The main thing I would do if I ran the railways would be to slash the crashworthness requirement for new trains, reducing weight and increasing capacity. Even the Mk1s were very safe, the Mk2s and Mk3s were even safer. The modern crashworthness standards are over the top and I imagine they are only justified as the true costs are not taken into account (extra fuel use, lost capacity etc, all over the lifetime of the trains). There has only been one serious accident involving post-privatisation stock, Greyrigg, and Mk2s and Mk3s have been shown to perform very well in derailments, and the Mk1s also have a good record (at least when you don't cook the statistics by adding in pre-Mk1s in). Crumple Zones particularly have no place on trains.
I can't disagree more with this. Mk1 and 2 stock was declared safe many years ago, but only when compared to their predecessors. By modern standards they're not good enough - especially at high speeds. Mk 3s have a good record, but even then, there have been some notable catastrophic structural failures - Look at Southall and Ladbroke Grove in particular. When the side of one carriage had an impact from the adjacent freight wagon, it created a weak point that moments later when subjected to an end load as the HST compacted, caused the whole carriage to buckle in half and folder over itself as the following carriages compressed it. This can be mitigated for with modern designs and materials. As good as Mk3's are, they can still be improved.
Mk 4s seem to have fared well considering they've been involved in some of the highest speed and most violent impacts - Great Heck in particular which is, at 142mph closing speed, the fastest collision in UK history. The DVT there was the principle crash structure - it absorbed so much of the initial energy (sadly at the cost of the driver), that the carriages had less to worry about. Still a lot of energy to be dissipated though. Had the 91 been leading, it's postulated that the impact wouldn't have occurred as the heavier loco wouldn't have been deflected like the DVT was on impact with the Land Rover. Even if it had been deflected and impacted with the freight loco, it's greater mass would have changed the dynamics of the impact - for better or worse can't be guessed. It would have been subjected to less deflection than the DVT but would not have absorbed as much energy. Either way, the fact is, it makes a difference.
This is well worth a read:
http://www.rail-reg.gov.uk/server/show/nav.1202
The Pendolino at Greyrigg stood up astonishingly well to the forces involved. There was no impact with another train, and the speed was around 90mph, which is significantly less than Great Heck, however, such was the integrity of the structures after the crash, it's clear they are very strong and well designed.
On the other end of the scale look at Clapham Junction - total carnage in a relatively modest speed impact because the trains were not up to the job structurally. Look at the photos and the failure mode of the carriage structures - it's frightening.
Can you imagine if new stock was designed with lower crashworthiness requirements, then it was subsequently involved in a collision in which 20 people died, and it was revealed to the public that this was the case and some or all of the lives would have been saved if the standards were met to previously held levels? Outcry.
Your last comment "Crumple zones particularly have no place on trains" leads me to think you don't fully understand the dynamics of an impact. Crashworthiness is the ability of the structure of a vehicle (car, train, truck, bus, whatever) to protect the occupants by limiting the impact forces that reach them, and managing the energy round the passenger structure to ensure that the "survival space" is not compromised. It's simple, basic physics that you need to disperse the energy over as long a distance as is possible to give you the maximum time to absorb the energy and therefore reduce the acceleration (which is what creates the large forces that cause the problems). Crumple zones may cost a few seats in a carriage, but at the cost of a few lives in a crash? That's a no-brainer in my book. It's the exact same principle that allows thousands of people to walk out of car crashes that would have been fatal 20 years ago when they had steel ladder chassis that directed the forces straight into the body shell where they were sitting.
The other thing in crashworthiness is bogie retention - vehicles shedding bogies has long been identified as a major issue in collisions. Those bogies are very heavy and can cause huge secondary impacts after the initial one. Same goes with the wheelsets within them.
Interiors need to be carefully designed - seats and tables coming loose can cause serious injuries and death.
Windows breaking too easily allowing passengers to literally fall out of moving carriages (as happened at Heck and Hatfield) while the train is moving is an issue. You can't stop flying bits of steel breaking glass of course, but if the structure that retains the window is not stiff enough (and it is a main part of the survival space) then the displacement of the structure from impact forces at the end of the carriage, will either smash the window or let it pop out before anything has even hit it. This is a fundamental failure of the worst kind.
Same goes for access doors - they must have a structure that can allow them to be opened by either passengers or emergency services after an impact. As these are often located in the crumple zones, this is a particular challenge.
Articulated bogies - MAJOR advantage - huge gains in crashworthiness from using them where possible. They allow a much stronger link between vehicles, they don't dislodge as easily and they give the train a much more rigid formation. Look at the TGV and Eurostar derailments of the past - some at 300km/h!!! The trains often remain upright, in line and in formation. Read this:
http://www.railfaneurope.net/tgv/wrecks.html
So - reducing crashworthiness is a MAJOR MAJOR mistake. Some impacts simply cannot be accounted for. Some, like Great Heck and Eschede in Germany have impacts so violent that no crashworthiness can reasonably contain the damage that will result, but think of the death tolls in those incidents if the structures were designed to accommodate a few more passengers at the expense of structural integrity... it'd be like a plane crash.
.
