It’s certainly an essential part of the process in a pure turbojet engine. But like the camshaft and valves in an automotive ICE or diesel engine, it’s a necessary evil that consumes power that in a counterfactual magical world you could avoid diverting and instead use to make your engine yield more useful results.
As @Francis_Vaughan almost says, in a steam turbine the production of the working gas is external to the “engine”. And as he did say, the whole goal of the turbine there is to extract the absolute practical maximum of the energy in that gas. Rather than the pure turbojet turbine’s goal of extracting the bare minimum needed to make the rest of the engine’s compress-burn-expand cycle work in a self-sustaining fashion.
The more modern turbofan jet engines are in a sense a hybrid. The so-called “core” of compressor, combustor and turbine work as I described above: trying to extract the minimum energy needed to make the cycle self-sustaining. Then a separate fan is bolted on the front turned by a separate turbine bolted on the back.
That fan + secondary turbine rig is indeed designed to extract the absolute maximum power possible consistent with two limitations:
Like a steam turbine, the environment you’re exhausting into limits you to below the ideal theoretical results you’d get exhausting into a vacuum at 0K. But unlike a steam turbine aboard a ship or fixed plant that exhausts into a condenser maintaining constant parameters in a closed system, a jet engine exhausts into the ambient atmosphere where we encounter very different combos of temperature and pressure. From sea level in the Persian Gulf at ~1050 millibars at 50C with 95+% humidity to high altitude at 200mb at -60C with 2% humidity. Given that (so far at least) all this is done with fixed geometry airfoils, a system that works acceptably everywhere is probably not optimal everywhere.
Unlike a steam turbine, you can’t have the power turbine be so much of an obstacle to flow that it clogs up the output of the upstream core more than that core can stand. Much of the black magic of modern jet engine design is right there; making these two separate systems with often opposing demands co-exist and couple efficiently both at steady state operations and during transients of starting through to idle and idle through to full power.
Which bit about transients leads to your second point.
The rotating mass of the core and of the fan system is big and heavy, despite being built as lightly as possible given the durability goals. Engines on 737- or A320-sized airplanes weight ~2500Kg / 5,000# each. The Trent 900 on the A380 we discussed earlier is about 6,500Kg / 14,000# each. I don’t have a formal rule of thumb for the rotating mass, but big picture I’d bet it’s about 1/3rd, so a little under a tonne/ton per small engine and 2+ tonnes/tons for a big one.
For very round numbers jet engines at idle are spinning the core around 50% of max RPM and the fan at around 20% of max RPM. Takeoff is close to 100% on both and cruise is in the mid-high 90s on both. In absolute terms core RPM redlines are in 20,000-30,000 RPM bracket and fan RPM redlines are in the 12,000-18,000 RPM bracket.
So yeah, lots of heavy stuff spinning at insane speeds under fine tolerances in beastly heat & highly reactive chemical conditions.
Modern engines can get from idle to full power or vice versa in 5-7 seconds. The acceleration is very non-linear. Getting from e.g. 50% RPM to 80% RPM takes most of the time and getting from 80 to 100% RPM is the last little snip of time. The thrust output vs. RPM is also very non-linear with not much change happening in the lower RPM ranges and substantially all the incremental power occurring in the last 10-15%.
Part of the reason airliners have such large flaps is to allow the wing to fly more slowly so takeoff and landing speeds are low enough that runways don’t always need to be 3 miles long to accelerate or stop.
But another large part for landing is to generate a lot of drag so we can leave the power set pretty high during approach and landing so that we have very fast thrust response if we need to go around. Pushing up the throttles and having to wait 5 or 7 seconds for meaningful power would often be a few seconds too many. This was a very severe problem with the early (<1960) military jets of all nations.
Making the engine acceleration rate fast enough, but keeping the core/fan coupling efficiency high for steady state cruise is another constraint on design. During steady state ops the core needs some excess capacity to add first fuel to the existing airflow, that grows the fire, that grows the turbine RPM which then grows the compressor RPM to add more air to further grow the fire, all while being obstructed by the fan turbine extracting some of that extra output to accelerate the fan.
But every bit of that excess capacity represents lost potential efficiency during steady state ops. Much black magic is hidden in there. It’s very much akin to impedance mismatches and their baleful effects on large transients in non-digital electronics design.