Thursday, August 25, 2011

The armchair bus planner

He idly leafed through the Melway street directory as a sedentary diversion from gardening outside. Such an activity could have been done anytime since about 1980 when the directory started showing bus routes as well as streets.

The page turning stopped at Map 17. Straight streets, but bent bus routes. Almost as soon as the bus would have got up speed it was time to turn the corner, hand straps swinging from side to side.

This was Hadfield, built by the Housing Commission in the 1950s. Nine miles north of the Melbourne CBD and a mile from railways either side. Boundary Road betrays a former outgrown limit, while North, South, East and West Streets reassuringly prevent those in between from getting lost.

Less known than Pascoe Vale to the south or Glenroy to the west but still deserving capitals, Hadfield is a real suburb. As opposed to a mere locality like the nearby obscure lower-case Westbreen. The ‘50s were probably the golden age of the small lettered locality name. Like Bellfield, Pennydale, Coatesville or Coonans Hill. Built after the railway ceased being a necessity for a suburb but before children and housewives routinely motored beyond it.

Three of the area’s four bus routes (513, 527, 534) go to Coburg at the south-east of the map. Three (513, 534, 536) also run to Glenroy Station, back a page to Map 16. Glenroy roughly marks the limit of pre WWII development and forms the edge of the inner suburban fare zone. Two run to quiet Gowrie while none run to busy Broadmeadows, the area's main suburban centre just off the map.

The map showed that bent routes were typical – buses would typically head west, north and then west or east. A couple seemed unsure of themselves. 534 might as well join the communists. Once out of Coburg it sways wildly left, right, left, until deciding far left, to Glenroy, was its destiny. Route 536, slightly to the north, is so indecisive that, for good or evell, trips alternate between parallel streets.

536 veers off Map 17, with a small incursion into Map 6, to 17’s north-west. Map 6 is dominated by Broadmeadows. Shopping centre, library, town hall, schools, Centrelink, police and courts. One could almost live one’s entire life there, and some may have. Plus a major railway station and two orbital SmartBus routes, including the main route to the airport.

‘Broady’ has long had big things planned for it. It’s in successive government plans as a district centre, principal activity centre or suburban central business district. The idea is that people from surrounding suburbs would take jobs there rather than in Melbourne CBD, thus decongesting trains and shortening commutes.

To make the diagonal trip from Map 6 to Map 17, one may take Widford Road as this passes over the Metropolitan Ring Road. North of the Ring Road is a community centre and a shop or two. Bus Route 538 runs behind there via back streets, no doubt to the chagrin of Campbellfield commuters. Most homes in the area are near Widford Street. Hold that thought, for unlike chewing gum on a bus seat, it may come in use later.

Flicking across to Map 7, one sees the 538 paralleling the 902 along Camp Road. There’s only a barracks and a business park, so a single route every fifteen minutes, such as the 902 SmartBus, should suffice.

Ahead and to the left is the Campbellfield Shopping Centre. It’s got the beginnings of a transit oriented hub. Just fifteen kilometres due north of the CBD on its most important road, industrial or millitary land that could suit higher and better uses, an orbital SmartBus and a train that passes (but does not stop). Here the 538 runs north along Sydney Road, and then east, eventually terminating in Campbellfield’s Somerset Estate, most definitely not designed for transport other than the car.

Despite the neighbourhood being off the highway and not on the way to anywhere else, Route 538 is not alone in these parts; Route 531 between North Coburg and Upfield also goes there. Timetables for both are time-capsules; a reminder of how buses in Melbourne used to be. Route 538 runs every 40 minutes until 7pm weekdays and Saturday mornings, while the 531 is weekdays only, operating hourly until 8pm. Two limited service routes serving the same low-density residential streets. Reread that last sentence and hold that thought.

The armchair planner’s mind sped and his felt pen quivered, with blobs forming overleaf. A messy side-to-side red has been drawn, running east, then north, then east on Map 7, over the entire Route 538. Why keep it when it almost entirely overlaps other routes?

During a candid moment he professes a slight satisfaction in wielding the pen, not unlike a dentist after an extraction or a surgeon after cutting a cancer. Although, perhaps unlike the last two, pecuniary interest is not involved. This joy is not because the planner revels in denying people a bus. Rather it is due to the opportunity made; after the cull can come the creating (with some compensating).

It remains likely that some from Somerset will still wish to visit Broadmeadows, preferably via a direct means. Deleting 538, at first glance, appears to remove this option. However the 531, which we’re keeping, terminates at Upfield Station. As does 540, which runs to Broadmeadows.

Joining the two would restore this link, albeit via the slower north than the faster south. And pay homage to existing practice, footnoted in the timetable, of the 538 and 540 being linked. Loadings should be reasonably bidirectional – some will be travelling south to the tram and Coburg, while others will be heading north to Upfield station and Broadmeadows.

Then there is the matter of service levels. Would Somerset residents prefer two limited service routes or one full service route? The latter since no stops are being denied a service. So the resources from the deleted 538 could boost the 531, giving it new weekend and early evening service, much like the 540 that it could join.

The armchair planner is tantalised by these apparently easy gains. A simpler network, fewer routes, the same coverage and better operating hours. Why did someone not think of this before?

Such euphoria potentially precedes pricking by the practicalities.

Not immediately obvious is that Routes 531 and 538/540 are run by different companies. This may test the art of the possible in management where the dreamers are fewer or don’t make the decisions.

Neither was much thought given to route lengths and service frequencies; a forty minute service may neatly meet every second train but is inefficient if the route is 41 minutes long. Then there’s other questions like whether Route 540’s current twenty minute frequency would apply over the entire extended route, none of it or only the current portion.

Still these are minor things for the dreamer, who by now has turned back a map. For he did not forget that Widford’s fate was left hanging on Map 6. 538’s deletion makes it deserving of a service to compensate, but only if made an economical part of something bigger. Unlike some previous excursions, which left a trail of unsolved problems across the pages, fixing this pleasingly returns us to where we started on Map 17, with some Map 16 changes on the side. Such circles are usually bad for the routes themselves but are good for their planning.

So the dreaming resumes. Possibilities include improving access to Broadmeadows by breaking the Ring Road’s historical barrier on route planning. And could services be improved if buses could be sped up by running routes not requiring as many turns? Maybe more straight routes and fewer L and S shaped routes.

The armchair planner partly closes his eyes. This blurs the smaller streets. He tries to imagine the main roads of Map 17.

Clear north-south routes include West Street/Cumberland Rd, and East Street/Sussex Street. One or two others are ‘maybes’.

East-west routes include Hilton Street/Box Forest Road and Boundary Road. There are others but they may be in other route’s catchments. Successful bus routes must serve several markets, for instance local shoppers as well as train connections. Given existing off-peak train frequencies, it was sometimes considered acceptable to break the grid to feed the main centres. For instance Glenroy instead of Oak Park or Coburg instead of Pascoe Vale. A bit like now but straighter.

Like a spider starting a web, his first move was a straight drop down (more or less). From the twig up at Widford Street, the bus from Broadmeadows turns right at Daley, then Morley, West, Cumberland and then east to Coburg Station, via portions of Gaffney, O’Hea and/or Bell. Hence it serves catchments of Route 536 (Morley St), 534 (West St) and 513 (most of the rest) in a roughly north – south alignment.

This modified 513 has the makings of a very strong route; Major stops at Broadmeadows and Coburg, two medium sized shopping areas in between and high directness. Given that it ends in distant Greensborough or Eltham it’s rather long, but that’s a problem for another day and another map. Glenroy loses this route but there remain others and Broadmeadows ought to attract more passengers.

The other key north-south route is 527. It’s already quite direct (with a wobble to meet Tram 55) and serves Coburg and beyond to its south. In the north though it ends with a whimper, skirting a school, a cemetery and the sometimes as quiet Gowrie Station.

‘527 deserves better than Gowrie’, the armchair planer thought. If not Broadmeadows, at least Glenroy. The aptly named Middle Street might get it there via the West Street Shops and Glenroy Road to the station. However traffic calming measures and residents may not agree. Either South or Hilton Street may be alternatives, with the former sacrificing coverage and the latter directness. Again the route has become longer and remedial treatment may be required on another map.

Sussex Street (Route 534) is a third possible north-south corridor. It’s nearer midway between Cumberland St and the railway line, so ought to be better than the parallel Derby Street, just 300 metres away. However, unless dead people can be persuaded to ride buses, Sussex Street loses its catchment north of Boundary Road. And buses on Sussex miss the Melville Road tram. It may yet deserve a service but only from the fragments left after the rest of the network has been built.

From north to south the first east-west route is the alternating 536. It’s more frequent than the 538 but it still lacks evening, Sunday and public holiday service. While some of its western catchment would be served by the extended 513, its east still needs service. The 536 could be kept but straightened, with routing via Hilton Street being the lesser evil.

Some numbers enter the dreamer’s mind. Train frequency: 20 minutes. Bus 536 frequency: 30 minutes. Existing bus run time: 20 minutes. Buses needed if route run independently (which it isn’t): 2. But if straightening could reduce the bus run time, could the two run a 20 minute headway harmonised with trains? Or if there’s more demand elsewhere, pull one off and leave a 40 minute service, meeting every second train? The hard question here is whether those in North East Glenroy will accept a faster and better connected route in exchange for a longer walk to it.

The next big thick east-west line is Rhodes Parade/Pascoe Street/Boundary Road. Its west is seved by part of the 513. Its middle has no service, though there is one in nearby parallel South Street. Its east, near Merlynston Station, has the wavering 534 from Coburg (a relatively recent gain).

Given the first move was to straighten the 513, it’s only fair that the 534 get similar treatment. The grand swap commenced above can now be completed. 513, as discussed before, now up West St to Broadmeadows. Taking its place could be 534 routed west to Glenroy via Rhodes and Plumpton.

Dreamers are prone to thinking too much of what can (should?) be without sufficiently appreciating what’s there. An example was a fleeting thought to finish the modified 534 at Oak Park instead of Glenroy. However this was soon dropped. For part of the route (as 513) is within walking distance of Oak Park Station and the route’s shopper function is best met by keeping the Glenroy connection (Oak Park having few shops). The 534’s wavering section to Coburg is for now left alone but has been marked for later scrutiny.

It was getting late and way past bedtime. Washing up from breakfast remained. Lunch was a fridge visit. And tea had not been eaten.

But a wish to avoid a sleepless night would require that Map 17’s loose ends were at least listed if not solved. There is no loss so great as a forgotten idea. These night thoughts though are the craziest – radical, controversial and often half-baked. Who was it that said nothing good happens after 2am? Luckily none can be acted upon until daylight brings reason.

For example, could 536 have been extended from Gowrie to Fawkner shops on Jukes Road? Yes Fawkner gains but it’s a slow zig-zag across Sydney Road that would likely cost frequency.

And does Broadmeadows deserve a second major route from the south? It may even be an even stronger terminus than the Glenroy suggested for Route 527. If routed via Hilton St its East Street coverage would be restored and 536 could be moved north to nearer its current route.

Then there’s Pascoe Vale Station. Like Oak Park it has no bus to the east. Would a bus from it, via Gaffney Street, to Coburg work? Economy may dictate that this be a straighter redirected 534 from Coburg, with service withdrawn from parts near Merlynston. The Boundary Rd bus would need a new number (perhaps 537) and possibly a southward turn via Sussex and Shorts for coverage sake.

The armchair planner sees both pros and cons in these late bursts of radicalism. By this hour he has not the energy to seek more compelling evidence. Although his gut does see more winners than losers.

Legibility is certainly better. Two major routes north - south; Broadmeadows to Coburg. Three lesser routes east-west; linking stations on each side. Most stops remain in use and route kilometres is much the same as now.

Though the actual streets to use are not set in stone, the big picture network appears roughly sensible, with connectivity, bus running times and their efficient usage major caveats. In nutting these out the result of the dream is now subject to judgement, study, consultation and culling in the pursuit of a network that works.

Friday, August 19, 2011

The shape of Melbourne’s streets

A ramble through the street directory

It is not possible to talk about bus network design without referring to street planning, for streets represent the fist of cards from which bus route planners must choose.


The Central Business District forms our first grid, with its main thoroughfare and tram spine aligned NNW to SSE, though for simplicity I will describe this as north-south. Its blocks are rectangular, with the ‘little’ streets feeding into the axis streets.

To the north almost all city streets continue into others (albeit angled) as they become West Melbourne, Parkville or Carlton. To the south most streets are blocked by the river, with Swanston and Spencer Streets being exceptions. However walkers enjoy greater permeability with some pedestrian-only bridges to Southbank.

State Parliament and Southern Cross Station bookend the CBD to the east and west respectively. Bourke Street, the central partially malled shopping street, must yield to these buildings at both ends. Whereas Collins Street, its classier southern neighbour, was privileged enough to recently gain an extension at its Docklands end. Its ‘Paris end’ also exits the grid, its leftward veer dividing the politics from the money. Similar easterly access, though with a rightward turn, exists off Lonsdale and Latrobe Streets.

Much of Flinders Street is denied a river vista by Federation Square and the two-block station named after it. The building’s narrowness suits the need to maximise platform space; open skies being an advantage in the steam age. Further west, the railway viaduct squashes the river vista from this quarter of low repute, austere pavements and architectual mistakes. However Flinders, like Collins, is unconstrained by Hoddle, with its trams running east, west and then north.

Docklands is remote, windswept and a mystery to many who don’t work there. Shopping and stadium stunt access to it, via Bourke and Lonsdale, at Spencer. While the former allows pedestrian access, this walk offers insufficient instant gratification during the two block walk; the steep stairs, the railway overpass, Docklands stadium, some ramps and a pause for a busy road deter all but the leisured curious (and tall train spotters). Colour vison is wasted here; everything is a shade of grey, except for one or two still visible brown remnants from the 1970s. However those leaving Docklands are better rewarded as the bridge provides a vista down Bourke Street to St Patrick’s Cathedral.

Docklands’ north-south routes are for driving more than walking, while its finer-grained areas are beyond lunchtime range of unrushed walkers from outside. It could yet become a medieval village with few from outside. Waterfront City attempts a grid but highways or water stymie egress from most directions; it’s not like Elizabeth or Swanston where one can march north or south until one is footsore, as Henry Bolte would wish. Its east-west streets bear the same names as in the CBD proper, but not always the same vistas or tram routes.

The redeveloped Southbank features a somewhat disordered grid. Its thorougfares are either pedestrian or car; different to say Collins or Elizabeth Street where three or four modes mix. Its key pedestrian way is along the river with frequent bridges from the north bank. Vehicle access predominates to the south where roads distribute traffic from the West Gate Freeway. Unlike Docklands, Southbank shows its face to the CBD (via the river) and, its South Wharf portion excepted, is more accessible.


A kilometre or two away from the Melbourne Town Hall, particularly to the north and east, the angled CBD mesh gives way to a coarser but more extensive net whose streets are almost exactly north-south and east-west. The south-south-easterly St Kilda Road makes it gently diagonal in that direction, with junctions with numerous east-west roads (which are exceptionally supplied with trams thanks to the Prahran and Malvern Tramways Trust).

Grid spacing is closest in the old areas of Brunswick, Fitzroy, Richmond, previously home to small factories and the working class. Occasionally these fine-grained areas are interrupted by 1960s housing commission towers on ‘superblocks’.

Spacing widens with distance but the grid becomes no less perfect. It increases from about a half-mile in the 1920s suburbs (10 kilometres distant) to about a mile by 20 kilometres distant. These are all key roads, mostly with building frontages, and also occasionally with service roads further out. Beyond established housing a coarse grid exists as rural roads, no doubt becoming tomorrow’s arteries as suburbanisation encroaches.

Spacings between traffic lights and the roads themselves also widen with increasing distance from the CBD. For buses wide grids mean that service may also be needed on intermediate streets to provide reasonable coverage; something that 1km grids in suburbs like Mount Waverley only just avoids. But whatever their width, grids allow faster speeds and more legible bus routes; one traffic signal and no turns beats two junctions and two turns.

In some directions, particularly to the west, north and north-east, the grid may be interrupted by freeways, rivers, parks or scrap land. Fawkner to Reservoir is short geographically but distant by road, for example. The Yarra also divides and sparsens the suburbs, with only a few favoured roads granted a bridge.

Sometimes, like the bisected Bourke Street in Docklands, both halves may be similarly named, indicating either a former or intended joining. For example the Grieve Parades in Altona, which could make a fine Altona – Toyota – Sunshine industrial bus route if joined. Or the Balmorals, Crowns, Graces and Highs over in Altona Meadows. But hop over to Hoppers and the mile grid reasserts itself with the parallel Tarneit, Derrimut, Morris, Leakes, Sayers, Hogans and Heaths Roads.

Up in Sunshine West, Glengala Avenue goes some way before it hits the freeway. Had it hypothetically continued it may have given Derrimut a bus earlier (possibly an extended Route 454) instead of requiring a wait for a new route (400).

This example demonstrates that discussions about freeways and their form are not only about infrastructure priorities (eg freeways versus railways) but also about the importance of long-distance orbital versus more regionalised road access (that a half to one mile at-grade grid provides). Due to the lack of the latter in parts of the west, the Western Ring Road has been bigger for the west than Eastlink for the east, with its more contiguous development and continuous roads.

The southern portion of Sunshine West does not have an equivalent spine to Glengala Road in the north. The Avenue (which has a bus) forms the makings of a spine but its short streets to the north limits its legible pedestrian catchment. Access is possible into Wright Street but catchment streets are widely spaced compared to the less direct Talintyre Road with more branches. Neither Wright nor Talintyre have buses so some homes exceed 400 metres from one. Whereas streets with the best of both worlds (eg direct road with closely spaced side streets but not necessarily particularly high density) like Sydney Road in Coburg may have facilitated an accessible, fast and well-patronised service.

While four-way intersections are the rule, there are more at some locations. And not just at Five Ways, outside Cranbourne. In the inner-east, all roads lead to Camberwell. Camberwell’s centrality is so great, or the take-up of motoring in this affluent area so early, that it sapped surrounding suburbs of significant shopping strips. In contrast more homes are walkable from a supermarket along the less affluent Dandenong and Frankston lines.

Multi-way junctions attract the map viewer as much as they are cursed by drivers. Eyes scanning a map are naturally drawn to St Kilda Junction, Reservoir and the notorious but soon to be removed Haymarket roundabout in Parkville. Two or three roads may meet obliquely, slicing the suburban grid. Examples includer Camberwell (again), Kew, Reservoir and Footscray. However mass motorisation caused prewar accessibility to become congestion and multi-way junctions have not been favoured since.

Inside grids

There are differences as to what’s inside the grids. Inner suburbs have grids within grids, with older working-class suburbs like Brunswick, Collingwood and Richmond markedly denser than spacious Mont Albert or Malvern East.

Other planners filled their grids with curves, not unlike an arched window. That off Bay Street in Brighton boasts an inner, middle and outer circle. Albion’s Selwyn Street more successfully encloses its park but does not form the suburb’s centre. Western Glenroy’s arch is irregular, while Sunshine West’s is a neat rounded square.

Circles are rare enough to confer novelty, like knots in a timber beam, but aren’t confined to the poshest suburbs. Albert Park’s grand St Vincent Gardens is the full circle (or oval), unlike Brighton semi. Altona North must feel impoverished with just its tiny circle (called ‘The Circle’) with attempted polygons surrounding. Nevertheless it forms the hub of local commerce. St Albans’ circle daintily sidesteps its main road junction. It forms a six slice cake, with the railway splitting it cleanly in half. Commerce occupies one slice with reserves in the centre.

Central reserves are the common pattern; circle centres are too oddly shaped to subdivide and they provide a pleasant outlook for homes lucky enough to front them. Homes also provides passive surveillance for park users.

Centres however vary greatly in size. Unlike the scraps of land in St Albans’, the park in the aptly named Park Orchard’s circle is large and unbroken. It forms a large steering wheel, though three of the four spokes are footpaths rather than roads.

Not that distant is Circle Ridge in Chirnside Park. I’m guessing this surrounds a peak given the street name. If this is the case then its aim may be to angle houses to look outwards at the view instead of the usual inward park. Pakenham’s version has a literary bent and encloses a lake.

More common than circles in the postwar suburbs are curvilinear streets and then culs-de-sac. These were considered to offer greater interest to the resident and visitor while respecting land contours and calming traffic for childrens’ benefit. Unfortunately though the layout induces vehicle usage due to low pedestrian legibility and permeability. The ‘exclusive’ golf estates like Sanctuary Lakes, Chirnside Park and the unserveable Sandhurst are the biggest offenders, but the pattern is widespread.

There are sometimes access ways at the end of culs-de-sac but these lack street frontages and may be perceived as unsafe. Since local police and residents groups often support closure of access ways and alleys (to reduce assaults, theft and graffiti), such access is less permanent as that via grid streets with street frontages.

The busier arterials in these areas may lack building frontages (more the case in newer cities like Canberra than Melbourne), or, where there are still buildings, they are set back from the street with service lanes and large front parking areas. Intersections are widely spaced, and in the worst cases are controlled by roundabouts. Distances are too far for people to talk or even recognise one another in the next building or across the street, from one shop door to another.

And even if they could (eg through a GPS-based mobile phone app linked to Facebook), the limited pedestrian access only grudgingly provided at widely spaced intersections means that such roads cannot sustain the form of street life written about by Jane Jacobs. Neither this nor the curvilinear distributor street nor the cul-de-sac suit efficient ‘last mile’ bus routes; one reason for the success of trams, along with their service levels, is the walkable grid that feeds each stop.

New urbanist planning has influenced new residential street layouts more than locating new shopping centres near stations or making industrial employment areas walkable. Modern estates now have straighter or at least connected streets compared to those of the 1980s. Altona Meadows was an early example, with a dense grid not unlike older suburbs. This layout allows a single bus route (411/412) to serve a suprisingly large catchment. And while weekend services are limited, its 20 minute weekday frequency is high for a 1970s-1980s suburb.

This evolution is clearest in established outer suburbs that have grown continously for a century or more. Central Cranbourne, Pakenham or Werribee both have old-style grids. Around 1 to 2 km out several kilometres of cul-de-sacs start. Even further out is often a modified grid, less regular than the old but with a legible central street that could take a tolerably direct bus.

However there is still difficulty in connecting adjacent estates. Bridges are expensive and we're not so gung-ho about draining swamps these days. Hence Shearwater Drive in Pakenham peters out into a park and creek and there appears no provision to connect it to Meeking Drive (and thus link it to the proposed Cardinia Road Railway Station). Buses instead would need to weave on and off the main highway, slowing travel and likely partly duplicating other routes.

Street design for walkability and transit has improved over the last thirty years but not over the last hundred. And decades of impermeable layouts and controlled access roads has left us with a legacy of culs-de-sac to open, super-blocks to bust and roundabouts to remove to improve accessibility.

Friday, August 05, 2011

Road transport's most civilised invention?

Safe, reliable and direct pedestrian access to transport stops is a more fundamental need than niceties like seats, shelters and real-time displays. The latter all improve the passenger experience but are of little use if access to the stop is difficult.

There are numerous ways to improve pedestrian access across roads. Their cost varies from negligible (for a zebra crossing) to many million (for an overpass). In between are treatments involving median strips or traffic lights. Their effectiveness (as measured in terms of average and maximum wait times) likewise varies. There are also trade-offs, mainly based on whether we prefer roads to maximise throughput of car traffic or facilitate access via a variety of modes.

The table below compares how well various access measures assist the passenger cross a road to reach a stop. Assessment criteria included waiting time, whether guaranteed access is provided (a roundabout or an unsignalised road that offers an indefinite stream of unbroken cars does not, for example) and cost.

While each method has its strengths and weaknesses, the humble zebra crossing stands out as the highest and best means of pedestrian access. They are also very cheap to build. I would go so far to nominate the zebra crossing as road transport’s most civisised invention.

How did I reach the three criteria?

Firstly access time. A critical part of making public transport more attractive is speeding random-arrival end-to-end travel times. Along with bus or tram priority, out of vehicle components of travel time are the easiest and cheapest to improve through attention to co-ordinated timetabling, service frequency, easier interchange and pedestrian access.

Then there’s the measure of whether guaranteed access is provided. This affects all walking trips but is particularly critical for access to public transport due to its reliance on timetables and the large time consequences (up to an hour) if a service is missed. Unlike signalised intersections, underpasses and zebra crossings, busy roundabouts or unsignalised intersections with continually flowing traffic offer no such guarantees so are major barriers to pedestrian movement.

Lastly there is cost, for which I make no apology in including. Especially in a dispersed city with thousands of intersections, a large number of low-cost improvements would probably benefit more people than a few very expensive projects. For example, 100 new zebra crossings may be possible for the cost of a single elaborate pedestrian bridge, while for motorists there are similar trade-offs between new road/rail grade seperations (cheaper) versus new bypasses or freeways (more expensive). Bus/tram priority at intersections and roundabout removals are similar low cost/high gain projects.

Zebra crossings have another virtue in that they only slow road traffic when used. Due to this and their low construction cost it is best to err on the side of too many zebra crossings than too few. And what some may see as too many assists walkability as it makes a neighbourhood more permeable on foot.

If a new crossing is so heavily used that motor traffic is significantly delayed, it should not be condemned. Rather it is evidence that it filled a previous unmet access need and has encouraged people to walk rather than drive for short trips. Most people are pragmatic rather than ideological in their transport choices, so an improvement in walking access should result in more walking.

Zebra crossings undoubtedly have their detractors.

Busy crossings may be seen as impediments to motorists, much like pedestrians see roundabouts, freeways and long ramps to overpasses. For they, along with traffic calming and even some road rules, challenge the doctrine of the ‘open road’, a presumptuous and romantic relic from early last century’s gentleman motorists and their clubs.

Mid last century’s ‘scientific’ traffic engineers also had little time for the zebra crossing due to their ‘inefficient’ obstruction of car traffic. Grade seperating various road users became the fashion. Trams were to be dismantled or buried, cars kept at level and pedestrians confined to overpasses in various proposals for central Melbourne (including from the RACV). 1950s futurist images often showed a fourth level; a swarm of commuters in helicopters.

While private motoring took off, mass grade seperations in our city centres did not. Costs were prohibitive. Road – road and road – rail seperations in established areas often cause overshadowing and urban blight, with Sunshine, Oakleigh and Huntingdale being prominent Melbourne examples.

The closing decades of the 1900s was marked by a reaction against the traffic engineer’s dominance of city planning. Examples include the 1970s freeway revolts and 1990s ‘new urbanism’ movements. Urban amenity was considered sufficiently important for the King Street bridge in Melbourne’s CBD to be removed and levelled, while pedestrian access to Southbank was improved as part if its redevelopment.

Environmental, security and physical benefits are often-cited advantages of more people walking more often. However in daily life these ‘warm and fuzzy’ factors are unlikely to influence behaviour beyond the minority who think strongly about these things.

Much greater success is likely if walking becomes attractive to the ‘transport pragmatists’ who will use whichever mode best suits the trip at hand. It is here that road transport’s most civilised invention may have benefits disproportionate to its small cost.