Being a signals engineer I have come across this argument many times.
There are two things that, in my mind, would make the highways (and alot of other things) a better place.
Essential services - schools, social services, waste collection, highways - should be carried out at a regional level and without political interference. Alot of highway authorities are too eagerly run by local politicans, who in alot of cases ignore the advice of their officers. It is the political interference that creates alot of the problems...
".. someone died here, therefore we must have a <insert remedy here>.."
A regional agency, divorced from local political wrangling would make our jobs much more efficient.
The other big problem is officers unwilling to consider ideas outside of the box. Again this stems from the points made above. Local officers become attached to the local preferences. And there is alot more power devolved to local highway authorities nowadays than there ever used to be. Gone are the days where traffic signals and speed limits needed to be approved by the DfT. At least with a regional agency there would be more consistency and funding could be applied without political interference. I also think there should be competency tests for Engineers to ensure that a) their practices are still "current" or at the very least innovative and b) that they make reference to other country's (and indeed our own) research before applying methods from their own little toolkit.
There need to be national RULES, not guidance. Any regional departure should be considered by a national forum.
Engineers have a wealth of information and research available to them, but very few use it preferring to go with how they have been trained or with local practices.
And of course, there is the issue of funding. It is much cheaper to install traffic signals than, say, a roundabout. But then, more accidents occur at roundabouts than signals - so you have to look at the total whole-life societal costs and the bigger picture than just simply your perceived delay at a set of signals. As well as that, there is the direct issue of maintenance. No-one wants to pay for more tax, but electricity bills have increased, maintenance bills have increased and material costs have increased. And these have to be met by budgets that are not increasing in line.
That means things like slot cutting, new controller configurations, or minor improvements have to go on a wish list. That means as traffic flows change, or detection fails, we cannot keep up so the operational efficiency of the junction falls.
My own personal feeling is that all signals should be fitted with MOVA (intelligent vehicle detection) and that at night they should rest on all red. MOVA detectors are generally at least 100m from a junction (you can tell a MOVA site as the loops in the road look like squashed diamond shapes, rather than the chevron or square-types found at ordinary sites) so if a junction is resting on red, you will be detected further from the stop line and there should be no need to slow down much. Also, Engineers should put more time and effort in to making the junction as flexible as possible rather than just "getting it to work". It costs significantly more initially, but the benefits to the driver however small, are those that are perceived the most. Fixed time should never be used.
But on the flip side of that, in busy periods, co-ordinating of junctions may mean that you are held at red at a particular junction for the greater benefit of the whole network. For instance, during peak periods I run a number of junctions on an arterial route at an artifically low cycle time. This holds alot of traffic in that part of the network to prevent two other areas from getting grid-locked, overall improving the throughput of the network. Now to the drivers at those junctions it may seem like they are being unnecessarily delayed, but the benefits to the whole network means that less people are delayed.
Rant over...