Board index FlightGear Development New features

Could this be the a new goal of FGFS ?

Discussion and requests for new features. Please note that FlightGear developers are volunteers and may or may not be able to consider these requests.

Re: Could this be the a new goal of FGFS ?

Postby Thorsten » Mon Mar 14, 2016 5:53 am

I always had the impression that Rembrandt was abandoned too early, when it was still able to show its true potential. Sure it's expensive at this level of CPU and can lower the frame rate (not sure though if with better management of multiple CPUs can improve it by changing, in a proper mode, the code).


I think you fail to realize that doing multiple passes over the scenery and over a texture buffer is primarily expensive on the GPU fragment pipeline.

Not only that, but the note of X-Plane, says the approach is "expensive" (quality costs is not free ...), but in anticipation of the improvement in the quality of CPU-GPU, will become more and more usable.


Maybe so, but I've also a list of features lined up I'd like to code then which I still think are more important than what deferred rendering provides - so consider that improvement already part of the ALS plan.

every user should have the freedom to exploit his system as he wishes with the frame-rate they want


I don't think anyone is keeping you from anything. In case this needs to be spelled out - FG is OpenSource. You can not only access the effects and shaders freely, you can access all the rendering source code. You can change the rendering in any way you like.

Just go ahead and use whatever rendering scheme you want.

(On the other hand, every developer is free to work on whatever he wishes to do with his time - so don't expect me to code things I consider not worth the effort...)

I've already said when I pointed out that it would be useful to make a ALS-Rembrandt merge.


Yes, it's a fictional thing that primarily people who never have done any rendering code think great and none of the people who do FG rendering code have touched.

Just based on what expertise would you say this is useful? Do you have an idea of the fragment pipeline load of deferred rendering? Of ALS? A rough number of how many instructions per pixel ALS executes for terrain vs. sky? How many Rembrandt does? How many passes over the scene Rembrandt computes?

It's so cheap to stand on a lofty position and tell everyone what would be useful, knowing you'll never code it anyway...
Thorsten
 
Posts: 11378
Joined: Mon Nov 02, 2009 8:33 am

Re: Could this be the a new goal of FGFS ?

Postby Thorsten » Mon Mar 14, 2016 6:10 am

I know this story, but unfortunately does not define a clear structure of the files that combine to create the effects. I think it would be useful to include a preamble that clarifies the locations of the files and what parts of them are to be changed to change the effects. Not only that, but it lacks an essential fact, when the change of the effect is displayed? You must restart each time the program? Same goes for other similar articles concerning ALS effects on the aircraft.


That's how we document things.

The article in question contains the specifics of particular effects. If you're interested in the general structure of landclass definitions, you need to read that documentation. If you don't know what an effect is, you need to read up on the rendering pipeline. If you don't know how the xml format works, you need to google that documentation. And if you don't know how to edit a file, you need to google that.

If you would pack all the things that a user could possible not understand into every article on the wiki, the length would grow by a factor 10 easily and you couldn't find anything in the stream of repetitive information.

The wiki is a technical documentation, not a beginner's course in how to tinker with files. You're supposed to find the information you need yourself and learn from the existing examples, not to expect to be taken by the hand and carefully be spoon-fed every bit of wisdom.
Thorsten
 
Posts: 11378
Joined: Mon Nov 02, 2009 8:33 am

Re: Could this be the a new goal of FGFS ?

Postby abassign » Mon Mar 14, 2016 9:33 am

Thorsten wrote in Mon Mar 14, 2016 6:10 am:
The wiki is a technical documentation, not a beginner's course in how to tinker with files. You're supposed to find the information you need yourself and learn from the existing examples, not to expect to be taken by the hand and carefully be spoon-fed every bit of wisdom.


Wiki allows you to insert links that refer to insights creating a sort of reticulated database. Therefore it is the easiest thing in the world, if you know how to insert something like [[link ...]], it does not matter to repeat, you just put the right references.
However, currently it lacking, within the impressive work done, the framework that unites all the objects and by an order, a picture that everyone has to rebuild reading the technical documentation and groped, if he can, to rebuild. This does not seem a good approach ... and frankly I think it's one of the reasons for the difficulty of having new programmers capable of working with ALS and unfortunately not only with it.

I think that would be very good if all those who know the program or parts of it are always encouraged to do a job more complete documentation, that the object on which they work not only documents, but also where does through the appropriate [[... link ...]] and at the same time to check if what is written is contextually "self-consistent".
abassign
 
Posts: 830
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Re: Could this be the a new goal of FGFS ?

Postby Thorsten » Mon Mar 14, 2016 9:48 am

It's a wiki - you can add all the links you like. My time is better spent coding something than to polish documentation for you ;-)


frankly I think it's one of the reasons for the difficulty of having new programmers capable of working with ALS and unfortunately not only with it.


I guess at the end of the day, in other applications you actually need to edit shader code if you want to customize effects. In FG you just pass a few parameters in documented slots. If you feel that's too complicated, then I guess I can't help you any further :-) A large number of aircraft developers manages to use the effect framework just fine, so please don't confuse yourself with everyone else.
Thorsten
 
Posts: 11378
Joined: Mon Nov 02, 2009 8:33 am

Re: Could this be the a new goal of FGFS ?

Postby abassign » Mon Mar 14, 2016 1:21 pm

Thorsten wrote in Mon Mar 14, 2016 9:48 am:It's a wiki - you can add all the links you like. My time is better spent coding something than to polish documentation for you ;-)
I guess at the end of the day, in other applications you actually need to edit shader code if you want to customize effects. In FG you just pass a few parameters in documented slots. If you feel that's too complicated, then I guess I can't help you any further :-) A large number of aircraft developers manages to use the effect framework just fine, so please don't confuse yourself with everyone else.


I am pleased that the 10% he does it, I'm sorry that 90% did not want to make the same path of that 10%. For your information if Linux was developed with your mindset ... would not have the code that currently runs about 90% of the CPU in the world.
I find it very sad what you've written and frankly I'm surprised that a person of your ability is not able to do what qualitative leap that would make your good work accessible to a rich audience of developers.
But you know the world is not always what you think, so I have to have patience and hope that others can understand that one of the most important new features of FGFS is precisely a radical revision of the technical documentation.
abassign
 
Posts: 830
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Re: Could this be the a new goal of FGFS ?

Postby Thorsten » Mon Mar 14, 2016 1:26 pm

I find it very sad what you've written and frankly I'm surprised that a person of your ability is not able to do what qualitative leap that would make your good work accessible to a rich audience of developers.


I predict that the output of developers who can't work from the existing examples and documentation will be very small - so I prefer to focus to help those developers who are actually making an effort :-)

I think you may be confusing a question of ability with a cost/benefit analysis here.
Thorsten
 
Posts: 11378
Joined: Mon Nov 02, 2009 8:33 am

Re: Could this be the a new goal of FGFS ?

Postby Johan G » Mon Mar 14, 2016 1:44 pm

@abasign: You are free to start reading in on any subject and starting to improve the documentation. ;) Unfortunately for many subjects that do include reading though configuration files and source code (due to lack of other documentation :roll: ).

On the other hand I would like to point out that I think the wiki articles Thorsten have done most work on are more or less exemplary compared to some other articles.* ;) Ok, I find some a bit lengthy though, but complex things need the complexities explained.

* But I guess that in his profession work would not be done and finished without having a comprehensible paper written and published. ;)
Low-level flying — It's all fun and games till someone looses an engine. (Paraphrased from a YouTube video)
Improving the Dassault Mirage F1 (Wiki, Forum, GitLab. Work in slow progress)
Johan G
Moderator
 
Posts: 5546
Joined: Fri Aug 06, 2010 5:33 pm
Location: Sweden
Callsign: SE-JG
IRC name: Johan_G
Version: 3.0.0
OS: Windows 7, 32 bit

Re: Could this be the a new goal of FGFS ?

Postby Hooray » Tue Mar 15, 2016 8:03 pm

yes, please feel free to get involved in maintaining the wiki - there's no point telling us how to do it, Thorsten is already among the tiny 1% of contributors who regularly, and almost religiously, document their work and update their articles.

If more people were doing that to a similar degree, the wiki would not need to contain any quotes at all, but most people prefer to post relevant information in the scope of discussions like this one...
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 11493
Joined: Tue Mar 25, 2008 8:40 am

Re: Could this be the a new goal of FGFS ?

Postby abassign » Tue Mar 15, 2016 11:00 pm

Hooray wrote in Tue Mar 15, 2016 8:03 pm:yes, please feel free to get involved in maintaining the wiki - there's no point telling us how to do it, Thorsten is already among the tiny 1% of contributors who regularly, and almost religiously, document their work and update their articles. If more people were doing that to a similar degree, the wiki would not need to contain any quotes at all, but most people prefer to post relevant information in the scope of discussions like this one...


Of course it's true, and those who deny!
But if the documentation is missing its placement within the complicated code FGFS, in few words: "...the documentation is missing the context and so who uses such documentation can not do much unless you first make the considerable effort to reconstruct the context..."
If FGFS wants to be an encrypted program to a select few super-programmers, this approach is understandable, but if FGFS want to operate within a context of normal-programmers this approach is ridiculous.
The lack of such a framework embarrasses anyone who wants to bring to the code, and frankly it is not nice to hear, who will demand to know this context, the usual mantra: "You're incompetent!"
While, in truth, unable is the one who does not want to understand that his way of documenting is totally inadequate and requires large and unnecessary efforts for understanding that can be avoided with a few and essential information. Programming is not a gymnastic activities, for which it is useful to take a preparatory exercise, it is an activity that is based on logic and completeness of information.
A trivial example, I want to change some parameters of an ALS effect to figure out what you can do ... there is not any information properly structured explaining the relationships between various files that build the effect, simply because whoever wrote the documentation of 'effect think that those who read it will know all aspects ... it is a behavior similar to those documents a class of an object without showing the hierarchy of that class and then the class relationships with other classes.
The problem of the lack of a proper and complete documentation of the ALS is just one of the cases, but since the ALS is just one of the aspects that I was going analyzing in the context of this post, it seems right and proper to highlight this lack.
abassign
 
Posts: 830
Joined: Mon Feb 27, 2012 5:09 pm
Location: Italy (living 5 Km from airport LIME)
Callsign: I-BASSY
Version: 2018.3
OS: Linux Mint 19. x

Re: Could this be the a new goal of FGFS ?

Postby Thorsten » Wed Mar 16, 2016 6:06 am

there is not any information properly structured explaining the relationships between various files that build the effect, simply because whoever wrote the documentation of 'effect think that those who read it will know all aspects ...


'properly structured' is an elastic wording - however information is presented you can always claim it's not properly structured. There is no such thing as 'the proper way' to structure information. That's the reason there's often different textbooks on the same subject - because two authors have a different idea what the proper way to present something is.

For reference, the structure of a landclass definition is outlined in:

https://sourceforge.net/p/flightgear/fg ... .materials

The structure of an effect is documented in:

https://sourceforge.net/p/flightgear/fg ... ME.effects

There's a whole wiki page gathering links to OpenGL resources and tutorials here:

http://wiki.flightgear.org/GLSL_Shader_ ... _Resources

It rather seems to be a case of you not taking the 15 minutes to look through the available resources (or taking the minute to make a forum post asking someone for pointers) than a question of an actual lack of documentation.

whoever wrote the documentation of 'effect think that those who read it will know all aspects


Yes, documentation usually makes an assumption about prior knowledge of the reader. For instance the documentation silently assumes you have the knowledge to use a text editor to locate, open and modify and xml file and can preserve the structure of xml tags. If you do not have that knowledge, you can't work with the documentation - so you're expected to read up on it.

Documentation of a particular effect assumes you have the required prior knowledge, i.e. you know what an effect is in general, how to assign it to a surface, ... If you don't know that, you need to read the effect documentation first. The effect documentation in turn assumes you have some prior knowledge of how a rendering pipeline works - if you don't know that, you need to read through a description of the rendering pipeline. That in turn is going to assume you know what variables and other data types are - if you don't know that you need to read a general introduction to computing.

You seem to assume that whenever you have a specific question coming from a specific amount of prior knowledge, there's a streamlined wiki article available that starts from precisely your level of knowledge and answers precisely your question. That's not going to happen - it's simply not sustainable in terms of the time it takes.

If you don't understand what an effect is and how it is implemented, you should not be trying to modify ALS specifics of effects and complain if it doesn't work, it's as simple as that, you should read up on effects first. If you don't understand what a pointer is, you shouldn't modify C++ code and complain that it crashes either.

The truth of the matter is that many aircraft developers have not had problems learning how to apply effects to their 3d models, that we have always answered specific questions in the forum when something was obscure, and that this has worked well. The truth of the matter is also that by your own description, you haven't even been trying to follow the instructions given in the documentation or asked what they mean - you've followed your own judgement, it didn't work and then you complained.

That's self-inflicted, sorry. If you manifestly disregard documentation, if you ask for things to be announced that have been announced in newsletter, FG webpage and the forum, you'll have a hard time convincing anyone that more documentation is the solution.

Please just do what others do, look through what's available first, try to understand it, then ask specific questions giving the details needed.
Thorsten
 
Posts: 11378
Joined: Mon Nov 02, 2009 8:33 am

Previous

Return to New features

Who is online

Users browsing this forum: No registered users and 1 guest