Added peer review
This commit is contained in:
		
							
								
								
									
										2
									
								
								.gitignore
									
									
									
									
										vendored
									
									
								
							
							
						
						
									
										2
									
								
								.gitignore
									
									
									
									
										vendored
									
									
								
							@@ -1 +1 @@
 | 
				
			|||||||
main.pdf
 | 
					*.pdf
 | 
				
			||||||
 
 | 
				
			|||||||
							
								
								
									
										30
									
								
								review.typ
									
									
									
									
									
										Normal file
									
								
							
							
						
						
									
										30
									
								
								review.typ
									
									
									
									
									
										Normal file
									
								
							@@ -0,0 +1,30 @@
 | 
				
			|||||||
 | 
					#align(center, [
 | 
				
			||||||
 | 
					  = Review of "Abstract Interpretations, Numerical Domains" by Silvia Gaspari
 | 
				
			||||||
 | 
					  == Matthias Veigel
 | 
				
			||||||
 | 
					])
 | 
				
			||||||
 | 
					=== Summary
 | 
				
			||||||
 | 
					The paper deals with numerical domains and how they work.
 | 
				
			||||||
 | 
					The domains are all based on a concept called latice, which is an ordered set of sets with #sym.tack.t as empty set and #sym.tack.b as set which includes everything from the other sets.
 | 
				
			||||||
 | 
					The different domains specified in this paper are: Interval: number #sym.in [a,b], Polyhedron: lattice of intervals, Zone: x #sym.in [a,b] and x - y < constant, Octagon: Extension of zone domain with #sym.plus.minus x #sym.plus.minus y < constant, Pentagon: x,y are program variables and x #sym.in interval and x < y -> y as upper bound.
 | 
				
			||||||
 | 
					These domains can then be used to constrain variables.
 | 
				
			||||||
 | 
					=== Previous experience
 | 
				
			||||||
 | 
					I have no previous experience with lattices or numerical domains.
 | 
				
			||||||
 | 
					=== General Feedback
 | 
				
			||||||
 | 
					#[
 | 
				
			||||||
 | 
					  #set list(marker: [+])
 | 
				
			||||||
 | 
					  - Good Structure, Domains ordered from least to most complex
 | 
				
			||||||
 | 
					  - Detailed explanation of domains
 | 
				
			||||||
 | 
					  - Methodology is clearly described
 | 
				
			||||||
 | 
					]
 | 
				
			||||||
 | 
					
 | 
				
			||||||
 | 
					#[
 | 
				
			||||||
 | 
					  #set list(marker: [--])
 | 
				
			||||||
 | 
					  - Many typos
 | 
				
			||||||
 | 
					  - Incomplete sentences
 | 
				
			||||||
 | 
					  - Inconsistent capitalization
 | 
				
			||||||
 | 
					  - The text of figure 1 and figure 2 is hard to read, because it is too small.
 | 
				
			||||||
 | 
					  - Figure 1: 0 and +/- are too close together.
 | 
				
			||||||
 | 
					  - Section 4.2: Should use #sym.NN,#sym.ZZ,#sym.QQ,#sym.RR instead of N,Z,Q,R (possible with `\mathbb{N}` in latex, `#sym.NN` in typst)
 | 
				
			||||||
 | 
					]
 | 
				
			||||||
 | 
					
 | 
				
			||||||
 | 
					The paper has a good structure and the different domains are clearly described. A big problem in the paper is the missing explanation for lattices, since most of the paper is build on this concept. I also noticed a few logic mistakes in section 4.1: arithmetic operators: (+) - (-) should result in (+) not #sym.tack.b, (-) - (-) should result in #sym.tack.b not (-), (+) \* (+) should result in (+) not (-). The papers are also not really evaluated yet, only the content in them is summarized currently. Another useful thing to include would be the table from 0:25 in https://www.youtube.com/watch?v=jWUXclaKppk for initial understanding and comparison of the numerical domains.
 | 
				
			||||||
		Reference in New Issue
	
	Block a user