Experimental Celestia

[Paolo]

Hi Fridger Glad to see you here again

I strongly agree with DJ as to providing the /full/ Celestia code including GUI. Quite a few modifications e.g. in celestiacore.cpp have interfered in the past with the existing GUI codes. So one HAS to check the GUI stuff in parallel.

Of course, but is celestiacore that has to manage the thing keeping GUI compatibility, so it is unnecessary to put GUI stuff in Experimental Celestia.

Also there are very important missing coding tasks that refer to GUI work: notably an extension of Celestia's browser to be able to profit from all the great multiple star enhancements, including all my 1000's of binary orbits. One urgently needs a browser to find and administrate these binaries!

Also a graphical tool for loading and unloading of specific catalog data is very overdue. Similarly, grid display enhancements and cursor coordinate readouts...All these things need some work at least on the OS-dependent GUI code.

Hey, hey, hey! Big calm here In your third post in this thread you've written that you are not interested in joining because:

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...

And now you are asking for exactly this kind of drifting toward GUI programming!
I agree with you more than perfectly that these are long waited things to do. But you have written a complete to-do list for the official Celestia! I think its too early for this! And for now it is out of the Experimental Celestia scope.

By the way, this "personal page" you requested including photo and lots of personal information, in my view is just too "hot" for the internet! This sort of thing is highly unusual for good reasons. In the Celestia developer team we were "friends" and collaborated well also without all this superfluous info. If there is desire for exchanging private information, everyone can always do that via PM|email.

Some of the fields like photo are signed as not mandatory at all. And moreover like the other documents in the forum it is a draft. If the majority of the members will decide to change them, they will be changed. For now the team its me alone so its good to me. If you will be in the team we should even agree upon something completely different. If we will be at least three it will be possible to vote.

Another thought: knowing Chris and his generic reluctance of working too much on other people's code (patches) my prediction is that a reunification of Experimental Celestia and Offiicial Celestia will become effectively impracticable after a decent amount of new code has been accumulated.

Why not? It?€™s simply a matter of opportunity and of amount of work. Having a lot of people contributing this should not be a problem at all.

Since the only astrophysics expert around seems to be DJ, I am having great doubts, who else might be able to take responsability for the new code's correctness as to physics and astrophysics!?

Beside DoctorJoe? First of all I have an idea. You. Maybe? As DoctorJoe has said if it is difficult to participate for newbies and newcomers with lot of spare time it is more frustrating for scientist with no sparetime. Reducing the difficulties perhaps someone elese will be interested.

My galaxy development work with Toti (and Chris) will be committed /as usual/ into the /official/ Celestia CVS tree, once it has reached a certain level of performance. This also holds for my updates of Titan, Japetus etc stuff. I shall continue in this forum to report in regular intervals on respective progress. Chris definitely plans to release version 1.40 in the near future.

I hope as soon as possible, so I'll get the galaxy code from the official CVS. But I would like to get access to the code, if within a couple of months nothing will happen.

I have been a developer and author of Celestia since > 3 years. Despite the arguments presented, I see no convincing reason to commit my Celestia stuff elsewhere. As soon as I find some time, I shall discuss with Chris about his views in connection with Experimental Celestia and the official one...
Nevertheless, I wish good luck to all involved in the new Celestia forking project.

I would prefer a lot if you will join the project, but for now I accept the most of your comments as positive and encouraging. So thank you.

Do you plan to make the new CVS code downloadable to all people outside the team?

CELUI CVS that hosts Experimental Celestia is reachable as anonymous user exactly like the Celestia CVS one.

Kind regards.

[doctorjoe]

Basically, what I'm looking for is some place to put my celestia patches. I don't care where it is, as long as people can get to them.

Among the patches that I've written are a fix to the foreground orbit problem, a fix to linux gcc 4.0 that makes it possible to view 3ds models, a fix to what seems to be a major bug in the orbit anomaly code, precession of axes etc. etc.

The problem with checking in the code into the main branch is that I've been here for over a month, and no one has given me a procedure for doing so. I do not object to code review, and reworking my patches to make sure that are stable, but no one has contacted me about it.

I had hoped that I would be able to check in the patches into the experimental celestia branch, but to do that I will need that the experimental branch include a full copy of the celestia code. Without this I'll be spending most of my time dealing with messing with build issues. I'm compiling my patches locally with KDE, and without the KDE gui compiled in, it is pretty hard for me to build/debug the branch. Also, a glut-only or developer-only policy makes it impossible for most people to work with the code.

I'd rather not do a fork of the code since I'd rather code. The trouble is that right now, I *am* basically maintaining a fork of the code.

So what do I do?

[Paolo]

Hi DoctorJoe

I'm compiling my patches locally with KDE, and without the KDE gui compiled in, it is pretty hard for me to build/debug the branch. Also, a glut-only or developer-only policy makes it impossible for most people to work with the code.

Sorry me but perhaps I'm missing completely this point. Can you please explain why you and the most of the people can't work with the code using the GLUT-only? If this is a real so big problem, of course I'll change the specification of Experimental-Celestia!!!!!
I've set this specification beacuse using MinGW DevC++ I've created the project file and compiled a fully functional GLUT-Celestia for Windows in less than one hour!
I've set this constrain in order to simplify the things! If this is a real so big obstacle of course we can discuss about this!

[doctorjoe]

Here are some of the issues.....

1) I like using the KDE version of celestia, and would prefer to develop on that platform. Among the things that makes it easier is that it save configuration information and makes it much nicer to access functionality and generally use.

2) More to the point you want people compiling with different UI frameworks to make sure that code doesn't break the UI frameworks. Also multi-platform development generally is good at making sure that you have good clean code, so any thing that encourages people to develop the celestia engine on different platforms should be encouraged.

3) Also, you just want more people running the code.

4) The version in celui is missing all of the build files needed to easily build things on linux. Makefile.am and the like. I like working on scientific code. I hate dealing with Makefiles. The celestia build system is set up so that you can set up a linux build with KDE by typing three or four commands.

5) Also, if you want to prevent a irreversible fork from the main line, it's best to include the gui so that mainline changes in the gui code can be copied over and vice versa. What will happen if you develop on one platform only is that you'll find that when it comes time to remerge, that there are hundreds of little incompatiblilities that will make remerging impossible. If you keep and track GUI changes from the main line code, this is much less likely to happen.

My suggestion is to do a straight copy of the mainline code tree, and then have it as policy that any changes in the mainline code tree changes (including GUI changes) make it into the experimental branch.

[t00fri]

Hi DoctorJoe

I'm compiling my patches locally with KDE, and without the KDE gui compiled in, it is pretty hard for me to build/debug the branch. Also, a glut-only or developer-only policy makes it impossible for most people to work with the code.

Sorry me but perhaps I'm missing completely this point. Can you please explain why you and the most of the people can't work with the code using the GLUT-only? If this is a real so big problem, of course I'll change the specification of Experimental-Celestia!!!!!
I've set this specification beacuse using MinGW DevC++ I've created the project file and compiled a fully functional GLUT-Celestia for Windows in less than one hour!
I've set this constrain in order to simplify the things! If this is a real so big obstacle of course we can discuss about this!

Well the issue does not affect me directly, but I would feel like DJ. My CVS machinery is completely set up and automatized. Why should one want to take this apart? Adapting the autotool configure scripts to new situations can be and indeed was a pain some years ago, when we enlaged the CVS tree to more than one Linux flavor.

You should not forget: There is only one Windows environment, but many Linux flavors! The distributors compete in wildly incorporating patches to their programs and also to their build environment. Notably the portability of autotool related stuff tends to suffer badly...

Many GUI call-back flags are managed in celestiacore.cpp. It is virtually imposible to keep track reliably without actual verification by operating the GUI.

Another example: the extensive testing of our present galaxy project profits a lot e.g. from being able to use the KDE bookmark bar. Toti also uses KDE. I can store there a fairly large set of benchmark galaxies, the rendering performance and correct alignment of which I can quickly check by just hitting those bookmark buttons in turn. Similarly, the KDE GUI has complete programmability of its keys via a graphical interface. I can manage /render different types of deep sky objects /separately/ in the Menue etc...I bet your GLUT version does not (yet) offer an automatic saving of the complete configuration. That's quite essential, however, for testing etc.

We are just USED to all these conveniences during our work...

Since we are doing major code restructurings (e.g. an octree template class serving both the star data base (float) and a new DSO data base (double)), we put temporarily the command line, object selection and labelling out of function for simplicity. The KDE bookmark facility was essential indeed during that exploratory period of octree testing and optimization...

Just some practical examples...

Bye Fridger

[doctorjoe]

Most non-technical users use windows
Most scientists use linux
Most graphic artists use Macs

[Paolo]

Fridger, DoctorJoe

Your statements seems reasonable to me, but implies a real fork of Celestia! Are you sure that you really want this!

Moreover you are both carrying on the KDE environment. I know that it is the most advanced and has many advantages. But the others? MacOSX has a remarkable team of developers right now. So I'm not worried about this. The only one that is uncovered is the mistreated Windows version that right now is already far behind the others, despite it is the most used.
I've choosen GLUT to don't increase these differences. Instead it seems to me that you are pushing for your favorite one.

Beside this it is not a problem for me to include the entire code with all the GUI stuff if the majority of the team wants this. Experimental Celestia will be the thing that their members will want it to be.

But I'm still alone here in the team.
Official invitation was made, and I haven't received any official subscriptions.

In order to be in the team people has to explicitly agree with the defined policy as stated in CELUI forums. I've updated the drafts but I haven't seen any post there with comments. Have I to think that these documents are perfect? So can I consider that everyone agrees with them? Moreover can I consider them definitive and mandatory? If so can I consider that the violators will be excluded from the team and banned from the CELUI forums?

Or is there someone that has something to say?

[t00fri]

Yes me! But this time it's a little less serious...

After a long day and reading your requirements again that request every team member's incognito to be lifted...

I thought it's about time that I get rid of this whole scientist fuzz around myself once for all, and let you all have a glimpse on my real identity:

http://www.maths.monash.edu.au/~johnl/JL%20on%20guard%202.jpg

So you can see now clearly the reason why I am devoted to fighting (mostly with words, of course... )

Enjoy,
Bye DoctorFridger

[Paolo]

DoctorFridger you was very kind and self-ironc in posting this photo! I have appreciated it.
I have to say that I have imagined you being almost exactly like you appear in this photo. Perhaps a little bit older and with a white scientist coat. But exactly like in the photo, serius and tenaciously determined.
I'm starting to think that sooner or later you'll be in the team. And Chris too. Why not? It is only an opportunity to have a lot of fun without having a lot of care about the official distribution.

Returning to the serious things. Have you got any further suggestion?
Does anyone have something to say about Experimental Celestia Policy and rules?

[t00fri]

:D
DoctorFridger you was very kind and self-ironc in posting this photo! I have appreciated it.
I have to say that I have imagined you being almost exactly like you appear in this photo.

Too bad

Let me add one remark, trying to defend my "breathtaking beauty":
On that photo above this unfavorably wide "cross section of mine"




is entirely accidental due to some unexpected gravitational
lensing...So it is definitely virtual!

Bye Fridger

[Paolo]

Too bad..

Very nice! Yesterday I've missed the detail on the bottom, the photo is so big!

I want to say something more about lifting the cyberspace incognito. I would ask to you how many times in these years you had to explain with lack of details who you are, which is your job and why you have credit, in order to get reason of arrogant and silly forum newbie? And of me too indeed
Experimental Celestia members will have their post to explain exactly and with plenty of detail who they are.(e.g. let see the mine.

BTW I have added and updated some other documents in Policy forums.

[doctorjoe]

Not including the full code will increase the chances that you have a permanent fork as you want to minimize the differences between the branch celestia and the main line. If you include the full celestia source you minimize the chances of an accidental incompatibility.

Basically, the chances of a permanent fork are low if in the next three to six months, the current mainline starts accepting patches from the experimental branch. If that doesn't happen in the next six months, then chances are that the experimental branch will replace the main line branch as the experimental branch will have all of the fixes of the main branch, but not vice-versa. If people start developing on the experimental branch and it starts getting better than the mainline branch, so be it. This has happened numerous times with open source projects. (egcs vs. gcc, amavisd, apache)

We'll only run into problems if there are some fundamental rearchitecture that affects one branch but not the other. That isn't likely for at least six months to a year, since celestia basically has a good architecture and any basic redesign will take about six months to one year of discussion.

As far as I can tell there is no active development on the main line CVS branch. There seems to be active development off-line, but it is not being incorporated into the main CVS branch. Right now there are effectively six or seven forks, as people are developing offline. This is a very, very bad situation and if we can reduce the number of forks to two, this will be an improvement.

I don't see why we should be shy about starting a fork of celestia if that is what is necessary. The basic problem is that external patches are not being incorporated into the main branch in a timely fashion. This problem needs to be resolved. I really don't care how it is resolved.

[t00fri]

As far as I can tell there is no active development on the main line CVS
branch. There seems to be active development off-line, but it is not
being incorporated into the main CVS branch. Right now there are
effectively six or seven forks, as people are developing offline. This is
a very, very bad situation and if we can reduce the number of forks to
two, this will be an improvement.

I don't see why we should be shy about starting a fork of celestia if
that is what is necessary. The basic problem is that external patches
are not being incorporated into the main branch in a timely fashion.
This problem needs to be resolved. I really don't care how it is
resolved.

We have always been conservative as to submitting the results to
CVS. Since besides pure coding, Celestia has a very diverse high-tech
astro-mechanical engine, it has turned out safer to first work out a
satisfactory working sub-package.

During the last three years we have always collaborated in small
groups or worked individually in case of major NEW implementations.

Note, I am NOT talking here about /patches/ to existing stuff, but
really new things that need lots of R&D first! That's the stuff
that really pushes Celestia ahead...

Take our mass-rendering project of 10000+ galaxies as a typical case.
A similarly big challenge was our joint work on binary star systems.
Altogether, we had many of such challenges in the past and there will
be more in the future!

For the galaxy project, there was a lot of research needed first to
make sure that all the required data exist. Next came accuracy studies
in numerous plots of distributions and a reliability evaluation through
extensive studies of the literature.

My catalog now involves a combination of 4 different methods of
determining galaxy distances, using many PERL-linked catalogs , for
example. All this work is clearly UNSUITABLE for being dumped right
away to CVS.

Next came lots of experimentation of how to do the best and fastest
rendering of the Hubble-type template galaxies. Chris is working on
ellipticals, Toti started with some very simple /external/ vectorizing
code for spiral templates, which is however, pretty unsatisfactory.
With Runar, I am now trying to scan and exploit the possibilities to
vectorize via Adobe illustrator etc....

Then came all the experimentation with the best octree index variables
and a reorganization of the octree into a template class for both the
star data base and the DSO one.

I hope this clearly illustrates that it would lead to CVS chaos if we
would dump every one of these different attempts to CVS at an early
stage.

What I clearly admit is that the SF Celestia developer list is
essentially dead. That IS a pity. Because previously,
--despite R&D work in small subgroups -- people gave
status reports in that list, such that other interested
developers were able to follow and contribute as well.

Although I tend to have regular contact to Chris it is
now mainly through private email.

Bye Fridger

[t00fri]

Paulo,

I had another reading through your many new "How to behave" documents in your SF site. Honestly, I think every /German/ administration officer would be exceedingly happy with this very large amount of rules and definitions you set up .

In the official Celestia team there was not a single one and YET things worked exceedingly well for > 3 years. The result is visible in form of the present Celestia code .

It's certainly all up to you, and not of my business...
I just thought it might be perhaps useful, if I let you know my personal opinion on all this...

Also to get back to your "personal page":

It appears that many people like(d) to read my ~2800 posts (with > 100000 hits in my own threads) or even sometimes listen to what I have to say. Let me emphasize that this is NOT because I told them all my academic titles, my professional responsabilities ...or even showed them my CV or my photo!

In quite a few cases (including mine) it is probably even better NOT to dwell on these aspects! Everyone, I guess who gets involved with me in a little discussion about theoretical physics in general, math or cosmology/astrophysics in particular will quite soon realize that I know what I am talking about. That's what matters HERE and in any Celestia development team.

Take Grant, for example: he is actually NOT an astrophysicist (as surely many might have thought) , but rather a medical doctor (consultant) by profession. Yet, I have rarely met people with his deep expertise in astro-mechanics, astrophysics, atmospheric physics, optical phenomena etc!

And I guess I am able to judge...It was always very enjoyable for me to get involved with him in any kind of collaboration or scientific dispute.

THAT's what counts!

I am really sorry that he left our community...

Bye Fridger

[doctorjoe]

In cases where there is large scale code changes, one can always create a new CVS branch so that the code is at least visible to outsiders.

But having said that it seems to me that the development process is fairly seriously broken. There is a patch out there that fixes a long standing problem with foreground orbits along with about six or seven other critical bugs. (Such as the fact that rings and virtual textures don't work at all on older cards, or that linux gcc 4.0 doesn't render 3ds models, or that the orbit algorithm is either ineffecient or wrong)

There is no process that I can see for getting that patch into the main line development tree. Note that it is not that I object to code reviews or making changes. The problem is that no one has let me know what changes I have to make in order to get the patches committed, and this is getting me seriously frustrated.

[t00fri]

In cases where there is large scale code changes, one can always create a new CVS branch so that the code is at least visible to outsiders.

But having said that it seems to me that the development process is fairly seriously broken. There is a patch out there that fixes a long standing problem with foreground orbits along with about six or seven other critical bugs. (Such as the fact that rings and virtual textures don't work at all on older cards, or that linux gcc 4.0 doesn't render 3ds models, or that the orbit algorithm is either ineffecient or wrong)

There is no process that I can see for getting that patch into the main line development tree. Note that it is not that I object to code reviews or making changes. The problem is that no one has let me know what changes I have to make in order to get the patches committed, and this is getting me seriously frustrated.

DJ,

yesterday, I wrote a rather "straight" letter to Chris where I asked for his ideas given the present development "dilemma", including the Experimental-Celestia fork. In that letter I have also addressed the present highly unsatisfactury situation with your patches and his NIL reaction to them, so far.

Of course, I have tested your patches quite a while ago, but I cannot give green light for their incorporation into CVS. I could always commit them, but I think that's not the right procedure...

Actually, I have that patch integrated into my running KDE-version of Celestia.

In fact, I was the first one in the team to point out and document that orbit foregound problem, more than a year ago in the developers list. Chris knew why it happened but did not care to do anything about it since...

Bye Fridger

[Paolo]

Hi Fridger

I had another reading through your many new "How to behave" documents in your SF site. Honestly, I think every /German/ administration officer would be exceedingly happy with this very large amount of rules and definitions you set up .

I like /german/ approach in coding of rules indeed.

In the official Celestia team there was not a single one and YET things worked exceedingly well for > 3 years. The result is visible in form of the present Celestia code .

You are right, but all these rules are strongly recommended by SourceForge documents. You have an example in the CVS usage thread.

It's certainly all up to you, and not of my business...
I just thought it might be perhaps useful, if I let you know my personal opinion on all this...

Since up to now I don't know where Experimental Celestia will lead and what kind of people will be involved, I prefer to add clear rules to the team behaviour (obviously if any team will be created, I'm still alone here ) So everyone that is interested will know exactly what he is facing and what will happen joining the team.

Also to get back to your "personal page":

It appears that many people like(d) to read my ~2800 posts (with > 100000 hits in my own threads) or even sometimes listen to what I have to say. Let me emphasize that this is NOT because I told them all my academic titles, my professional responsabilities ...or even showed them my CV or my photo!

In quite a few cases (including mine) it is probably even better NOT to dwell on these aspects! Everyone, I guess who gets involved with me in a little discussion about theoretical physics in general, math or cosmology/astrophysics in particular will quite soon realize that I know what I am talking about. That's what matters HERE and in any Celestia development team.

And so you did it again. Its wrong to continue to repeat these things. Trust me! Its annoying and seems arrogant even if you think to be kind doing so! This is noise in the communication process. [Edited]
BTW not very detailed info (CV like) is present even in this forum in your user profile. Have you ever asked yourself why? Perhaps because phpBB is the sum of years of experiences of many-many communication specialists in managing bulleting board systems. If they included some CV like fields is because are really useful. I've only extended a bit this concept.

Take Grant, for example: he is actually NOT an astrophysicist (as surely many might have thought) , but rather a medical doctor (consultant) by profession. Yet, I have rarely met people with his deep expertise in astro-mechanics, astrophysics, atmospheric physics, optical phenomena etc!

And I guess I am able to judge...It was always very enjoyable for me to get involved with him in any kind of collaboration or scientific dispute.

THAT's what counts!

I am really sorry that he left our community...

I think that we all are sad about his departure. Perhaps if a fresh wind will renew Celestia he will return back. But have you asked why he is missing you so much? Because he, like Chris, Selden, Christophe (and some others participating these forums), has a very very kind communication style, that allows to go straight to the point. So working with them is easy, enjoyable and satisfactoring because the communication process is really effective.

Kind regards

[Paolo]

Hi doctorjoe

There is no process that I can see for getting that patch into the main line development tree. Note that it is not that I object to code reviews or making changes. The problem is that no one has let me know what changes I have to make in order to get the patches committed, and this is getting me seriously frustrated.

I have to say that one of the most important reasons why I have decided to start this thread and be involved in this /venture/ is to solve at least partially your comprehensible frustration!
I've followed the Celestia Development Mailing just after having registered in these forums. I've red all of your posts.
If this whole thread and Experimental Celestia will end up in a soap bubble explosion but official Celestia will restart including you and some other developers I will be happy the same! Experimental Celestia is not for my personal glory but for the future of Celestia!

Kind regards.

[hank]

There is no process that I can see for getting that patch into the main line development tree. Note that it is not that I object to code reviews or making changes. The problem is that no one has let me know what changes I have to make in order to get the patches committed, and this is getting me seriously frustrated.

Of course, I have tested your patches quite a while ago, but I cannot give green light for their incorporation into CVS. I could always commit them, but I think that's not the right procedure...

Fridger,

Why is that not the right procdure? What is the right procedure?

- Hank

[t00fri]

Hi Fridger

...

Also to get back to your "personal page":

It appears that many people like(d) to read my ~2800 posts (with > 100000 hits in my own threads) or even sometimes listen to what I have to say. Let me emphasize that this is NOT because I told them all my academic titles, my professional responsabilities ...or even showed them my CV or my photo!

In quite a few cases (including mine) it is probably even better NOT to dwell on these aspects! Everyone, I guess who gets involved with me in a little discussion about theoretical physics in general, math or cosmology/astrophysics in particular will quite soon realize that I know what I am talking about. That's what matters HERE and in any Celestia development team.

And so you did it again. Its wrong to continue to repeat these things. Trust me! Its annoying and seems arrogant even if you think to be kind doing so! This is noise in the communication process.
...

Paulo,

just so you know: you completely misunderstood the purpose of my paragraph above.

I am going to get peacefully out of this discussion NOW...

Good luck,

Bye Fridger