Category: Uncategorized

A few years ago we did something that traditional mithai houses are not supposed to do. We wrote the recipes down.

All of them. Properly documented, standardised, held in a system rather than in a handful of heads — quantities, sequences, timings, the lot. It is the reason we can now produce almost any order within twenty-four hours instead of scheduling around who happens to be working that week.

The interesting part was not what went into the system. It was what would not go in.

Five names

The chain in our kitchen runs like this. Our head chef is Shakil Ustadh. He learned from Irfan Ustadh, over roughly ten years. Irfan learned from Yameen Ustadh. Yameen learned from S. Abdul Khaliq — the man Old Delhi knew as the halwa-maker by the Ghanta Ghar — and Abdul Khaliq learned from his father, S. Abdul Wahid.

Five names. Not five owners or five signatures on a document: five people who each stood next to the one before them, for years, at a hot pan.

Ten years is worth pausing on. That is not a course or a qualification. That is a decade of turning up to someone else’s kitchen and watching them work, and it is the standard unit of measurement in this trade. Everything else — the shop, the name, the city, the country — changed at least once. The pans moved twelve hundred kilometres in 1947.

What an ustad is, and is not

The word gets translated as “master,” which is close but carries the wrong weight in English. An ustad is not a title you are awarded or a rank you are promoted into. It is what you are called by the people who learned from you. You do not become one by being good; you become one by having taught someone who is.

The relationship on the other side is shagird — apprentice, though again the English is thin. A shagird does not enrol, does not pay fees, and is not taught in any organised sequence. For a long stretch at the start, they are barely taught at all. They fetch, they clean, they carry, they watch. The watching is the actual curriculum, and nobody announces this.

What the system holds

A great deal, as it turns out. Far more than the romantic version of this story would allow.

Ratios are ratios. Sequences are sequences. The order in which ghee goes into habshi halwa, the proportions of khoya to sugar, the point at which nuts are added, how long a batch rests — all of that is writable, and writing it down made us better rather than worse. A documented process is consistent across branches. It survives someone being ill. It lets a new person be useful in months rather than years.

Anyone who tells you traditional food cannot be systematised is usually protecting something other than the food.

What would not go in

Then there is the residue — the part of the process that resisted every attempt to turn it into an instruction.

Habshi halwa takes about nine hours, and the wheat is sprouted and fermented before cooking even begins. How far to let that fermentation go is the first thing that will not reduce to a number. It is a judgement about smell and weather. Two days in Karachi in June is not two days in Karachi in January. You can write “48 hours” into a system, and the system will be wrong for most of the year, and someone still has to walk over and decide.

The milk reduces for hours, and readiness is a matter of how it moves — how it falls off the edge of the stirrer, how the surface behaves when the pan is disturbed. There is a sound to a pan that is going well and a slightly different sound to one about to catch. An ustad hears it several seconds before a shagird sees it. Those several seconds are not in any document we have written, and we tried.

The colour has to be judged in whatever light the kitchen happens to have that day. The final texture is assessed by dragging a spoon through and watching how slowly the channel closes.

None of this is mystical. It is ordinary expertise, the same kind a good carpenter or a good doctor has, and it has the same property: the expert cannot fully articulate it, because most of it stopped being conscious years ago. Writing the recipes down did not capture it. What writing them down did was show us precisely where it lives, which we had never had to locate before.

The man who carried it across

There is one link in this chain that deserves more attention than it has ever had.

Yameen Ustadh learned his craft in Delhi, in the shop beside the Ghanta Ghar, before 1947. He learned it directly from Abdul Khaliq. And when Partition came and the family’s world was cut in half, the knowledge did not travel only in the family — it travelled in him too.

Think about what that means. A recipe that had been in one family for generations survived the loss of its shop, its city and its country partly because a man who was not related to anyone in that family had spent years learning to hear a pan properly, and then went on teaching what he knew. Half of what we make today reaches us through him rather than through our own bloodline.

Farrukh knew him. Yameen Ustadh died around twenty years ago, and the man he remembers is easy enough to picture: a simple man, bearded, short and slenderly built, wheatish in complexion. Mild in his manner. He worked away quietly and made very little fuss about anything.

Except the halwa. About the halwa he was pedantic to a degree everyone in the factory noticed — every small detail had to be right, and he would not be moved past one that was not.

That combination is worth sitting with, because it is the whole of him. A man with no interest in asserting himself anywhere else, who was immovable about the temperature of a pan.

He also kept a cupboard.

It sat in the factory and he had the only key. Inside were the ingredients that went into habshi halwa — the specific things, in the specific proportions, that made ours ours. Nobody else on the floor knew what was in it. Irfan knew, because Irfan was his shagird. The family knew. That was the entire circle.

It is difficult to convey to anyone who has not worked in a kitchen like that how normal this was. Knowledge was not a document, it was a responsibility, and it was held by a person who had earned it. A locked cupboard was not paranoia. It was the physical form of a trust, and Yameen carried it for decades without ever writing any of it down.

No photograph of Yameen Ustadh is known to exist. What survives of his face is the paragraph above, held between the two people who worked beside him. Irfan Ustadh is still with us, though he is old now and no longer at the pan.

A sixth name, and a shorter line

The chain above is the working chain — the one that runs through the kitchen and produces what we sell. But it is not the only line back to Delhi.

Farrukh Ahmed, who runs the business today, learned this craft himself. He learned it from Irfan Ustadh, and he learned it from his own grandfather, S. Abdul Khaliq — the man behind the counter at the Ghanta Ghar.

That is a shorter line to Delhi than anyone else alive has. Yameen learned from Abdul Khaliq and passed it to Irfan, who passed it to Shakil: three steps. Farrukh learned from Abdul Khaliq directly: one.

It is also, incidentally, why the documentation project happened at all. Someone who has stood at the pan knows which parts of the process are genuinely load-bearing and which are habit. You cannot systematise a craft you have only managed.

Fifteen people

There are around fifteen people training in our kitchen right now.

We mention the number because the usual version of this story ends with a lament — the old ways dying, nobody willing to learn, the last ustad taking it all with him. That is a real problem in a good many trades. It is not ours, and pretending otherwise would be a more flattering story than the true one.

What the system did was change what the training is for. When the recipes lived only in people’s heads, a decade of apprenticeship went partly into memorising things a document can hold perfectly well. Now it does not have to. The ratios are written down, so the years go into the part that is not: learning to smell fermentation, hear a pan, read a colour in bad light.

Put plainly, writing the recipes down did not replace the ustads. It got out of their way.

The chain still has to be walked, one person at a time, next to a hot pan, for years. What has changed is that fifteen people are currently walking it, and the thing they are learning is now the thing that actually needs a human.

The cupboard and the system

It is worth sitting with the distance between those two things.

For most of this business’s life, what made our habshi halwa ours was kept in a locked cupboard by a gentle man who had the only key, and the circle who knew its contents could be counted on one hand. If Yameen had been hit by a bus, a meaningful part of a two-hundred-year-old craft would have gone with him. Everyone accepted that, because there was no alternative anyone could imagine.

Today the same knowledge is documented, standardised, and legible to anyone in the kitchen who needs it. Nothing is behind a lock. If any one of us disappeared tomorrow, the halwa would be identical on Thursday.

Something was lost in that, and it would be dishonest to pretend otherwise. There was a dignity in the cupboard — in knowledge as something entrusted to a person rather than stored in a database. Yameen’s authority came from what he carried.

But rather more was gained, and the thing that was gained is the thing he actually cared about: the halwa still tastes the way he made it, and it will keep doing so long after everyone who remembers him is gone. He spent his life being a single point of failure for a recipe. Nobody has to do that any more.

What could not be put in the system — the smell of the fermentation, the sound of the pan — is still being handed over the way he handed it to Irfan. Fifteen people at a time.

Which is why we wrote down what he looked like. Bearded, short, slender, wheatish, mild — and completely unyielding about a detail nobody else in the room could see. It is not a photograph. It is what there is, and it is now written somewhere other than in two men’s memories.

If you knew Yameen Ustadh, or worked in the Ghanta Ghar shop before 1947, or your family kept a chain like this one, we would like to hear from you.

Almost everyone ordering wedding mithai for the first time asks the same question, and it is the wrong one.

The question is “how much mithai do I need for 400 guests?” The answer is that guest count is not what determines the order, and working from it is how people end up with either an embarrassing shortfall or forty kilos nobody eats.

Mithai at a Pakistani wedding does not work like catering. It is not portioned per head at a table. It moves in boxes, to households, as an obligation — and the number that matters is how many households you are sending to, not how many people sit down to dinner.

Count households, not guests

Write out the list of families you are sending mithai to. Not individuals — families. A couple with three children and a grandmother is one box, not five. The uncle who is coming alone but whose married sons are not invited may be three boxes, because you are sending to their homes too.

That list is almost always shorter than the guest list and longer than people expect, because it includes households who are not attending at all: relatives in other cities, the family your mother has exchanged sweets with for thirty years, neighbours, your father’s colleagues, the people who sent mithai at your cousin’s wedding and must now be sent it back.

Do this list first, in writing, before you look at a single catalogue. Everything else follows from it.

The points where mithai actually appears

Most Pakistani weddings involve mithai at four or five distinct moments, and they are not interchangeable. Ordering everything at once as a single undifferentiated quantity is the second most common mistake.

The rishta and the mangni. The first exchange, and the most scrutinised. This is not a distribution — it is a small number of arrangements, presented, and looked at closely by people forming a first impression of your family. Quality matters more than quantity here by a wide margin.

Mayun and mehndi. Informal, high-energy, and typically served rather than boxed. Smaller pieces, things people can eat standing up with their hands.

The main event. Where the bulk of the boxed distribution happens, and where the household list does its work.

The walima. Frequently a second, smaller round for the groom’s side’s own circle.

The sending-out. The one people forget until three days beforehand. Boxes going to households that are not attending, often in other cities, often needing to travel. This is where shelf life and packing stop being details.

Working quantities

As a starting point, for boxed distribution to households:

• Close family and the households whose opinion you actually care about: 1–2 kg per household.
• The wider circle — neighbours, colleagues, extended relatives: 500 g to 1 kg.
• Courtesy sends, where the gesture is the point: 250–500 g.

Then add ten to fifteen percent. Not because the arithmetic is wrong, but because it is never the arithmetic that fails — it is the four households somebody remembers on the day. Running out is a genuine social problem in a way that a small surplus is not.

For served mithai at mehndi or mayun, plan on roughly 100–150 g per attending guest, and expect that to skew high if the event runs late.

Choosing by how far it has to travel

The practical constraint nobody mentions is that different mithai keeps for very different lengths of time, and wedding orders almost always involve something sitting in a car in the sun.

Habshi halwa and the dense halwas keep for around a month without refrigeration and travel extremely well. This is why they have historically been the sweet you send to another city, and why they are the safest choice for the sending-out list. They also read as serious — a slab of properly made habshi halwa is understood by anyone who knows mithai to be a nine-hour product, and that is legible to the recipient in a way that a mixed box is not.

Khoya-based mithai — barfi and its relatives — is best inside a few days and much happier in winter than in summer. Ideal for local distribution, risky for anything going far in warm weather.

Anything with fresh cream or chhena should be treated as same-day. Beautiful on a table at the event; a bad idea in a box on a five-hour drive.

If your wedding falls in the hot months, weight the order toward the things that keep and be honest with yourself about how long boxes will sit before they are opened.

When to order

Here we should be straight about something, because a lot of wedding advice online assumes constraints that do not apply to us.

We can produce most wedding orders in 24 hours, habshi halwa included, in the middle of season. Orders above roughly 500 kg, or with substantial customisation, need 48. Production capacity is genuinely not your bottleneck.

That matters because the thing that actually goes wrong with wedding mithai is never the kitchen. It is the list. Households get remembered late, guest numbers move in the final fortnight, and somebody’s side of the family turns out to have been counted differently by two people. If production took six weeks, every one of those would be a disaster. It takes a day, so they are phone calls.

What is worth doing early is the deciding, not the ordering:

• Early: write the household list and agree it with whoever else has opinions. This is the slow part, and nobody can do it for you.
• Early: settle packaging and any customisation. Design decisions have longer lead times than the sweets do.
• A week out: confirm quantities and delivery dates per event.
• Any time after that: additions, within the 24 to 48 hour window above.

One scheduling note worth flagging for anyone planning around early 2027: Ramadan is expected to begin in February, and weddings generally do not run during it. That compresses the whole season into the months before, so the December and January weeks will be busier than usual. If your wedding falls in that window, bring every date above forward.

The mistakes that cost people

Ordering from the guest list. Covered above, and worth repeating because it is the one that causes real distress.

Treating every household identically. A single box size across the whole list is efficient and slightly insulting at both ends — too little for the people closest to you, too much for a courtesy send.

Leaving packaging to the end. How a box looks when it is handed over is a substantial part of what you are actually giving. Decide this early, not in the last week when options have narrowed.

Forgetting the out-of-city list entirely. This is the single most common late panic, and it is also the hardest to fix quickly because those boxes need to travel.

Panicking about production time instead of the list. People routinely place a rushed, badly thought-through order six weeks early because they are afraid of missing a deadline that does not exist. The sweets take a day. The list takes a month of arguing. Spend your time on the part that is actually hard.

Talking to us

Bring the household list, the dates of each event, and where the boxes need to reach. From that we can usually put together a plan in one conversation, including what to send where and what will still be good when it arrives.

We have been making mithai for Karachi and Lahore weddings for a long time, and the recipes we work from came down the same way the advice above did — through people who have watched a great many families get this right and a few get it wrong.

In a lane off Ballimaran in Old Delhi, a man named Firoz Ahmed spends nine hours at a time stirring a copper pan of habshi halwa. He is not related to our family. He has probably never met anyone from it. But the recipe he works from came to him through a chain of hands that begins in our great-grandfather’s shop, and the packaging he sells it in carries a picture of the Ghanta Ghar clock tower — the landmark our family’s shop stood beside, and the name Delhi still uses for the halwa.

That is the strange afterlife of a recipe. Partition split a family business across a new border in 1947, and eighty years later both halves are still making the same sweet, twelve hundred kilometres apart, neither quite aware of the other.

The shop by the clock tower

Before Partition the family shop stood in Chandni Chowk, in the stretch between Kucha Rehman and Kucha Ustad Daag, close enough to the Town Hall clock tower that customers stopped using the family name altogether. They asked for the halwa from Ghanta Ghar. The shop’s proprietor, Abdul Khaliq, became known across Old Delhi simply as Ghantaghar Habshi Halwa Wale — the habshi halwa man from the clock tower.

By family account, the business had been there a long time before that. We say 1830, and we should be honest about what that number is: it is family tradition, carried down in the telling rather than in a document. We have not found a deed or a ledger that fixes the year, and we would rather say so than pretend to a precision we cannot show. What is not in doubt is that the shop was long established well before Partition, that it sat beside the Ghanta Ghar, and that its halwa was well enough known for the neighbourhood to name it after a clock tower.

1947, and a recipe left behind

When Partition came, Abdul Khaliq migrated to Pakistan. He did not take the shop with him, and he could not. What he left behind were his workers — the men who had stood at those pans beside him, who knew the timing and the colour and the smell of the thing better than any written recipe could hold.

They kept making it.

The chain is documented by Delhi Food Walks, who traced it in detail: one of Abdul Khaliq’s workers taught the recipe to a man named Hanif, and Hanif taught his son Firoz. Firoz Ahmed makes it today at Punjabi Phatak in Ballimaran. Hanif’s brothers opened Hanif Dairy in Gali Qasimjan, near the Hamdard Dawakhana. Taqi Ahmed opened Ahmed Dairy nearby. An entire small constellation of Old Delhi halwa shops traces back to one shop’s kitchen staff, still working from what they learned there.

We find this genuinely moving rather than proprietary. Nobody in Delhi owes our family anything. The recipe was not stolen; it was left, in the only way a recipe can be left — in the hands of the people who already knew it. That Firoz Ahmed puts the Ghanta Ghar on his packaging is not a claim against us. It is the same act of remembering that made us keep the name on our side of the border.

The other half of the chain

In Karachi the family started again. The knowledge that travelled was the same kind that stayed behind: not written down, but held in the hands of the men who had learned it.

Our current head chef is Shakil Ustadh. He learned from Irfan Ustadh. Irfan learned from Yameen Ustadh. Yameen learned from S. Abdul Khaliq who learned from his father S. Abdul Wahid himself. Five names, and every link in that chain is a person who stood next to the previous one at a hot pan for years until the timing became something in the wrists rather than something in the memory.

This is why we are named for Abdul Wahid rather than for his son. Abdul Khaliq is the man Delhi remembers, because he was the one behind the counter when the shop became famous. Abdul Wahid was his father, further back up the same line — and the person from whom the transmission we can still trace, hand to hand, actually begins.

What habshi halwa actually is

For anyone who has only encountered the word on a menu: habshi halwa is not a quick sweet. Ours takes about nine hours in the pan, and there is no way to shorten it that does not produce something else.

It begins with wheat, sprouted and left to ferment before it is ever cooked — that fermentation is where the deep, faintly sour backnote comes from, the thing that separates habshi halwa from the sweeter halwas it sits beside. Milk is reduced alongside it for hours until it thickens and darkens. Ghee goes in slowly, in stages, and the mixture is worked continuously; stop stirring and it catches on the bottom and the whole batch takes on a burnt edge you cannot cook out. Khoya, sugar and nuts come in toward the end. The colour deepens from pale to a dark, almost mahogany brown, and the texture turns dense and elastic enough that a spoon dragged through it leaves a channel that closes slowly.

The name is generally understood to refer to that darkness. It keeps for about a month without refrigeration, which is why it was historically a winter sweet, and why the Delhi shops still sell most of theirs between October and March. It travelled well, which mattered when sweets moved by train.

Why we are writing this down

Most of what we know about our own history exists as things said at family gatherings. That is normal for businesses of this age in this part of the world, and it is also fragile. The generation that could describe the Chandni Chowk shop from memory is nearly gone.

So we are writing down what we have, marking clearly which parts are documented and which are family tradition, and we would rather be corrected than be tidy. If you have photographs of the Ghanta Ghar shop, or a relative who bought halwa there before 1947, or anything at all that touches this story — we would genuinely like to hear from you.

And if you are in Old Delhi, go to Ballimaran and buy some halwa from Firoz Ahmed. It is the same recipe. It has simply been in other good hands for eighty years.

---

A note on sources. The account of what became of the Delhi shop after 1947 — the recipe passing from one of Abdul Khaliq’s workers to Hanif, and from Hanif to his son Firoz Ahmed, along with the shops at Punjabi Phatak, Gali Qasimjan and nearby — is drawn from Delhi Food Walks, who documented that chain on the ground in Old Delhi. We are grateful for their work. Without it we would have no independent record of what happened to the shop our family left behind.

The rest — the 1830 date, the family lineage, the chain of ustads — is family account rather than documented history, and is marked as such above. If you can correct or add to any of it, we would like to hear from you.