a smaller highly competent gang over a larger less competent one...

I for my part will immediately leave the Celestia community (like Grant did!) in case the basic focus in development goals changes significantly. Of course, I am always open for new ideas within Celestia's framework, as long as they are put forward on a competent level...

--I do not share much enthusiasm for getting involved at this "late stage" with a selfmade cross-platform GUI toolkit, like Paulo's. I consider the presently adopted choice much superior, since we can rely on professionally maintained and platform-optimized toolkits via our respective team experts. Most of Celestia's engine is platform-independent, anyhow.

I am not a devoted and very competent GUI coder (in scientific programming, GUI's are mostly irrelevant ). But after all, I coded much of the initial gtk+-1.x Linux GUI and hence, if there are GUI things that /have/ to be incorporated, I can always do it (or send an email to Christophe for help )

--As concerns myself, I am pretty busy with Toti (and in principle also with Chris) since > 4 weeks, developping the galaxy mass-rendering project. With Toti, this pleasantly works on a regular (daily) PM + patch exchange basis, while with Chris I echange mail about 1-2 times/week, sending my views and results, and getting his promises back .

I still believe that (at least in periods of bad weather) Chris is quite interested to go on and seems to enjoy the various challenges of our present joined project.

So I'll prefer to leave things for a (little) while without much change. If stagnation from Chris' side will persist, I shall either leave the community for good or go on developping for myself whatever /I/ consider interesting...

Actually, I am pretty sure that Chris has long stopped following even roughly what's going on in this forum.

Bye Fridger

t00fri wrote:...Chris has selected the original developer team both according to required abilities, but also such that there was a quite uniform agreement about Celestia's design philosophy among the developers!...

I think a quick thread can be done to see what are the actual orientations, but I believe the general philosophy hasn't change so much... Anyhow, what still from the original dev team? I never heard a word of Clint Weisbrod, neither from Bob Ippolito, Grant has left and Chris is on climb...
Soon or later the dev team will have to be refresh with new coders, and so, soon or later a new discussion on goals will be necessary, perhaps only to redefine the actuals ones as a continuum...

In all case, if people decide to create such parallele project, this don't necessary mean the death of the actual developpement. As already said, if the purpose is really to merge again with the main Celestia, there is no fear to have about this "semi" fork...

Now Friger, can you ask Chris for his opinion and his view (projects) on Celestia? (if he can do this personnaly on the forum, better...)

paolo wrote:...The less desirable thing is that many coders work on the same thing creating conflicts or without previously agreeing about important structural decisions...

Is this has already happened? Is this an actual problem?

For the speed of the forum, yes, it's clearely annoying...
The communication "behind the scene" seems quite normal to me (example Fridger/Toti/Chris), but it's true that others coders cannot give their opinions in real time; this should change...

Bye

The SF Developers in alphabetical order (except for myself ):



With exception of Don G, they all have 'write permission' for Celestia at SF, i.e. may submit code.

It is true that Clint Weisbrod, Don Goyette (he was "support manager" ) and Grant have left. Grant, nevertheless keeps updating on and off various data...

On the other hand quite a number of the developers NEVER ever appear in this forum, partly have never been even registered

Alex Burton
Clint Weisbrod
Bob Ippolito
Pat Suwalski

But nevertheless most of them have made or are still making really major code contributions to Celestia!

As you can also see, there are quite a few (potentially) active guys left

Bye Fridger

Fridger as usual you have a non optimistic view and it seems that you are scared about newness. But perhaps its caused by my poor ability in explaining the things.

t00fri wrote:after some thinking, I have decided that I shall NOT join such a venture.

Why so unconditional? It is an opportunity nothing more and nothing less. Surely it is not a venture.

I know my first post sounds a little bit strange and can be misunderstood even for the very poor English.
So first: the purpose of Experimental Celestia should be a cooperative and promotional project alternative to the original Celestia but not in competition with it (so it is not a real fork but a collateral-fork/branch).
Second I've never said that main Celestia's goals and purposes has to be reconsidered. Experimental Celestia should discard automatically the developments that are not considered useful nor coherent with the current purposes.

t00fri wrote:I do not share much enthusiasm for getting involved at this "late stage" with a selfmade cross-platform GUI toolkit, like Paulo's. I consider the presently adopted choice much superior, since we can rely on professionally maintained and platform-optimized toolkits via our respective team experts. Most of Celestia's engine is platform-independent, anyhow.

I've never said that my tiny selfmade cross-platform GUI toolkit has to be used. It is my personal toy for learning C++ and OpenGL. Perhaps in a far future it will be useful to create some kind of GUI for Celestia, but quite so far. I've only proposed to use my Sourceforge CVS space to host Experimental Celestia. But obviously you can ask to Sourceforge to open a brand new project and within a few days you'll have your own space, so you'll be the administrator and you'll select the developers you'll prefer. The only important thing is that there is somewhere a CVS with fresh Celestia code available for the community. So if you discuss with Toti about how to use the octree to optimize galaxy rendering we all will be able to see THE CODE and not only the screenshots.

t00fri wrote:...in scientific programming, GUI's are mostly irrelevant.

So this is the reason why I've said Experimental Celestia should be GLUT only. Cross-platform and GUI independent by definition. GUI part at this stage is irrelevant.

t00fri wrote:So I'll prefer to leave things for a (little) while without much change. If stagnation from Chris' side will persist, I shall either leave the community for good or go on developing for myself whatever /I/ consider interesting...

This should be vary sad. Instead than take the leadership you'll prefer to leave the project. So when the play become really tough you'll abandon the game. I don't believe in these words.

Hi Paulo,

please, don't misunderstand my previous post.

Fact is that I am presently not at all pessimistic. Developing the galaxy project is great fun. My considerations above were in NO way to be viewed as my general extrapolations for the future of Celestia. They were just a collection of reasons why --for now-- I don't feel like joining an "Experimental Celestia" fork. And that independent of that fork's future.

My main goals and considerations at the Celestia front are right now

-- to get Chris somehow back to Celestia coding . His general approach to Celestia suits mine perfectly. His graphical and astrophysical knowledge are outstanding and hence I don't want to miss him in any Celestia crew! I can see no one around here who could replace him in any way. He's always good for a (graphical) surprise and that's a lot of fun to share!

Chris would for SURE not participate at all in "Experimental Celestia" for a number of reasons that presumably I don't have to detail.
What would happen with the present "official" Celestia branch, if people and the rest of the developers would start moving e.g. to a new "experimental" forum at ML? Chris would probably completely decouple...That's what I do not want!

--In my previous coding projects, I have always collaborated with only one or at best two people. So currently, I am quite happy, as concerns my respective activities. If anyone cares to get the patches Toti and I are exchanging, that's of course welcome and no problem whatsoever.

--Suppose we start with 10 or more active new developers in a new "Experimental branch". Clearly, this would generate a huge flow of postings before everybody is in sync, ready to get to work...I neither got the time nor any devotion for this. Just look at some earlier discussions in this board ... We had this sort of chaos at times in the developer list when more people where active at the same time. One spends major parts of the evening answering mail .

Right now, all I care is to get some fascinating piece of work done elegantly and as well as we can manage.

--Who of these newcomers would get write permission?? Who would decide? CVS can get quite chaotic if MANY people start committing code without applying fairly HIGH THRESHOLDS...

--I also don't want to get involved with new discussions about "Addon managers" and other "service coding" for making Celestia easier to use, or internationalization and all that. I am fully aware these things are important for speading Celestia's usage even further.

But... it's just not were my interests lie.

Naturally such areas are however near at hand for new coders who know some allround GUI programming but little astronomy, astrophysics or cosmolgy. So I do expect that "Experimental celestia" would have a certain tendency drifting towards such activities...

+++++++++++++++++++++++
All these points should not serve to influence what other people want to do. I have only listed these issues to illustrate why I am personally not so hot moving in this direction.
+++++++++++++++++++++++

And after all, Paulo, you asked for everybody's /personal/ opinion ...

Bye Fridger

Hi Fridger

The fact is that your opinion has a heavy weight in the community. So if you are not interested in this opportunity I think that no one else will be interested in Experimental Celestia.

Beside this your last post has clarified a little bit the current situation and sounds to me somewhat scaring:

T00fri wrote:-- to get Chris somehow back to Celestia coding . His general approach to Celestia suits mine perfectly. His graphical and astrophysical knowledge are outstanding and hence I don't want to miss him in any Celestia crew! I can see no one around here who could replace him in any way. He's always good for a (graphical) surprise and that's a lot of fun to share!
Chris would for SURE not participate at all in "Experimental Celestia" for a number of reasons that presumably I don't have to detail.
What would happen with the present "official" Celestia branch, if people and the rest of the developers would start moving e.g. to a new "experimental" forum at ML? Chris would probably completely decouple...That's what I do not want!

Do you think that there is a real possibility that Chris won't retrun back on Celestia coding or that he will decouple at all his creature?
If this is true it's really time to fork Celestia. But absolutely I don't think so. I only would like to see development in progress waiting for the Chris' return.

Are a few months that I'm trying to spur on the development of Celestia (e.g. patch contest, this thread), but since now (beside the new galaxy rendering) my tentatives seems ineffective.
I would like to put new fuel to the Celestia development engine, but probably I haven't yet found the right fuel/comburent mixture to start ignition.

To everybody.
Come on share your opinions and please vote! It's a poll! So do not read only: Vote! If you agree in principle but you are not interested in joining post a message!

Hi Paulo,

I know you are trying hard, I am trying hard, each one of us with a somewhat different approach. I certainly appreciate your efforts!

Paolo wrote:Hi Fridger
...
Do you think that there is a real possibility that Chris won't retrun back on Celestia coding or that he will decouple at all his creature?

Well if only I knew that. I have asked Chris repeatedly, whether it was not better to close down. His answer was always that he was only /temporaily/ unable to work on Celestia. However this goes on since January...

Chris wrote:Sorry for the delayed response. I just returned from crevasse rescue practice on Mt. Rainier (another requirement of my mountaineering course.)
...

In the same way I keep getting enthusiastic mails like so

Chris wrote:Fridger,

The shader is coming along nicely. I'll send you some screen shots tonight.
...

Then again NOTHING until one week later

Chris wrote:Still no screenshots, I know . . . I was observing some artifacts near the edges of the galaxies and decided to switch to a new approach. Now, I'm using a volumetric approach . . .

<lots of interesting details>

I'll be gone climbing Mt. Baker over the next couple days, so no more Celestia work until Thursday night.

--Chris

Again NOTHING for more than a week after my next mail...

Well, I don't want to disclose too much of Chris' mails that are meant to be private, but I guess you can see the pattern...

Still I have some hope, ...but I am not sure, really.

Bye Fridger

I think people are forgeting that anything that makes it difficult for novices to participate also makes it difficult for experts to participate.

For example, I have a doctor of philosophy in astrophysics and over a decade of coding experience in scientific numerical coding. I've offered about ten different patches, one of which fixes a long standing bug. To this date, my patches have neither been committed to the main branch nor have I gotten any explanation for why they have not.

From a newbie point of view, the concern about keeping in "competent" people. People with deep expertise in astrophysics or c++ are not going to participate in this project unless you make it easy for them to do so, and anything you do to make it hard for a novice to submit code is also going to make it hard for someone who has a doctorate in theoretical astrophysics.

Also, the discussion on underlying philosophy seems to be highly misplaced. It should be rather obvious to an "expert newbie" what the style conventions and coding conventions are and what the goals of the project are. It isn't.

For example, if the object is astrophysical accuracy, then among my pet peeves is that:

* celestia doesn't handle precession of axes
* celestia doesn't have provision for displaying elliptical and galactic coordinates
* making the oblateness parameter not work just destroys realism for the gas giants.
* galactic distances should be in Kiloparsec or megaparsec.
* quaternion rotations all require conversion to matrices, this shouldn't be necessary
* distances, angle, and times should be encapsulated in classes. (Technically the time unit for celestia should be TDB which differs from UT by a few seconds.)

Also, I got into this project because 1.4.0pre6 would not display acceptably on my C600 laptop. I've fixed this problem.

Now what?

doctorjoe wrote:For example, I have a doctor of philosophy in astrophysics and over a decade of coding experience in scientific numerical coding. I've offered about ten different patches, one of which fixes a long standing bug. To this date, my patches have neither been committed to the main branch nor have I gotten any explanation for why they have not.

I actually have a fairly useless degree in film production and over a decade of USING code (probably written by people with decadeS of coding experience, though) so I'm probably completely useless to you guys.

However, I do happen to be associated with some VERY talented coders who I have been attempting to lure into the realm of Celestia. But I'm wary of offering to do several weeks laundry for them to get them to write a patch for any number of oft-requested features while not having a clear idea of how such patches get committed to the main branch.

So... since I'm here, I second the good doctor's post -- what is the process for getting 3rd party patches into the mix? Personally, I could do the laundry and get the patch for my own selfish uses. But if I'm gonna be washing someone else's socks, a greater good is going to have to come from it. Plus I like to share... if you'll let me.

doctorjoe wrote:I've offered about ten different patches, one of which fixes a long standing bug. To this date, my patches have neither been committed to the main branch nor have I gotten any explanation for why they have not.

I'd suggest you request CVS write access from Chris so you can commit the patches yourself.

- Hank

hank wrote:

doctorjoe wrote:I've offered about ten different patches, one of which fixes a long standing bug. To this date, my patches have neither been committed to the main branch nor have I gotten any explanation for why they have not.

I'd suggest you request CVS write access from Chris so you can commit the patches yourself.

- Hank

Hi Hank
All the question rounds around this point. The official development team is too closed. But I can perfectly agree with this. If Chris wants to keep control over the things, specially when he isn't involved directly, it is the only solution.
So this is another reason to branch the main development using another CVS. So new coders will be able to demonstrate that are skilled enough to be admitted in the official development team. When Chris will trust them enough the whole problem will be solved and Experimental Celestia should even close.

Hey GUYS don't Forget! VOTE!

The poll is running from some days so it is time to make some considerations. When the forums are reachable the speed continues to be unacceptably slow (2 minutes average to display the page!) however the number of visits to the poll thread is high. Despite the number of views the number of votes is low but the percentages are encouraging.
Perhaps the poll is missing of an important choice that should be: "Yes, I'm not able to join but I think it should be useful.". With this choice probably the number of votes should be higher.
Except for the shareable doubts raised by Fridger seems that a number of developers should be interested in joining this project.

A critical mass of developers that are interested in joining is going to be reached so I think that the project should start. I've already activated a new free forum for Experimental Celestia to make some tests http://www.freeforumzone.com/viewForum.aspx?f=68903. It is not phpBB but it is not very different. The speed seems at least the same as the fastest shatters.net despite the presence of advertising banners.

Before kick off the Experimental Celestia I'll wait for a while (at least until all July 3rd) to allow the poll running a bit more and see if someone else will vote, if there will be other positive or negative comments or best of all if Chris will be back!

If Experimental Celestia will start under my coordination (since hosted in CELUI CVS) the first thing I'll do is to write documents for the basic:
- purposes of the project, (essentially the things explained in this thread),
- requirements to be in the team (interests, skills, availability),
- roles for participants,
- rules for the communication between developers (How to communicate, when and what),
- rules for CVS update and commit of code (e.g. to avoid uncontrolled, changes and deviations, only some file will be released for updating other will be locked),
- requirements for documentation of the code changes, of the tools usage, etc.

These documents will be voted and approved, then the project will start. It sounds like politics because it is exactly this: "Politics". IMHO communities requires clear rules, laws and even disciplinary punishments that has to be agreed democratically.

Oh! BTW the last thing is the most important. If you haven't already did it, please VOTE!

After some hours of deep testing I've decided to abandon freeforumszone.com to host Experimental Celestia forums.
Instead I'll focus my attention to:

http://www.freeforumworld.com
http://www.forumforfree.com
http://www.phpbbforfree.com

All of them uses phpBB so it will be more friendly.
I'm trying to understand which has the best performances.
Has anyone experienced one of these?

I've choosen this one:
http://p2.forumforfree.com/index.php?mforum=celui
So now I'll start the configurations.

Paolo,

Did you consider http://www.freebb.com? (No ads.)

Hank

Many thanks Hank!

http://celui.18.freebb.com

Without ads it is better!