None of the proofs below are original. My goal here is to practice reading terse homological algebra proofs from Stacks Project and experiment with presenting them in a much more verbose way.

Lemma SP.12.5.11

link

w ker v y x ⊕ y x z f u =( g,f ) g r i h q p v =( k, − h ) k

TODO


Lemma SP.12.5.12

link

We have e:ker⁡f→ker⁡ke : \ker f \to \ker k (with j∘e=g∘ij \circ e = g \circ i) as the coastriction of g∘ig \circ i to ker⁡k.\ker k. This is valid since

k∘g∘i=h∘f∘i=h∘0=0.k \circ g \circ i = h \circ f \circ i = h \circ 0 = 0.

Since we have 0:ker⁡k→y0 : \ker k \to y and j:ker⁡k→xj : \ker k \to x the pullback UP of ww gives us t=⟨j,0⟩.t = \langle j, 0 \rangle.

Since f∘t=0,f \circ t = 0, we have a coastriction u:ker⁡k→ker⁡fu : \ker k \to \ker f such that i∘u=t.i \circ u = t.

ker f w y ker k x z i e = ker k | ( g ◦ i ) f g h u = ker f | t t = j, 0 j k

To show isomorphism, we want u∘e=id⁡ker⁡fu \circ e = \id_{\ker f} and e∘u=id⁡ker⁡k.e \circ u = \id_{\ker k}.

For the first case, we can send probes into xx and yy and use the joint monomorphism property of a pullback.

Start with g∘i∘u∘e,g \circ i \circ u \circ e, the xx-probe. Since i∘u=t,i \circ u = t, this is just g∘t∘e.g \circ t \circ e.

! ker f w y ker k x z i e f g h u t = i ◦ u j k

Now use j=g∘t.j = g \circ t.

! ker f w y ker k x z i e f g h u t j = g ◦ t k

Now j∘e=g∘i.j \circ e = g \circ i.

! ker f w y ker k x z i e f g h u t j k

We have what we need from the first probe, g∘i∘u∘e=g∘i.g \circ i \circ u \circ e = g \circ i.

Now for the second probe, we want f∘i∘u∘e.f \circ i \circ u \circ e. This is simple.

! ker f w y ker k x z i 0 e f g h u t j k

Since f∘i=0,f \circ i = 0, it doesn’t matter whether we include an u∘eu \circ e or not, the result remains 0.0.

So, since f∘i∘u∘e=f∘if \circ i \circ u \circ e = f \circ i and g∘i∘u∘e=g∘i,g \circ i \circ u \circ e = g \circ i, we can conclude that i∘u∘e=ii \circ u \circ e = i via the joint monomorphism property of w.w. Since ii itself s a monomorphism, this further implies u∘e=id⁡ker⁡f.u \circ e = \id_{\ker f}.

To prove e∘u=id⁡ker⁡ke \circ u = \id_{\ker k} we will consider j∘e∘u.j \circ e \circ u. But since j∘e=g∘ij \circ e = g \circ i this is the same as g∘i∘u.g \circ i \circ u.

ker f w y ! ker k x z i e f g h u t j k

But now we see that i∘u=t.i \circ u = t.

ker f w y ! ker k x z i e f g h u t j k

Now we use j=g∘t.j = g \circ t.

ker f w y ! ker k x z i e f g h u t j k

So j∘e∘u=jj \circ e \circ u = j and since jj is mono, we have e∘u=id⁡ker⁡k.e \circ u = \id_{\ker k}.

Thus we have an isomorphism ker⁡f≅ker⁡k.\ker f \cong \ker k.