If I want to create my own bitmap/picture file reader (in any format, .jpg, .tga, .bmp, etc…), what kind of information I need to extract from a picture to later use texture map?
Is there a information page that give specification on what openGL needs to get before creating a Texture map object?
For the specification, look at the man page for glTexImage2D: you need to supply the width and height in pixels, the format and type of the data, and a pointer to the data.
It is certainly possible to write your own image loader for every file type and parse this information out, but other than for the experience, why would you? Use QuickTime or NSImage or other preexisting image libraries.