Would it behoove the standardization of some kind of rubric to codify scientific papers on publication where they clearly xml-style state dependencies, conclusi

Would it behoove the standardization of some kind of rubric to codify scientific papers on publication where they clearly xml-style state dependencies, conclusions, conditionalities, need for further research, keywords etc? Or is the state of natural language processing adequate to parse that information out of the publications the way they are?