What would be needed to go to a full 1.5.0 release?

[ElChristou]

...A "Celestial Matters" fork of Celestia? That's up to the Celestial Matters team to decide (I only suggest them because that is where active development, the required knowledge, and the commitment to Celestia lies), but without the community supporting it, Celestia is on perilous grounds...

Chaos, not so fast; it's not CM goal to become a fork from Celestia. Actually Fridger do have his Cosmo project, but for now it's not really the continuity of Celestia. Now concerning Runar and I we don't have the skills to pretend working actively on such project (by actively I mean at code level).

For now all the workforce is here. Of course a fork is possible, but what if we do the same errors? Perso I would like a big brainstorming first to define a structure which will prevent any trouble. Something that would gives one the liberty to take a big break for example without creating the problem we are dealing with...

[ajtribick]

Do you know guys some examples that could serve us to build some thinking upon?

First example that comes to my mind is Oolite (a game based on original Elite). The program is copyrighted to the original developer and contributors but released under the GPL, as stated in their license file.

The original developer has since left the project and development is now led by someone else.

However I have only been involved in the community since quite a while after the transition so I don't know what happened back then.

[ElChristou]

Do you know guys some examples that could serve us to build some thinking upon?

Stellarium ?
NeoOffice ?

Stellarium is a one man show if I'm not wrong; NeoOffice I don't know... What would be nice is to find examples where the copyright is shared in a way or another...

[ajtribick]

Chaos, not so fast; it's not CM goal to become a fork from Celestia. Actually Fridger do have his Cosmo project, but for now it's not really the continuity of Celestia. Now concerning Runar and I we don't have the skills to pretend working actively on such project (by actively I mean at code level).

I had a feeling that would be the response to that suggestion.

Ah well.

[hank]

I don't really think it's necessary to fork the code as long as Chris does not object to what we're doing and authorized developers are available to do the commits at Sourceforge. What is necessary is finding enough volunteers to do the work.

- Hank

[Fenerit]

Who, amongst you, by coding, is able to add features to Celestia? If someone know how to add a feature request, he may do it. So think the user when see an open project. Look at Sourceforge and in particulary beyond the first page of whatever project. The majority of the active projects (that is, the projects which are up to date) didn't go ahead the second. The developer compile and share the new files with that feature and so on the others with the owns; or not? If no one is able to do this he is not a developer, as me. So, you know the rules.

[dirkpitt]

Adding new features and committing them to CVS sounds easier than you think. However, no good software engineering project of this size would allow you to commit such changes without a good amount of super review and testing.

The complication here is, if I were to submit a patch against the parts of Celestia that Chris originally wrote, Chris is the best person to review such a patch. Other devs would have to first read Chris' code and try to understand it (bear in mind Chris has several times more graphics industry experience than the rest of us), and then decide whether to approve such a patch, or if not, suggest a better patch. This may be very difficult, especially with architectural changes.

Architectural changes deal with problems (bugs, or lack of a certain feature, etc) that lie so high up that a complete code rewrite may be required. In such cases, feedback from the original architect (Chris) is often required in order to fix the issue.

The reverse is also true - Chris could turn this project into a 1 man show, but since there are large swaths of the code (especially the UI) that he didn't write, he'll eventually have a tough time dealing without us.

So either way, it'd be a much better off for all of us if Chris were to come back into the picture.

[Fenerit]

Adding new features and committing them to CVS sounds easier than you think. However, no good software engineering project of this size would allow you to commit such changes without a good amount of super review and testing.

The complication here is, if I were to submit a patch against the parts of Celestia that Chris originally wrote, Chris is the best person to review such a patch. Other devs would have to first read Chris' code and try to understand it (bear in mind Chris has several times more graphics industry experience than the rest of us), and then decide whether to approve such a patch, or if not, suggest a better patch. This may be very difficult, especially with architectural changes.

...

Sure. Meantime, though, the user can test that changes; seem to me is here the subject of this thread. What I want to say is that when I've seen Vincent wished to add the sound, for example, he did it. What file was? The main file? Ok, what can he do to the textures? Nothing. To Add-ons? Nothing; if only: better management of 3d shaders. And so on; even with other files. Being all already happened, expecially for what concern the Mac. Probably is not the place here on Shatters as well as CelestialMatters or Sourceforge, but what say Hank appear from the fact that at least here, the "Purgatory" has the number of topic more great than they of the "Developer Talk" or the "Celestia Scripting"; when should be the contrary

Moreover, there is even the possibility of an end of whatsoever project due to the basic fact that the program to be complete. If the user in writing a letter use a simple program, the developers can do evolving that program in a big suite. Well, go ahead; meanwhile the former will change the program because he didn't need of a supreme program to write few lines of text; and he'll search a small and complete (for his task) program.

But... What happen?... Incredible; I've seen the red presence of Chris amongst the registered users!

[ElChristou]

...So either way, it'd be a much better off for all of us if Chris were to come back into the picture.

Of course, Chris is a master piece here. We are all thankful for this great soft, we do want him to be back because we perfectly know how important he is. Now our problem is that Chris is just not here and apparently don't care about what we have to say.
So again the problem is simple, or we bear his attitude without a word and Celestia will survive as long as possible in this present dev formula, or we decide to find another solution for a more efficient dev.

Perso I think that "waiting to see" is illusory and Fridger is probably right about Chris motivations. If this is the case it's a really serious problem because then Celestia is no more independent. Someone paying will decide what must be done or not and we, we will be only a bunch of beta tester working for free. Again, perso I won't follow this road...

[hank]

So either way, it'd be a much better off for all of us if Chris were to come back into the picture.

Certainly it would be a big help (a cosmic-scale understatement) if Chris were able to participate. But obviously we can't depend on that.

Of course all code changes would need to be carefully reviewed and thoroughly tested. The developer team would need to become much less compartmentalized. More than one developer would need to become familiar with each area of the code. And more developers would need to be recruited.

- Hank

[ElChristou]

I don't really think it's necessary to fork the code as long as Chris does not object to what we're doing and authorized developers are available to do the commits at Sourceforge. What is necessary is finding enough volunteers to do the work.

- Hank

Let's say we find the volunteers to do such debugging to release 1.5, then, what next? We will have a few more guys waiting with us Chris for 1.5.1?

[hank]

Let's say we find the volunteers to do such debugging to release 1.5, then, what next? We will have a few more guys waiting with us Chris for 1.5.1?

The same team of volunteers that completes 1.5.0 would proceed to 1.5.1 without hesitation. The intent is to establish a continuing, sustainable development process. Otherwise it's not worth the effort.

- Hank

[Fenerit]

Let's say we find the volunteers to do such debugging to release 1.5, then, what next? We will have a few more guys waiting with us Chris for 1.5.1?

The same team of volunteers that completes 1.5.0 would proceed to 1.5.1 without hesitation. The intent is to establish a continuing, sustainable development process. Otherwise it's not worth the effort.

- Hank

Right. I'll quit my presence here saying that for the texture makers or for the add-ons makers is irrelevant to test the new versions during the develop of a texture or of an add-on. They'll act on the last stable version once tested and they'll test the new apart. In this case the "stable" version is as equivalent as if the programs were "complete". Try to forget Chris from your discussion, though; because when he returns in the discourses, the discourses cannot go ahead and nothing new can arise in matter of facts.

[MKruer]

Sound to me that there is an unofficial game plan to fix only existing bugs and to release a 1.5.0 build.

Lets keep it simple.

Cham had 15 or so on the first page, some of them have been committed to the bug tracker already. http://celestiaproject.net/forum/viewtopic.php?t=11508 the rest are at https://sourceforge.net/tracker/?atid=1 ... unc=browse

If we can all accomplish this by say end of the year, I think that we would be in a good position. If Chris comes back during that time, great, if not, we will not have deviated from his plan, we will have only fill in the missing part or bugs that he left with.

Edit, all Cham bugs are in the Bug tracker under a single topic.
https://sourceforge.net/tracker/index.p ... tid=121302

[ElChristou]

Let's say we find the volunteers to do such debugging to release 1.5, then, what next? We will have a few more guys waiting with us Chris for 1.5.1?

The same team of volunteers that completes 1.5.0 would proceed to 1.5.1 without hesitation. The intent is to establish a continuing, sustainable development process. Otherwise it's not worth the effort.

- Hank

But Hank, do you think people will spend their time on debugging and dev knowing that the day Chris will come back he can say "Hep, this, this and that were not in my plan, this here should have been done this way, and this one, no really not elegant at all. I must remove those last change to be able to code what I had in mind..."?

We can try to make him come back with a 1.5 release, but do you really think we can go this way for a long time?

[hank]

But Hank, do you think people will spend their time on debugging and dev knowing that the day Chris will come back he can say "Hep, this, this and that were not in my plan, this here should have been done this way, and this one, no really not elegant at all. I must remove those last change to be able to code what I had in mind..."?

We can try to make him come back with a 1.5 release, but do you really think we can go this way for a long time?

The purpose is not to "try to make Chris come back" with a 1.5 release, but to establish a sustainable development process for continuing Celestia development.

If Chris is able to resume active involvement with Celestia at some point, as I hope he will, he will hopefully recognize the value of the new sustainable development process, and will choose to participate in that effort. Changes proposed by him would be subject to community review and approval just like any others, although of course his ideas would certainly carry great weight. If that is not acceptable to him, and he prefers to continue on his own, he is free to go his own way. In that case we would have to fork the code. But I really hope that won't happen.

- Hank

[t00fri]

Changes proposed by him would be subject to community review and approval just like any others, although of course his ideas would certainly carry great weight. If that is not acceptable to him, and he prefers to continue on his own, he is free to go his own way. In that case we would have to fork the code. But I really hope that won't happen.

- Hank

All this must be a joke.

What do you mean concisely with "community review"?

Are you really saying that Chris should have his future OpenGL code reviewed by someone here, knowing just a fraction of what he knows in the area?

It's about like having some of my physics/astrophysics/cosmology arguments reviewed by Chris or by you perhaps? Chris owns Celestia's Copyright, full stop.

Or should we adopt the reviewer scheme of E-Bay, where on average NOONE understands anything about the matter he/she is reviewing about?

Bye Fridger

[hank]

All this must be a joke.

What do you mean concisely with "community review"?

Are you really saying that Chris should have his future OpenGL code reviewed by someone here, knowing just a fraction of what he knows in the area?

It's about like having some of my physics/astrophysics/cosmology arguments reviewed by Chris or by you perhaps? ;-) Chris owns Celestia's Copyright, full stop.

Or should we adopt the reviewer scheme of E-Bay, where on average NOONE understands anything about the matter he/she is reviewing about?

Bye Fridger

Fridger,

Peer reviews are always useful. Even experts sometimes make mistakes or overlook things. And the reviewers learn as well. This is very important. The fact that currently there is no one here other than Chris with substantial OpenGL competence is one of our biggest problems. If other developers had become familiar with Chris's code through code reviews, we'd be in a much better position to carry on in his absence.

- Hank

[t00fri]

All this must be a joke.

What do you mean concisely with "community review"?

Are you really saying that Chris should have his future OpenGL code reviewed by someone here, knowing just a fraction of what he knows in the area?

It's about like having some of my physics/astrophysics/cosmology arguments reviewed by Chris or by you perhaps? Chris owns Celestia's Copyright, full stop.

Or should we adopt the reviewer scheme of E-Bay, where on average NOONE understands anything about the matter he/she is reviewing about?

Bye Fridger

Fridger,

Peer reviews are always useful. Even experts sometimes make mistakes or overlook things. And the reviewers learn as well. This is very important. The fact that currently there is no one here other than Chris with substantial OpenGL competence is one of our biggest problems. If other developers had become familiar with Chris's code through code reviews, we'd be in a much better position to carry on in his absence.

- Hank

Hank,

as an experienced scientist I live with peer reviews of publications and research results since decades. So I know very well about this scheme, without which science could not prevail!

But unlike this community, peer reviewing in science involves scientists with PROVEN knowledge in their field of research.

Do you have some personal experience with professional level peer reviewing? If so let me know. Since you make such general statements above , it seems you must have plenty of experience?

Your general "sustained community" scenario outlined above gives no clue about how to assert the competence of the reviewers in relation to that of the people to be reviewed.

Sorry, but all this talking appears very "blue eyed" to me.

Bye Fridger

[hank]

as an experienced scientist I live with peer reviews of publications and research results since decades. So I know very well about this scheme, without which science could not prevail!

But unlike this community, peer reviewing in science involves scientists with PROVEN knowledge in their field of research.

Do you have some personal experience with professional level peer reviewing? If so let me know. Since you make such general statements above , it seems you must have plenty of experience?

Your general "sustained community" scenario outlined above gives no clue about how to assert the competence of the reviewers in relation to that of the people to be reviewed.

Sorry, but all this talking appears very "blue eyed" to me.

Bye Fridger

Fridger,

In software development, peer reviews are perhaps less competitive and more collaborative than in your field of scientific research. The reviewers and the people whose work is being reviewed are members of a team who share a common objective in assuring quality and improving productivity. They aren't generally focused on asserting their competence or proving their knowledge. Those things are usually self-evident anyway. Of course less experienced team members generally defer to and learn from those with more experience, and more experienced team members share their knowledge, to mutual benefit.

- Hank