working with on interval
as approaches , function oscillates with ever-increasing frequency



by definition
which gives

zeros happen at , for
for each interval , area has sign
sum is essentially
and first interval is largest, which in this case is
which has sign of
all contributions are smaller in magnitude and alternate; so

together,


basic riemann sum form is
but actually, use rectangles at middle, so

main problem is that whatever is chosen, it’ll be too big for the increasingly small curves when going
i.e., need to scale somehow

instead of having and , take some and
so now have some

and then for each section, want to have some bars, so
so each has a

and then have the actual function like

trying to graph this out in desmos, first need to compute the position of , say using function

taking in endpoints, . trying to normalize from to ,
so , , ,
and since discrete, need to get nearest , so

the computation for a given rectangle height is the same—using some
combining, over range , use to get s

so now, in desmos, represent this as
and want to put these x values into , so

and highlight above and below axes with or




implies
excluding and , working with , so
and since not all on a set of s, strictly
and thus, , and by extent,

with , turns into
so

splitting at zeros,
pairing to , gives

over a cycle, changes sign, stays the same, is divided by , but same otherwise
so
as one,
and combining,
and factoring,

when choosing s.t. , then also (as always)
and more importantly
so everything is positive and
so

already know
now know
together,