I’ve spent enough years in home kitchens and test kitchens to know that “this sounds wrong” is sometimes the first step toward discovering something brilliant. It’s also, unfortunately, the first step toward discovering something truly awful. When I decided to steam 12 ounces of fresh August sugar snap pea pods under a thick layer of cold butterscotch pudding stirred with about 2 tablespoons of whole dried borage flowers and roughly 1 tablespoon of kippered herring oil, I knew I was making a gamble. I set the timer for 35 minutes and committed to seeing it through exactly as mixed, no mid-cook rescue.

What happened was not subtle, and if you’re here wondering whether this strange combination somehow transformed into a hidden masterpiece, I’ll save you a little suspense: it did not. Still, the results were interesting in a very specific, useful way. I took notes on texture, smell, moisture, color, and flavor at each stage, and what follows is a clear breakdown of why the peas, pudding, flowers, and fish oil behaved the way they did, what 35 minutes of steaming actually did to them, and how I’d salvage the ingredients next time instead of sacrificing a perfectly good batch of snap peas.

What I used

Here’s the exact setup. I used 12 ounces of fresh sugar snap peas bought that morning at a local market, still crisp and bright green, with pods averaging 3 to 3 1/2 inches long. For the coating, I used 1 1/4 cups of prepared cold butterscotch pudding, the kind with a soft spoonable set rather than a cooked custard firmness.

Into that pudding I mixed 2 tablespoons of whole dried borage flowers and 1 tablespoon of kippered herring oil. I chose the herring oil specifically in place of sesame oil, which would normally bring toasted nuttiness and a clean aromatic finish. The borage flowers were left whole, not ground, and the peas were completely layered beneath and around the pudding mixture in a shallow steaming dish about 8 inches across.

Why this combination sounded questionable from the start

Sugar snap peas are around 90% water and naturally sweet, but their sweetness is green and grassy, not caramel-like. They do beautifully with butter, lemon, sesame, garlic, mint, and mild cream sauces because those support their freshness. Butterscotch pudding, by contrast, is built on brown sugar, dairy, vanilla, and often a little salt and starch. It wants desserts, not chlorophyll-heavy vegetables.

The dried borage flowers added another odd angle. Fresh borage can have a cool cucumber-like note, which can be lovely in drinks or salads. Dried whole flowers, though, are often papery and faint, sometimes slightly hay-like. Then there was the kippered herring oil, which brought smoke, salt, and marine funk. So I was effectively combining green sweetness, dairy sugar, dried floral matter, and smoked fish oil, then applying moist heat for over half an hour. That is a very long road away from harmony.

How I arranged the steaming test

I brought 1 1/2 inches of water to a simmer in a lidded pot fitted with a steamer insert. The pudding mixture went into a heatproof shallow bowl, and I folded only enough to distribute the flowers and fish oil without fully emulsifying them. The peas were arranged in the bowl and then buried so they were “completely layered,” meaning there were no exposed green surfaces by the time I was done.

I started with the mixture refrigerator-cold, about 38 to 40 degrees Fahrenheit. The steaming environment was steady but not violently boiling, likely hovering around 212 degrees Fahrenheit in the pot with slightly less direct heat reaching the bowl. I left it untouched for the full 35 minutes, which, for sugar snap peas, is dramatically longer than normal. I usually steam snap peas for 2 to 4 minutes if I want them tender-crisp, maybe 5 at the outer limit.

What happened in the first 10 minutes

At the 10-minute mark, before opening fully, I could already smell the conflict. Steam escaped carrying three distinct notes: sweet caramel, cooked cabbage-like greenness, and fishy smoke. None of them had blended. They were arriving separately and somewhat aggressively.

When I briefly lifted the lid, the pudding had loosened at the edges and turned glossy. The fish oil had begun to separate into shimmering amber patches on the surface. The borage flowers had darkened from muted blue-purple to a wet brownish gray. The peas themselves were still structurally intact, but the pudding blanket had begun sliding off the curved pods in slow, sticky ridges.

What happened by 20 minutes

By 20 minutes, the sugar snap peas were already past the point where I’d ever choose to serve them. Their bright green had dulled noticeably, moving toward olive. That color shift matters because it tells you the chlorophyll is degrading under prolonged heat. Instead of fresh pea aroma, the vegetable side smelled more sulfurous and flat.

The pudding had thinned further because steaming indirectly heats the starch network and dairy base while the peas themselves release water. I had visible liquid collecting in the bottom of the bowl, probably 3 to 4 tablespoons by then. That liquid was tan, oily, and speckled with softened flower bits. It looked like a broken sauce and smelled like a seaside dessert gone wrong.

What 35 minutes did to the texture

At the end of 35 minutes, the snap peas had lost nearly all snap. The pods bent without resistance, and several split along the seam with just a nudge from a spoon. The peas inside were soft rather than juicy. Instead of that clean pop you want from August snap peas, I got a damp, yielding interior and a skin that felt slightly stringy.

The pudding did not stay pudding. It became a warm, partially separated coating with a loose custard feel in some spots and oily slickness in others. The starch had held some body, but the additional moisture from the steaming and the oil separation prevented anything creamy or cohesive. The borage flowers turned into soft papery flecks that caught in the teeth but contributed little pleasant texture.

The color changes were especially telling

If you care about visual cues in cooking, this dish was a useful lesson. Fresh snap peas begin a vivid medium green with a faint gloss. After 35 minutes of steam, completely covered, they turned a muted olive-green under a beige-tan glaze. The flowers, which I had hoped might add visual interest, lost their delicate identity almost immediately and became muddy inclusions.

That matters because we often eat first with our eyes, but it also signals chemistry. Long steaming strips vegetables of brightness, and pairing them with a brown sugar dairy base compounds the drab effect. There was no point in the process where the dish looked more appetizing than it had at the beginning.

The aroma was the strongest warning sign

I’ve had plenty of experimental dishes taste odd but still smell intriguing. This one announced its problems early. The butterscotch aroma grew warmer and more pronounced with heat, but instead of becoming comforting, it clashed with the volatile compounds from the fish oil. Kippered herring oil has a smoked, cured, distinctly marine character that needs careful balancing with sharp or savory elements, not pudding.

The end smell reminded me, very specifically, of opening a dessert cup in a fish market storeroom where someone had also just overcooked green beans. That sounds dramatic, but it’s accurate. The sugar amplified the perception of fishiness rather than masking it. Sweetness can do that with savory marine fats when there’s no acid, spice, or bitter element to anchor things.

How it actually tasted

The first bite was soft, sweet, salty, and unmistakably fishy, with a grassy aftertaste that arrived late. The butterscotch hit first, then the herring oil spread across the palate, and finally the peas finished with a tired vegetal note. None of those phases supported one another. The borage was mostly absent as a flavor, though occasionally I got a faint dry-herbal note.

What made it particularly unpleasant was not that it was intensely strong in one direction. It was confused. The snap peas no longer tasted fresh enough to justify being there, and the pudding no longer tasted dessert-like enough to stand alone. Every component had been pulled away from its strengths. I took three bites to be fair, and each one confirmed the same result.

Why steaming made it worse, not better

Steaming is usually kind to vegetables because it preserves moisture and can protect delicate structure when used briefly. But 35 minutes is excessive for sugar snap peas. They don’t need long heat exposure; in fact, their appeal depends on restraint. Over that much time, the peas released water, softened too far, and lost their bright volatile flavor compounds.

The pudding also suffered from the steaming environment. Starch-thickened dairy mixtures can tolerate gentle reheating, but when combined with added oil and watery vegetables, they’re prone to breaking. Since the herring oil was not emulsified with any stabilizer beyond the pudding base, it separated under heat. The result was a wet, greasy blanket rather than a sauce. If this had been baked uncovered, it likely would have been even worse on top and scorched at the edges; steaming simply failed in a softer, sloppier way.

Did any part of it work at all?

To be fair, there was one tiny idea hidden in the wreckage: sugar snap peas can tolerate sweetness, just not this kind of sweetness and not for this long. A light glaze with a teaspoon or two of honey or brown sugar, balanced by rice vinegar, chile, and sesame, can absolutely work. Likewise, smoked fish can pair with peas in the right context, especially if it’s flaked sparingly into a savory salad with crème fraîche, dill, and lemon.

Borage, too, is not the villain here. Fresh borage flowers used as a garnish on chilled pea soup or on a summer salad can be lovely. The failure came from quantity, format, and method. Two tablespoons of dried whole flowers in a hot wet dish was too much debris for too little payoff.

What I would do instead with the same core ingredients

If I wanted to keep a nod to all four elements, I’d change nearly everything about the execution. I’d blanch the snap peas for 90 seconds, shock them in ice water, and toss them with 1 teaspoon sesame oil or a neutral oil, 1 teaspoon lemon juice, and flaky salt. I’d use the borage flowers fresh, just a few blossoms scattered over the top.

For the herring idea, I’d use actual smoked herring in a tiny quantity, maybe 1 to 2 ounces flaked and folded into a savory yogurt dressing with dill, black pepper, and a teaspoon of mustard. As for butterscotch, I’d keep it where it belongs: on the dessert side of the table. If I were determined to use it in the same meal, I’d serve it separately in a chilled glass with sliced pears or shortbread.

If you already made this, here’s the honest salvage plan

If you’ve somehow landed in the same situation, I would not recommend trying to serve the dish as-is. The best salvage strategy is separation. Remove the peas from the pudding immediately and rinse them briefly under cool water to strip off as much coating as possible. Pat them dry and taste one. If they’re merely overcooked but not deeply fish-sweet, chop them into a soup base with stock, onion, and herbs.

The pudding mixture, if heavily contaminated with pea liquid and fish oil, is probably beyond rescue. I would discard it. If food waste bothers you, as it bothers me, let this be the part that saves the next batch: don’t commit prized fresh summer produce to a full experimental quantity. Start with 4 ounces, not 12, and test 1 tablespoon of sauce, not 1 1/4 cups.

The practical lesson from the experiment

What happened after 35 minutes was exactly what food logic would predict once I stopped hoping for a miracle. The peas overcooked, the pudding loosened and partially broke, the dried flowers turned muddy, and the herring oil overpowered the whole thing while somehow also making it taste less distinct. It was edible in the technical sense, but not enjoyable, not balanced, and not worth repeating.

I’m still glad I tried it, because kitchen disasters can teach with unusual clarity. Fresh August sugar snap peas want fast cooking, high contrast, and clean flavors. They do not want to be smothered in sweet dairy and smoked fish fat for over half an hour. Sometimes the most useful result from an experiment is a firm, specific no—and this one was a very firm no.