Oxford's roads have been relatively unchanged over the last fifty five years - the last significant main road build was Marston Ferry Rd, opened in 1971 and there has been almost no attempt at road-widening. The result is that a fairly simple street classification is possible: woonerf/mews, side-streets, main roads, and strategic highways. The first three of these are for providing access to and from homes and destinations, the last also serves to enable through trips.
The division between street types can be seen as a relationship between traffic volumes (and the structure of the road network) and local features, such as street widths and the presence or absence of footways, formal crossings, cycle infrastructure, traffic calming, and grade separation. Mismatches between these result in problems, and solving those requires a decision as to which category a street falls into.
| category | traffic volume | diagnostic features | design speed |
|---|---|---|---|
| mews/woonerf | <300cars/day | no footways | 5mph |
| side-streets | <2000cars/day | no formal crossings, no cycling infrastructure | 15mph |
| main roads | one lane each direction (ieally <12,000mv/day) | bus routes, priority crossings, cycle infrastructure, signal junctions | 20mph |
| strategic roads | any | speed limit >20mph, may be >2 traffic lanes, carries through traffic | any (but 30mph with significant at-grade crossings) |
This basic classification is based solely on movement features of the street - the place features of a street determine distinctions and layout inside these categories. Any of these kinds of streets could have shops and services on them, for example, but their implementation might be quite different - from a single shop tucked away in a woonerf with no fuss to something like a motorway services area off the A40.
Mews/woonerfs are small areas in residential areas with very low traffic volumes. No pedestrian infrastructure is provided - the entire space is effectively the footway and all modes share the same space. For this to work, Motor traffic volumes need to be under 300cars/day (30/hour in peak) and ideally lower, and design speeds have to be very low, perhaps as low as 5mph. There aren't many of these in Oxford: Reliance Way and York Place are examples.
Most of Oxford's streets are side-streets. These are mostly residential but can have significant retail/civic features. There will be footways but no cycling infrastructure, so traffic volumes need to be under 2000pcu/day (200/hour) and ideally under 1500/day. They should have 20mph speed limits and 15mph design speeds (design features to support low speeds may include narrow carriageways, speed humps, build outs, or other traffic calming features). Some support for crossings may be provided, but pedestrians should be able to cross anywhere. There may be some buses, but rarely more than one route.
Oxford's main roads mostly have A or B classifications. Main roads will have one lane each way and traffic volumes should ideally be under 12,000pcu/day, but most importantly kept low enough to avoid congestion or the need for separate turn lanes at junctions. They will have footways — wider ones where pedestrian flows are high — and zebra or signal crossings. They need cycling infrastructure along them, and main road junctions need pedestrian signal crossings and measures to separate cycling movements from motor traffic (in time or space). They should have 20mph speed limits and 20mph design speeds, with 30mph only where there is fully separated cycling infrastructure and no walking or cycling "exchange" across the street (in Oxford, probably only a stretch of Marston Ferry Rd). These will usually be major bus routes. Traffic calming is unlikely to be appropriate, but lane widths should be kept to 3m, or <6.1m for a two-way lane with no centre line, to help keep speeds down.
Then there are highways, or strategic through routes — Oxford's ring-road, the A40, the A4074, the B4044, etc. These may have two lanes in each direction and may carry very high levels of motor traffic. Speed limits can be 50mph, but should be dropped to 40mph wherever noise is a concern, and to 30mph wherever there is significant interaction with people walking and cycling (e.g. the at-grade signal crossings at Barton Park and Kiln Lane, Wolvercote and Cutteslowe roundabouts, etc). They need fully separated footways and cycle tracks -- shared paths where walking and cycling numbers are low -- and crossings on significant foot or cycle routes would be grade separated except for one-lane-each-way roads with 30mph speed limits, where there would be signal crossings.
And there are motorways or near-motorways, such as the M40 and A34. Here walking and cycling must be provided for by alternative routes, and all crossings grade-separated.
Walking
The key decisions are: Does this street need footways? Does this street need priority crossings? and Should crossings be grade-separated?
Cycling
The single most important number in LTN 1/20 isn't actually explicit, but is implicit in the green-yellow transition in the top-right of Figure 4-1. This is the level of motor traffic around which most people stop being willing to cycle mixed with motor traffic, even with 20mph speed limits and design speeds. This is a lot clearer in the draft Rural Design Guide, which says "Where cyclists are on-carriageway, traffic volumes may be up to 2000 pcu/day or 200 pcu/hour but should desirably be less".
If there are problems cycling somewhere in Oxford, the first question should be: "is this a side-street with too much traffic, which needs traffic reduction (and possibly traffic calming as well), or is it a main road which needs cycling infrastructure (or better cycling infrastructure) and a 20mph speed limit?
I understand the 'movement' basis of this piece but I question it.
I would say that you should never seek to split the urban situation into movement and the rest. It is what engineers and planners have sought throughout the 20th and this 21st century. Mostly they have made wrong decisions. They haven't understood scale for one thing. Oxford hasn't quite suffered the highway building that was planned from the late 1940s to the late 1970s but a city such as Birmingham has so suffered. Major sections of the city centre and suburbs were ripped down in the 60s. By c2000 most of the major city centre dual-carriageways were themselves ripped out.
What is needed is a mix of types and apart from motorways, all roads should be considered as 'mixed-priority routes' with some balance between place and movement.
Regarding main roads you say "but most importantly (passenger car units) kept low enough to avoid congestion." I understand that the motornormative DfT uses 'congestion reduction' as a key metric in supporting new infrastructure but the metric is crude. The way the measure is used is a direct benefit to commuting motorists while it gives no benefit to walking, cycling nor public transport. It has no place for the concept of 'place'. Note the process by which the B4044 Eynsham-Oxford road's 'Community Path' was excised from a Road Scheme bid to fit the funding algorithm.
You write:
"(Note that no one should ever put these (High Streets) into new developments: retail and civic centres should be on side-streets adjacent to main roads, not on the main roads themselves. No one wants to sit outside a cafe and enjoy the noise and pollution from non-stop motor traffic at short-range, or have to cross that motor traffic to get from one shop to another.)"
I disagree profoundly.
New developments should absolutely include a High St, as a part of a through route. It's the only way ... that local facilities might survive. The retail market uses footfall as a measure for pricing frontage values. Check it out? High Streets are best and side streets might be a quarter of the rental value (because they have a quarter of the people). Free-standing car-born shopping centres are a different thing. The number of cafes overlooking main roads, the very Banbury Rd and Cowley Roads that you mention, have such places. And this is the case throughout the world.
Graham
If you look at Dutch cities, the shopping streets are very much _not_ the main arterials. In Amsterdam, for example, there are office buildings or housing along Weesperstraat, Prins Hendrikkade, Wittenburgergracht, the S100, etc. — not shops.