What strikes you first in the old town is the straightness. Streets meet at right angles, the blocks are rectangles, and after two hundred and fifty years the plan still holds: you can walk from the sea to the edge of the historic core without meeting a curve. Every account of the place calls this the French grid, and the grid is indeed the reason the town reads as a single idea rather than a accretion.

But the grid is not the interesting part. The interesting part is the one line in it that is not a street.

The grid was drawn by the Dutch

The orthogonal plan does not come from France. It dates from the Dutch occupation of 1693 to 1699, when the town was taken from the French and a comprehensive grid was drawn up: large rectangular blocks, straight streets crossing at right angles, a simple structure that could be parcelled out to different communities. When the French regained the territory they carried the Dutch plan to completion rather than replacing it. The most French thing about Pondicherry's Frenchness was designed by somebody else.

The same is true of the boulevards. The four wide avenues that ring the historic core, and give the whole quarter its name, were not laid out as promenades. They were the fortifications: ramparts, bastions and a moat, in the manner of Vauban. The British destroyed them in 1820, and the town could not afford to rebuild. The moat was filled, or left to silt into a lane, and the circuit of the walls became a ring road. Every photograph of a tree-lined boulevard in this town is a photograph of a defensive ditch that nobody had the money to restore.

The canal was a drain before it was a border

The Grand Canal was cut around 1765 as a piece of storm-water engineering. It runs about 3.4 kilometres, from the Uppar — a tributary of the Sankaraparani — to the sea near the new lighthouse, and its job was to take the monsoon off the town: rainwater north into the Uppar drainage, rainwater south into the bay. The grid was laid so that every street drained into it, which is why the canal is the spine and the streets are the ribs.

Three details make it a machine rather than a ditch. First, the slope: a canal that drains by gravity needs a fall along its whole length, and that fall is the design. Second, the sluices. There were three of them on the seashore, and their purpose was to let sea water in — the channel was flushed by admitting the tide and then releasing it, so that the silt went out with the ebb. Third, the culverts, thirteen of them, carrying each street's drain through the banks.

A drain with sluices is not maintained by sweeping. It is maintained by the moon. Twice a day the tide offers to do the work, and all the town has to do is keep the gates operable and the bed clear enough for the water to move fast. It is an elegant arrangement, and it is completely invisible in a plan: on the map the canal is a strip, and the strip looks the same whether or not it works.

The grid, the canal, and the three sluices — a plan whose only non-street is a drain THE OLD TOWN, SCHEMATIC SEA Ville Noire dense lanes Ville Blanche gardens, wide streets three sluices, now gone from the Uppar The canal does three jobs and is one trench: it drains the town, it is flushed by the tide, and it separates the two quarters. 1765 drain, with sluices 1979 sewer for 117,000 2000s a stretch covered Each successor kept the line and dropped a working part. The line is what the map shows.
The blocks on the two banks are drawn at their real relative density: the same trench, a different grid on each side of it.

Two jobs, one trench

Because the canal already existed, and already ran between the sea and the inland, it made a convenient place to put people on either side of. To the east, facing the water, the Ville Blanche: wide streets, gardens behind high compound walls, villas with colonnaded porticoes. To the west, the Ville Noire: narrower lanes, denser blocks, row houses around temples that predated the French. The historians differ on which came first, the drain or the cordon, and on how sharply the line was enforced at any given decade. What is not in doubt is that the town got two things out of one excavation, and that the second thing cost nothing extra. A drain needs a line anyway; putting the social border in it means you do not have to build a wall.

That is the part worth sitting with. The engineering was not neutral and then later put to a political use. The engineering was cheap because it did both. A ditch that has to be dug anyway is a free border, and a border that has to be maintained anyway is a free drain. This is a common shape in colonial town planning — separate the two things you want separated with the one piece of infrastructure you were going to build regardless — and it is a reminder that the most efficient designs are often the ones whose efficiency is the problem.

Then the drain stopped draining

An underground sewerage system was built in 1979, designed for a population of 117,000. The town has long since passed that. When the sewers surcharge, the manholes overflow into the roadside drains, and the roadside drains — which are the grid's ribs — carry it into the canal. A survey by INTACH counted about 250 houses discharging waste directly into the channel. The sluices, which were the canal's self-cleaning mechanism, have been removed or damaged. A long stretch was covered over in the 2000s and built on: public toilets, parking, shops.

So the thing that was designed to take water off the streets is now, in the words of one urban planner quoted in the local press, a sewage line through the middle of town. And the streets that flood when it rains are the ones at the low end of the grid. The cause and the effect are the same piece of infrastructure, in two states.

What a restoration can and cannot put back

The canal is now under restoration: about ₹21 crore under the Smart Cities Mission, with INTACH as architectural consultant, and a stated intention to carry out the work without changing the original design. The scope is revealing, because it is mostly about the surface — clearing encroachments, a granite parapet with benches, pedestrianising the banks, facade improvement, avenue trees, planter boxes, and rebuilding the thirteen culverts.

Every one of those items improves the town. None of them is the thing that made the canal work. The working parts were a continuous fall along 3.4 kilometres, a set of gates that could be opened to the tide, and a desilting schedule frequent enough to keep the bed clear — and of those three, only the gates could even be drawn on a plan, and they no longer exist. A parapet, a bench and a row of trees restore the canal as an object. They do not restore it as a drain.

The temptation to say that the restoration will fail is one I want to resist, because the works are recent and no post-project drainage performance has been published. But the structure of the risk is clear enough, and it is the pillar's pattern exactly: the part that is visible — the line on the map, the alignment, the name, the banks you can walk along — is the part that gets conserved, because it is the part that can be seen. The part that did the work was a gradient measured in centimetres per metre, a tide, and somebody's job of clearing the bed twice a year. None of that is a shape. All of it is what a canal is.

Sources and checks

  1. The Hindu Once a stormwater drain, now an open sewer — the Grand Canal in Puducherry
  2. The Hindu After years of neglect, restoration of Grand Canal begins — ₹21 crore under the Smart Cities Mission
  3. INTACH Puducherry Urban heritage, mass tourism and gentrification in the Boulevard town of Pondicherry
  4. Sahapedia The Grid and the Canal — the layout of the heritage town
  5. Institut Français de Pondichéry The plan of a town — the morphology of Pondicherry and the covering of the Grand Canal
  6. Puducherry Smart City Development Ltd Grand Canal rejuvenation — project scope, culverts and public consultation

Last checked 3 November 2026 Yellowsea Huang Corrections