Infrastructure improvements on a busy railway are far more effective than restricting the service to certain types of trains, new or otherwise.
Extending platforms, (obviously including necessary adjustments to track layout and signalling) is a once and for ever investment, usually with an opportunity to improve station efficiency. With new trains costing from £1m per coach, any intensive service that requires that level of investment every 20-40 years will soon indicate which way is the most cost effective.
If Salford Crescent is so busy and important then the investment should come when (if) traffic reaches the point of need. There will inevitably be a period between the need and the implementation but if justification is there it will happen. That's exactly what happens where traffic levels are really high as in London. The Thameslink programme totals somewhere over £6B which includes trains, London Bridge rebuilding, Blackfriars rebuilding, St Pancras (Thameslink) rebuilding, and extension of platforms at all bar 3 or 4 of the stations north of London. Such is the volume that the cost per passenger is relatively low for a 50-100 year investment (trains excepted).
If traffic in Manchester really is as heavy in the North-west as you say, then longer trains are the only way, not dozens of little trains of odd lengths wasting paths just because the platforms are that way.