Usually, we do our multiplication like this, for example just some shifting around:
[1 0 0 dx] _ [px] _ [px+dx]
[0 1 0 dy] _ [py] _ [py+dy]
[0 0 1 dz] * [pz] = [pz+dz]
[0 0 0 1] __ [1] ___ [1]
We put the translation amount along the right hand edge of the matrix.

While programming around I forgot about that and put them on the lower line instead, on the places marked in bold above.

I wonder what that did to the Points? With the 3x3 in the left top being an identity matrix, and the d's being zero nothing happened to my point

[1 0 0 0] _ [1] _ [1]
[0 1 0 0] _ [1] _ [1]
[0 0 1 0] * [1] = [1]
[1 2 3 1] _ [1] _ [7]

except that 1 we always use in fourth place gets changed (meaning further multiplication with the point results in a mess?)

What am I doing by making the fourth coordinate change?

Usually, I do my matrix multiplication like this:

[1 0 0 dx] | [px] [px + dx] [0 1 0 dy] | [py] [py + dy] [0 0 1 dz] | [pz] = [pz + dz] [A B C 1] | [1] [1] 

Where the translation is along the right-hand edge of the matrix.

Once, I forgot about that and put them on the lower line instead, on the places marked as A B C above.

What did that do to the resulting points? With the upper-left 3x3 being an identity matrix, and the translation being zero, nothing happened to my point

[1 0 0 0] | [1] [1] [0 1 0 0] | [1] [1] [0 0 1 0] | [1] = [1] [1 2 3 1] | [1] [7] 

except that 1 we always use in fourth place gets changed. Does this mean further multiplication with the point results in a mess? What am I doing by changing the bottom row?