(semi)Auto Text -> (3d/anim opengl etc)graphics
Posted: Tue Jun 26, 2012 4:51 am
I was surfing internet for LISP and Prolog, and how to use that in Artificial intelligence for text processing (I need programming language to allow analyze the text - part-of-speech-tagging, finding noun-verb-phrases, and converting that into "chess notation" like N.V - so that it can be animated in some opengl/directx etc 3d graphics).
So I found this image -
http://2bbie.wordpress.com/2010/02/24/n ... sp-parte1/
there is shown some opengl.lsp
But I do not know how to work with opengl in any programming language actually.
So first of all, how can I use newlisp for opengl and directx for mac-os, iOS, Win, Android or maybe even Java mobiles? the problem is that I want to use some technologies and languages which are more or less stable and will not change just in one-two years so I have to learn some new technology then port it on another platfrom. I mean if Lisp is also an old language, and newlisp is a new one, still, what is the present and future of newlisp and how it can help in problem which I try to solve -
HOW TO CONVERT AUTOMATICALLY TEXT INTO MOVIE/GAME (3D ANIMATION OPENGL/DIRECTX) etc.
So I've read that AI languages - they can allow analyzing the text using some rules like "boy is a noun", "say is a verb", and then combine that into some "Noun-verb phrase" like "boy says".
So I want to do that automatically. I mean I am sure that there are different technologies to automatically convert text of one language into another language - machine translation (I know very well google translate),
so I guess they use some ai languages like lisp and prolog.
But they do not actually use graphics much.
Even mind-mapping software - they only may draw some simple 2d raster icons, which do not animate. So thus such programs like machine translating never use technologies like opengl/directx...
So I want to combine AI in computer games, for graphics/movies etc
with AI for language processng, and make that work in such program.
See for example - widgit -
http://drupal.widgit-online.com/node/add/pagesymbolbody
they automatically convert words into images, and that is nice! especially for children, it text is not very abstaract, and is more about some stories, or kind of scenario for some movie/game where are different persons are described,and what they do.
SO i made own program -
http://sourceforge.net/projects/vedaviz/
see it here -
it uses now around 1000 images from widgit (I got them by reading text of Bhagavad-Gita as it is with purports, and getting all images for all words which widgit provides)...
so problems here:
1 widgit doesn't provide images for each and every word.
2 widgit cannot analyze text using part-of-speech tagging, cannot find relationships between words, between sentences, phrases etc.
3 widgit cannot support opengl 3d animation graphics.
I searched for few similar programs -
I found nice one - called
SPINNAKER STORY MACHINE -
this one can read your input - you just type a text using some of around 40 words,
- among them are nouns, verbs, pronouns, and even proper names which you can type in manually.
but this program is also limited. it was written some 30 years ago.
but it was a success. it used just some 16KBytes memory, it could run on soe 3MHz computer etc,
and it could animate a sentence with nouns, verbs, numerals, articles, custom proper names etc - into some animated story. and it could replay it...
so time had passed of course. Now I see that there are chess programs - and they use some portable game notation PGN file format to record down chess games. So, the thing is that if we consider any story written in machine story to be written in story notation - then it can be saved into some file, reloaded etc. I mean that chess can be expanded into something greater - there can be not just words like King, Queen, bishop, horse etc, and not just actions to move (from LEtter1..8, to another-letter1..8) (x1y1, x2y2).
If chess is expanded - we get many computer games like strategy games and they can use many persons, many actions etc, many ai engines, and they use graphics like opengl directx.
so i see problem that program should have some database - maybe mysql or something else.
actually io was thinking of this,
and cocluded that there might be few modules:
1 computer linguistics module (lisp/prolog for ai and languages)
2 chess notation module (or similar to spinnaker story machine engine to do conversion between text and graphics)
this module is something which is common for visual programming languages like ALICE, scratch etc - they also can do programming, but that is just manual - you can find 3d objects you like, then you add there some properties, change actions, and then run compile - and that it works. it is complicated also.
But why not use thousands of free 3d studio max etc objects on internet? and convert them to opengl directx and show them for each NOUN found in text? And why not apply to each of those NOUNs (3d objetcs) a VERB or Action/script, just like in computer games ? there are so many games, thousands and they use different actions for different actors - do this and that, do here there etc. so this might be connected to text input.
3 there must be some module like in widgit that text is got from input, and then it is converted to some images
4 programming of graphics module - all that must go together - all those images, 3d objects, animation scripts and texts, and finally text is converted into game or movie. by game I mean something like chess game - anyone can open some famous chess game in Portable game notation format, and REDO all the moves and see how all the chess pieces do this and that... So I mean something like chess by saying game here - game which can save all moves, and then redo them. that means usage of NOUN.VERB notation - I am sure that lisp/prolog can handle this - there might be many algorythms for that already...
well maybe lisp is a bit different. if ProLOG is for logic, then maybe I should find some way to do that in prolog, but what about graphics? how can really newlisp help with that?
anyone welcome to answer, say a few words, or even criticize.
I just have some experience in programming using assembler, and (visual/q) basic (sine DOS versions and for win3.1), but also tried to learn other languages. Now I'm puzzled if I should switch to faster C/C++, but this seem to be not the best solution.
I look forward AI programming in lisp/prolog.
though I was adviced to use some LUA for opengl, but I also not sure how to connect that to lisp/prolog and natural language processing...
So, I had experience with programming, but not really with directx opengl. Now i want to use that technologies and thus to convert words into opengl/directx.
For example simplest thing I want to do in some newlisp - how can I get input in some noun.verb form (like in chess PGN format), and then show it on screen using opengl showtext functions (or directx? though it is only for win, but I also use macos and linux/android, so directx is no the best for every platform).
So here is a problem and you can help me with the solution...
yes, actually i was thinking... maybe this program can be done opensource,
thus different people can together work on this, so this is developed faster.
And also I want to know - maybe something like this already exist? So I do not invent wheel again?
I'm asking this becasue I know that there are some programs like autocad/3d studio max/blender - they may also use some command line or scripts like prolog/lisp (now sure exactly, but autocad - sure), so that I can type in some words, and they are converted into some 2d graphics, or even into 3d. But of course that is done in some format, in some programming language. So anyway. the thing that there text is connected with visualization that in graphics. And if user does this input manually - he can do some automation - record key presses, mouse clicks, and then rerun that - just like edit/redo history is used in chess etc.
so I find many similarities here. Also - there are 3d engines like NeoAxis using ogre3d etc. THere say one can load some example, and just click text on the screen, some 'control panel' to generate some objects, and they appear, do this and that. BUt the thing is that input - to click on this or that text or icon or menu command - is done manually. But I want to make it automatic - just like chess allows - just open file, and then show all that visually, like in turn-by-turn game...
for example I think of some special computer file format - to record all the events in game in some common file format - # Actor.Action # Actor.Action # Actor.Action # Actor.Action
like this. So it is like in chess..
so when I open some scenario for movie or game or cartoons (that actually can make millions of dollars, remember movies like the matrix), it should be automatically analized similarly to chess - find Noun+Verbs in each sentence, and then show them on screen in 3d, animate. just like in chess...
It must be very nice!!!! =-)
anyone, any ideas? any suggestiong?
I need something to simplify my work.
first, I need some dictionary of nouns and verbs.
they must have some simple fields - word+part-of-speech.
of course it is for one language, and also if same word may have few part-of-speech-tags - ok let that be. so there lips/prolog (newlisp) might help.
where can i get free source codes for this text processing? to analize text?
I hope there are some free engines, free libraries for that? especially for home/non-commercial/educational use?
If I don't earn money on this - can I use technology like this in my program?
I am interested in this because there are countless texts, and countless games/movies.
so, usually games and movies are done for some not very good scenario - I mean popular movie on each tv channel, but some movies games are nice of course (religious, philosophical, mythological or even historical).
so anyway, if those texts can be automatically analized - to find actors, find the scenes, find all instances of words/images in dictionary for each sentence etc... this would create some knowledge base or expert system. Like mind-map. So then it can be visualized.
Well, anyone please write something here. Can anyone give direct links for everything I need to write this program? AI for text, AI for graphics?
thank you in advance
So I found this image -
http://2bbie.wordpress.com/2010/02/24/n ... sp-parte1/
there is shown some opengl.lsp
But I do not know how to work with opengl in any programming language actually.
So first of all, how can I use newlisp for opengl and directx for mac-os, iOS, Win, Android or maybe even Java mobiles? the problem is that I want to use some technologies and languages which are more or less stable and will not change just in one-two years so I have to learn some new technology then port it on another platfrom. I mean if Lisp is also an old language, and newlisp is a new one, still, what is the present and future of newlisp and how it can help in problem which I try to solve -
HOW TO CONVERT AUTOMATICALLY TEXT INTO MOVIE/GAME (3D ANIMATION OPENGL/DIRECTX) etc.
So I've read that AI languages - they can allow analyzing the text using some rules like "boy is a noun", "say is a verb", and then combine that into some "Noun-verb phrase" like "boy says".
So I want to do that automatically. I mean I am sure that there are different technologies to automatically convert text of one language into another language - machine translation (I know very well google translate),
so I guess they use some ai languages like lisp and prolog.
But they do not actually use graphics much.
Even mind-mapping software - they only may draw some simple 2d raster icons, which do not animate. So thus such programs like machine translating never use technologies like opengl/directx...
So I want to combine AI in computer games, for graphics/movies etc
with AI for language processng, and make that work in such program.
See for example - widgit -
http://drupal.widgit-online.com/node/add/pagesymbolbody
they automatically convert words into images, and that is nice! especially for children, it text is not very abstaract, and is more about some stories, or kind of scenario for some movie/game where are different persons are described,and what they do.
SO i made own program -
http://sourceforge.net/projects/vedaviz/
see it here -
it uses now around 1000 images from widgit (I got them by reading text of Bhagavad-Gita as it is with purports, and getting all images for all words which widgit provides)...
so problems here:
1 widgit doesn't provide images for each and every word.
2 widgit cannot analyze text using part-of-speech tagging, cannot find relationships between words, between sentences, phrases etc.
3 widgit cannot support opengl 3d animation graphics.
I searched for few similar programs -
I found nice one - called
SPINNAKER STORY MACHINE -
this one can read your input - you just type a text using some of around 40 words,
- among them are nouns, verbs, pronouns, and even proper names which you can type in manually.
but this program is also limited. it was written some 30 years ago.
but it was a success. it used just some 16KBytes memory, it could run on soe 3MHz computer etc,
and it could animate a sentence with nouns, verbs, numerals, articles, custom proper names etc - into some animated story. and it could replay it...
so time had passed of course. Now I see that there are chess programs - and they use some portable game notation PGN file format to record down chess games. So, the thing is that if we consider any story written in machine story to be written in story notation - then it can be saved into some file, reloaded etc. I mean that chess can be expanded into something greater - there can be not just words like King, Queen, bishop, horse etc, and not just actions to move (from LEtter1..8, to another-letter1..8) (x1y1, x2y2).
If chess is expanded - we get many computer games like strategy games and they can use many persons, many actions etc, many ai engines, and they use graphics like opengl directx.
so i see problem that program should have some database - maybe mysql or something else.
actually io was thinking of this,
and cocluded that there might be few modules:
1 computer linguistics module (lisp/prolog for ai and languages)
2 chess notation module (or similar to spinnaker story machine engine to do conversion between text and graphics)
this module is something which is common for visual programming languages like ALICE, scratch etc - they also can do programming, but that is just manual - you can find 3d objects you like, then you add there some properties, change actions, and then run compile - and that it works. it is complicated also.
But why not use thousands of free 3d studio max etc objects on internet? and convert them to opengl directx and show them for each NOUN found in text? And why not apply to each of those NOUNs (3d objetcs) a VERB or Action/script, just like in computer games ? there are so many games, thousands and they use different actions for different actors - do this and that, do here there etc. so this might be connected to text input.
3 there must be some module like in widgit that text is got from input, and then it is converted to some images
4 programming of graphics module - all that must go together - all those images, 3d objects, animation scripts and texts, and finally text is converted into game or movie. by game I mean something like chess game - anyone can open some famous chess game in Portable game notation format, and REDO all the moves and see how all the chess pieces do this and that... So I mean something like chess by saying game here - game which can save all moves, and then redo them. that means usage of NOUN.VERB notation - I am sure that lisp/prolog can handle this - there might be many algorythms for that already...
well maybe lisp is a bit different. if ProLOG is for logic, then maybe I should find some way to do that in prolog, but what about graphics? how can really newlisp help with that?
anyone welcome to answer, say a few words, or even criticize.
I just have some experience in programming using assembler, and (visual/q) basic (sine DOS versions and for win3.1), but also tried to learn other languages. Now I'm puzzled if I should switch to faster C/C++, but this seem to be not the best solution.
I look forward AI programming in lisp/prolog.
though I was adviced to use some LUA for opengl, but I also not sure how to connect that to lisp/prolog and natural language processing...
So, I had experience with programming, but not really with directx opengl. Now i want to use that technologies and thus to convert words into opengl/directx.
For example simplest thing I want to do in some newlisp - how can I get input in some noun.verb form (like in chess PGN format), and then show it on screen using opengl showtext functions (or directx? though it is only for win, but I also use macos and linux/android, so directx is no the best for every platform).
So here is a problem and you can help me with the solution...
yes, actually i was thinking... maybe this program can be done opensource,
thus different people can together work on this, so this is developed faster.
And also I want to know - maybe something like this already exist? So I do not invent wheel again?
I'm asking this becasue I know that there are some programs like autocad/3d studio max/blender - they may also use some command line or scripts like prolog/lisp (now sure exactly, but autocad - sure), so that I can type in some words, and they are converted into some 2d graphics, or even into 3d. But of course that is done in some format, in some programming language. So anyway. the thing that there text is connected with visualization that in graphics. And if user does this input manually - he can do some automation - record key presses, mouse clicks, and then rerun that - just like edit/redo history is used in chess etc.
so I find many similarities here. Also - there are 3d engines like NeoAxis using ogre3d etc. THere say one can load some example, and just click text on the screen, some 'control panel' to generate some objects, and they appear, do this and that. BUt the thing is that input - to click on this or that text or icon or menu command - is done manually. But I want to make it automatic - just like chess allows - just open file, and then show all that visually, like in turn-by-turn game...
for example I think of some special computer file format - to record all the events in game in some common file format - # Actor.Action # Actor.Action # Actor.Action # Actor.Action
like this. So it is like in chess..
so when I open some scenario for movie or game or cartoons (that actually can make millions of dollars, remember movies like the matrix), it should be automatically analized similarly to chess - find Noun+Verbs in each sentence, and then show them on screen in 3d, animate. just like in chess...
It must be very nice!!!! =-)
anyone, any ideas? any suggestiong?
I need something to simplify my work.
first, I need some dictionary of nouns and verbs.
they must have some simple fields - word+part-of-speech.
of course it is for one language, and also if same word may have few part-of-speech-tags - ok let that be. so there lips/prolog (newlisp) might help.
where can i get free source codes for this text processing? to analize text?
I hope there are some free engines, free libraries for that? especially for home/non-commercial/educational use?
If I don't earn money on this - can I use technology like this in my program?
I am interested in this because there are countless texts, and countless games/movies.
so, usually games and movies are done for some not very good scenario - I mean popular movie on each tv channel, but some movies games are nice of course (religious, philosophical, mythological or even historical).
so anyway, if those texts can be automatically analized - to find actors, find the scenes, find all instances of words/images in dictionary for each sentence etc... this would create some knowledge base or expert system. Like mind-map. So then it can be visualized.
Well, anyone please write something here. Can anyone give direct links for everything I need to write this program? AI for text, AI for graphics?
thank you in advance