Picture the next ten years
and you probably picture
something new.
New language.
New paradigm.
A green field
and a clean start.
That is not what is coming.
What is coming
is the code that already exists,
still running,
with more of it every year.
The bank still runs it.
The hospital still runs it.
The thing your favourite new app
quietly calls
at the last step of checkout
still runs it.
Nothing gets replaced
as often as we announce.
It gets wrapped.
It gets fronted.
It gets a new interface
and the same heart.
So the valuable skill
is not only building the new.
It is walking into something
large, undocumented,
and load bearing,
and changing it
without breaking
the world that depends on it.
That is a craft,
and almost nobody teaches it.
It looks like reading
before writing.
Like finding the seam
instead of the rewrite.
Like adding one test
around behaviour you do not understand
just to hold it still
while you move.
Like asking who is upstream,
who is downstream,
and who will call at midnight
if you are wrong.
The people who can do this
will not run out of work,
because every year
adds another decade
of somebody else's decisions
to the pile.
And there is a second half,
which is about you.
Everything you write this year
becomes exactly that
for somebody else.
The clever thing
will be the confusing thing.
The undocumented choice
will be the load bearing mystery.
Your Tuesday shortcut
becomes their Thursday outage.
You cannot make it immortal.
You can make it legible.
Say why in the commit.
Leave the reason next to the code.
Delete what nothing calls,
so the map matches the ground.
The future is not a clean start.
It is other people's code,
and yours,
still running
long after the reason
has been forgotten.
Write like someone
will need to understand it
without you.
– Serguey Asael Shinder
Top comments (0)