Celestia 1.4 final?

[hank]

I think it would be very nice to have a new "official" release of Celestia (1.4, or should it be 2.0?). It would be great to have it out before the end of the year. So I'm wondering, what would we need to do to make that happen?

Obviously we'll need Chris to be involved for the official release, and we have no idea when (or if) he'll be able to participate. But it's not my purpose to revisit that issue now. Rather, I'm asking what tasks remain to be done to complete the new release?

I take it Toti and Fridger still have some work to do to finish up the new galaxy rendering code. Is there still work to do on any of the other new features? Any critical bugs to be fixed? Performance issues?

Once coding is complete, we'd need to integrate the code, build versions for each platform, and test, test, test. (I guess it would also be nice to have some up-to-date documentation.)

Are there any volunteers out there who could help make this happen? Or am I just dreaming?

- Hank

[Cham]

I'm all ready to test, test, test...

I'm waiting the OS X version of the new Ceelestia made by Fridger and Toti.

New features ? Geez, I have TONS to suggest !

-BumpMaps on asteroids,
-Real drop shadows of models on planets
-fix the horizon bugs on planets
- etc...

[hank]

New features ? Geez, I have TONS to suggest !

I think at this point getting a new release out is more important than adding new features. Unless they'd be trivial to implement, any new features beyond those already in work must wait for the following release.

- Hank

[Toti]

Certainly a 1.4.0 final release would be great. Really lots of features have been accumulating since 1.3.2.
There are a few other things in our immediate FT-TODO apart from the rendering code:
--Extensible binary DSO catalog (perhaps with cross-indexing) to speed up Celestia start up.
--Importance weighting of labels for large numbers of objects.

[t00fri]

Certainly a 1.4.0 final release would be great. Really lots of features have been accumulating since 1.3.2.
There are a few other things in our immediate FT-TODO apart from the rendering code:
--Extensible binary DSO catalog (perhaps with cross-indexing) to speed up Celestia start up.
--Importance weighting of labels for large numbers of objects.

And certainly to extend the scheme to

-- clusters and nebulae!
(-- asteroids + comets )

A sky display is not much /useful/ and very systematic if it only knows about galaxies but not of clusters and nebulae...

++++++++++++++++++
My preference right now would be to first get FT to some more advanced stage before trying to invoke Chris again. Remember what a lengthy affair it was before he finally had DrJoes patches implemented.
The FT code modifications are vastly larger and include significant modifications also to the code /structure/.
++++++++++++++++++

I am just tired of :" I will implement your code this weekend" and then NOTHING happening for four subsequent months or so!

Sorry for two days of additional delay with releasing FT1.1. I seriously hope to release it tonight.

The benefit of the delay is however now:
------------------------------------------------
-- a complete installer version for Windows that will not interfere with the default Celestia installation. FT1.1 will be installed in a directory called Celestia FT1.1.

--I found and eliminated the reason for the CTRL+V bug in Windows.
Chris' *.iss file lacked the ell_galaxy_arb.vp shader! While it was not used yet, the code tried to load it nevertheless and quit with a segfault if it was not present in the shader directory. So everyone playing with FT1 needs to add this shader by hand to 'shaders' and then the bug is gone! Everyone can download it from CVS.

As of FT1.1 everything is taken care of in this respect!

--Chris had forgotten about my two binary orbit files (visualbins.stc and spectbins.stc) in his installer *.iss file for Windows. So these were lacking accidentally in the recent 1.4.0pre versions. Since I now have included them, you will also get hundreds of new binary orbit data to play with...

Bye Fridger

[Boux]

t00fri, Toti,
Please don't hurry and roll-out prematurely, I am sure this release will be a major landmark
Will there be a Linux binary or will you provide a tarball for penguins to compile by themselves?
I have another question (be assured there is no rant, no second-thought or hidden agenda in this question):
Aren't you /de facto/ forking Celestia?

[t00fri]

t00fri, Toti,
Please don't hurry and roll-out prematurely, I am sure this release will be a major landmark
Will there be a Linux binary or will you provide a tarball for penguins to compile by themselves?
I have another question (be assured there is no rant, no second-thought or hidden agenda in this question):
Aren't you /de facto/ forking Celestia?

There will be NO binary for Penguins. But I made sure that FT1.1 compiles cleanly for all three supported Linux GUI environments:

Linux-KDE
Linux-Gnome
Linux-Gtk

As to the forking issue:

I have been fighting fiercly against ongoing forking efforts for many months, while I still had hope to get Chris back to Celestia development! Meanwhile, I am unfortunately convinced that effectively "the soup is cold", although -- if asked-- Chris might well deny.

So I would be most happy, if we could find with Chris an acceptable procedure to keep Celestia unified. In my view, presently, the best strategy seems to be to come up soon with an excellent, matured Celestia-FT release, such that Chris' only practical option would be to declare it official.

Bye Fridger

[Guckytos]

What i would really appreciate in a new official version is that it enables you to reload the stc, ssc, dsc-files.
So that if you see that something doesn't look right, you change it in an editor, and just reload, the corresponding file-type, instead of ending and restarting Celestia, the whole time.
It already works in the unoffcial version.

Well, just some personal experience, but nonetheless it would be good in an official version.

Regards,

Guckytos

[Boux]

... But I made sure that FT1.1 compiles cleanly for all three supported Linux GUI environments:
Linux-KDE
Linux-Gnome
Linux-Gtk

Thanks

[Boux]

As to the forking issue:
I have been fighting fiercly against ongoing forking efforts for many months, while I still had hope to get Chris back to Celestia development! Meanwhile, I am unfortunately convinced that effectively "the soup is cold", although -- if asked-- Chris might well deny.
So I would be most happy, if we could find with Chris an acceptable procedure to keep Celestia unified. In my view, presently, the best strategy seems to be to come up soon with an excellent, matured Celestia-FT release, such that Chris' only practical option would be to declare it official.

I see... in diplomatic terms, you mean it is not a fork but a /shortcut/

[hank]

It's been more than a year (22 Aug 2004) since the last official release of Celestia (1.3.2). And it's coming up on a year (11 Nov 2004) since the last official prerelease (1.4.0pre6). In the meantime a number of unofficial versions and patches have appeared in various places.

This isn't so much of a problem for experimenters who are willing and able to scrounge around for unofficial developments and install or build unofficial versions. But it is not so good for more casual or less technical users and for addon developers who want to address that larger audience. That's why I'd like to see a new official version.

I don't think we can rely on Chris to drive the process as he did in the past. But I'm hoping that if we could advance things to the point where we have an "excellent, matured" codebase (to use Fridger words) ready for release, then Chris would be willing and able to assist in making it official.

Although there are many additions and corrections that could be made to Celestia (and hopefully they will all be added eventually), if we don't draw the line somewhere we will never complete a new release. So at this point I believe we should be thinking in terms of what can be left out rather than what can be put in. Personally I feel the binary star orbit and galaxy rendering additions are more than enough to justify a new official release, but if some other things are absolutely essential, let's identify them here.

- Hank

[Cham]

I agree. WE MUST HAVE A NEW OFFICIAL VERSION NOW ! After a year, it is URGENT ! There's a lot of confusion out there about what is Celestia, in large part because of all the unofficial versions.

Chris, PLEEEEEASE ! Say something, for Chris sake !

[hank]

I agree. WE MUST HAVE A NEW OFFICIAL VERSION NOW ! After a year, it is URGENT ! There's a lot of confusion out there about what is Celestia, in large part because of all the unofficial versions.

Chris, PLEEEEEASE ! Say something, for Chris sake !

We'd all be very happy to see Chris return to active involvement with Celestia at any time. But let's not waste our energy bugging him about it until things are ready to go. Please, let's limit this thread to discussing the specific things that we need to do to get the release in shape.

- Hank

[t00fri]

I agree. WE MUST HAVE A NEW OFFICIAL VERSION NOW ! After a year, it is URGENT ! There's a lot of confusion out there about what is Celestia, in large part because of all the unofficial versions.

Chris, PLEEEEEASE ! Say something, for Chris sake !

We'd all be very happy to see Chris return to active involvement with Celestia at any time. But let's not waste our energy bugging him about it until things are ready to go. Please, let's limit this thread to discussing the specific things that we need to do to get the release in shape.

- Hank

After all this frustrating experience of the last 6 months, say, I am deeply convinced that Chris has since long completely different activities for his spare time. There is just some leftover "nostalgy" for Celestia. That's all.

+++++++++++++++
If we do not find a way to make Chris officially transfer the CVS decision making to someone else, NOTHING is gained, even if we manage to get his OK for Celestia FT to become official! At the same time other issues related to Copyrights etc must also be discussed.
++++++++++++++

We need a change of procedure, otherwise development REMAINS blocked and I shall leave Celestia. It's as simple as that. I am simply NOT willing to wait again for months each time before our code will be implemented into the CVS archive. NO WAY.

The present effort together with Toti to get Celestia FT somehow going, was for me the LAST try. Definitely the last one...

In view of all these delicate issues ahead, one has to contemplate most carefully what the best timing would be for entering the discussion about unifying the codes and releasing a new "official" version.

Bye Fridger

[Malenfant]

We need a change of procedure, otherwise development REMAINS blocked and I shall leave Celestia. It's as simple as that. I am simply NOT willing to wait again for months each time before our code will be implemented into the CVS archive. NO WAY.

I don't understand how this all works, but could't you just start a new CVS archive with your new code? Call it CelestiaFT 1.0 instead of Celestia 1.4.0 or something, but that way you wouldn't have to overwrite anything that Chris does, and you could have full control over what goes in and out of the code. Don't ask for Chris' permission (because you'd die of old age before you got it, probably) - just set up a separate archive, acknowledge somewhere that it's ultimately based on Chris' version of Celestia and then carry on development. AFAICT there's nothing stopping you from doing this - Celestia is open source and anyone has the right to continue to develop it. It seems to be only out of politeness that you want to give Chris a chance to respond here, but he's evidently not interested. My feeling is that he's had his chance and not taken it, so it's up to you now to move on as you feel best.

Either way, the current situation is clearly not acceptable for anyone.

[t00fri]

We need a change of procedure, otherwise development REMAINS blocked and I shall leave Celestia. It's as simple as that. I am simply NOT willing to wait again for months each time before our code will be implemented into the CVS archive. NO WAY.

I don't understand how this all works, but could't you just start a new CVS archive with your new code? Call it CelestiaFT 1.0 instead of Celestia 1.4.0 or something, but that way you wouldn't have to overwrite anything that Chris does, and you could have full control over what goes in and out of the code. Don't ask for Chris' permission (because you'd die of old age before you got it, probably) - just set up a separate archive, acknowledge somewhere that it's ultimately based on Chris' version of Celestia and then carry on development. AFAICT there's nothing stopping you from doing this - Celestia is open source and anyone has the right to continue to develop it. It seems to be only out of politeness that you want to give Chris a chance to respond here, but he's evidently not interested. My feeling is that he's had his chance and not taken it, so it's up to you now to move on as you feel best.

Either way, the current situation is clearly not acceptable for anyone.

Malenfant,

what you are proposing is what is called "code forking".

Of course this is a straightforward possibility. Yet you forget one little thing:

1) I have been working closely with Chris for four years and
we even have a little bit of a personal relation . I
just have little motivation to hurt his feelings and also I
respect his know-how very much.

2) I think he is just hard to replace, whence I have been
trying /very hard/ to get him back to Celestia
development.

Since (2) almost definitely failed, there is only (1) left that makes me hesitate!

But be assured a solution will arise in one way or another...

Bye Fridger

[Malenfant]

what you are proposing is what is called "code forking".

Oh, that's what forking means is it? right .

1) I have been working closely with Chris for four years and
we even have a little bit of a personal relation . I
just have little motivation to hurt his feelings and also I
respect his know-how very much.

True, but I think you have to consider that we haven't had any serious input from him for over a year now. If I was him I'd probably understand why you just gave up and moved on with a fork. Either way, his material on the original Celestia tree wouldn't be written over or anything and in the unlikely event that he returns then maybe the two versions could even be reintegrated. But until that time (which may not happen) clearly something has to happen to move the project forward. FT1 is a very promising step in the right direction and is better than sitting around waiting and getting frustrated. At least now something is happening.

Put it another way - would you prefer that the project stagnate completely and that everyone just gave up on it for fear of hurting Chris' feelings, or would you prefer that the project move on with FT and then other developers can start chipping in and feeling optimistic about it again? I think it's a cycle really - if a potential developer sees what looks like a dead or stalled project, they're not going to be likely to want to contribute, but if he sees an active project then I think he'll be more likely to take part.

2) I think he is just hard to replace, whence I have been
trying /very hard/ to get him back to Celestia
development.

And yet as you say that's failed, through no fault of your own. I think you and others here have tried enough to persuade him to return - it's clear that he isn't. Someone has got to take a deep breath and cut the cords that bind and move the project on into where it's actively being developed.

And to be honest, you and Toti have made these changes without his input. While perhaps the main rendering core of Celestia might not be changed because that's Chris' domain, there's still a lot of development that can continue around that.

[t00fri]

Malenfant,

most of what you wrote is somehow right. Still I know the people involved for quite a while and follow my own strategy. Just be patient still a little while. We are getting there

And meanwhile we are not sleeping either.


Bye Fridger

[wcomer]

If you fork the project then there are two outcomes:
1) Chris is a grown adult and handles it with grace and maturity.
2) Chris isn't and takes it poorly.

In the first case, the fork is the right decision and you've lost nothing.
In the second case, the fork is still the right decision because that sort of attitude is counter productive. And you would have gained something by dropping the dead weight and bad attitude.

I strongly believe that Chris would handle the situation very well and think the second outcome is unlikely. But either way, forking the project is the sensible choice.

cheers,
Walton

[rthorvald]

In the first case, the fork is the right decision and you've lost nothing. In the second case, the fork is still the right decision because that sort of attitude is counter productive.

This isn??t simple... Just think for a moment on what an enormous investment Celestia must be to Chris. Even if he is tired of the project, that does not go away. I think one will have to be very tactful here for the "official" development to be transferred to someone else. Maybe one idea would be to persuade him to fork Celestia himself, and then hand over the "new" Celestia to the other devs. Then he could keep his own "original" Celestia, while the "new" continues as an official Celestia Mark II or something.

There are other things to consider, too:
You cannot seriously believe Chris will hand over Celestia to someone else, and continue to host the project indefinitely... A new infrastucture for the devs might become neccecary (CVS, hosting, comm). Not to mention preserving the enormous knowledge base of the forums...

I certainly understand the hesitation!

BTW, i volunteer to help out with such mundane tasks if it comes to that.

-rthorvald