I'm studying for an oral qualifying exam and going through various past exams I find on the interwebs, including this mock exam from the University of Bath (Wayback Machine). One of the questions seems like it should be rather straightforward, but it's giving me trouble. The question goes as follows:
Four small squares are cut out from a large square and discarded. The edges of what remains are then identified according to the labels in the figure. Find the fundamental group of the resulting space $X$. [You may wish to introduce more edges to help your calculations, but be careful not to make unauthorized identifications.]
What would be a 'standard' approach to such a question? Here is what I have thought of so far and where I'm stuck:
Using Van Kampen's: Intuitively, I'd like to set $U$ to be the interior of the large square and $V$ to be the enlargement of the boundary of the large square. Then $U \cap V \simeq S^1$, but I get tangled up when it comes to making the edge identifications and determining what to do with $U$, plus, after edge identifications, $U$ and $V$ interact, and so $U \cap V$ might not actually be $S^1$ as I've thought it was. Is Van Kampen's the way to go here?
Since the hint mentions adding edges, is there some way to view this as a simplicial complex and compute the fundamental group from there? If so, how would I go about computing the fundamental group? (I am only familiar with computing homology for simplicial complexes, where I explicitly construct the boundary operators and compute the kernel/image of each to find the quotient groups.)
Thank you all for your help/advice!
UPDATES:
New Approach #1: I'm going to try and approach this by using Proposition 1.26 in Hatcher. That is, I'll view $X$ as attaching a 2-cell to the boundary in the picture (as suggested in the comments). First, I've added a basepoint $x_0 \in X$ and edges from this basepoint to the points labeled $C$ in each of the squares, as shown in this picture:
After doing this, the boundary of the space turns into the following picture, and, by contracting the edge $k$ to a point, this space deformation retracts to the wedge of five circles. I've also listed the attaching map.
By the proposition, we should be able to compute $\pi_1(X) = \pi_1(\partial X)/N$ where $N = \langle e\gamma e^{-1}f \gamma f^{-1} g \gamma g^{-1} h \gamma h^{-1} k \gamma^{-1} k^{-1} \rangle$. Writing $\pi_1(\partial X) = \langle \gamma, e, f, g, h \rangle$ and noting that $k$ is now the identity element, we get $\pi_1(X) = \langle \gamma, e, f, g, h \rangle / \langle e\gamma e^{-1}f \gamma f^{-1} g \gamma g^{-1} h \gamma h^{-1} \gamma^{-1} \rangle$.