How to write a research paper:

1. http://www.columbia.edu/cu/biology/ug/research/paper.html : okay, short, biology oriented

2. bombastic but makes its point well (plus a collection of other advice)

http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

3. http://tex.loria.fr/typographie/mathwriting.pdf

Mathematical writing by Don Knuth & al. (course notes, 118 pages) The
first 8 pages very good primer.

4. http://grammar.ccc.commnet.edu/grammar/composition/purpose.htm

generic (not necessarily scientific writing) but has some interesting bits (e.g "sense of purpose")

5. http://ccr.sigcomm.org/online/?q=node/618

Metaphoric, light, short, easy read. Fun. (about computer systems
papers, but applies well to any science/engineering research paper)

How to give a math presentation

1. Zermello's rules

2. Same as 2 above (Simon Peyton Jones)

http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

I'm not sure I agree with all they say. For instance, making slides by hand (NO), making them the night before (NO, you need to give a practice talk - or more!)

But the part about how to give your talk (in the slides.pdf) is good. 

also, the ideas from others are interesting.