update

diff --combined report2/results.tex
index b7a5b90,1d244a7..8b64e9a
--- 1/report2/results.tex
--- 2/report2/results.tex
+++ b/report2/results.tex
@@@ -56,10 -56,10 +56,10 @@@ Positional arguments
  Using the \texttt{-l} option the user can give a initial LURD to the program.
  LURD stands for Left, Up, Right, Down sequence and is a string consisting only
  of $l,u,r,d$ characters and it specifies the initial moves to be taken by the
- agent. To achieve this we simple for every character in the lurd update the
+ agent. To achieve this we simple for every character in the LURD update the
  initial state with the relative product of the transitions belonging to the
  correct direction. After that the problem is solved as usual and thus notifying
- the user wether the LURD leads to a solvable state. Note that when the LURD is
+ the user whether the LURD leads to a solvable state. Note that when the LURD is
  empty the program behaves like there was no LURD given.
  
  \subsection{Evaluation}
@@@ -67,14 -67,16 +67,19 @@@ The assignment was a very interesting a
  feel that we could not get the most out of it due to several reasons.
  
  First of all it was a hassle to get to know sylvan and to be able to use the
 -package. There were some trivial errors in the sylvan package that made
 +package. The most problematic was to understand how can we work with state transitions efficiently. The way it works with sylvan was not evident neither from the introductory lectures, nor from the available documentation. Also, there were some trivial errors in the sylvan package that made
  debugging a lot harder. One of those errors was malformed dot output for
  printing BDDs. In the meantime we have made a pullrequest to fix that.
  
  Secondly, due to the relatively short introduction to BDDs the gap between our
 -knowledge and the knowledge required to build such a solver is quite big. This
 -lead to startup problems.
 +knowlegde and the knowledge required to build such a solver is quite big. This
 +lead to startup problems (which we managed to solve, but lost too much time to it).
  
- Finally, we feel a bit frustrated that we only finished and tested the reachability of a goal state checking, but did not finish the extraction of solving sequences.
++Finally, we feel a bit frustrated that we only finished and tested the reachability of a goal state checking, but did not finish the extraction of solving sequences. We have a solution for that problem, which shows potential, but still contains bugs.
 +
- % Maybe something about the complexity of the assignment
++\subsubsection{Diviion of work}
+ The work was divided evenly in the beginning of the project. When we noticed
+ that we did not have enough time to implement two solutions or even a hybrid
+ approach we decided to go through with the solution that seemed the most
+ doable. In the next phase of the project Alexander did more of the programming
 -and Mart more of the report. Cumulatively we spent a lot of time on it.
++and Mart more on the report. Cumulatively we spent a lot of time on it.