About
Winemaking Recipes: an apology
Winemaking Recipes: An Apology — and What It Teaches Us About Knowledge Loss
Home winemaking runs on shared knowledge — recipes passed between neighbors, typed up late at night, posted to early internet forums by people who just wanted to help. That knowledge is fragile. A single server fire can erase years of work in seconds, and no backup system can fully replace the hours a person spent writing it all down. Jack Keller’s winemaking site was one of the most generous resources the home winemaking community ever had, and part of it burned — literally — when a server rack caught fire and took 720 recipes with it. That loss is worth understanding, because it points to something important about how we learn to make wine in the first place.
What Actually Happened
Jack Keller spent years building an alphabetical recipe library across two web servers. Most of the content lived on his internet service provider’s server. When that facility suffered a fire, the data was gone — no cloud backup, no redundant copy, no recovery option. Over 700 recipes and 14 pages on grape growing disappeared overnight.
He estimated it would take three to four months to reconstruct the missing material from scratch. That is not an exaggeration. Writing a single well-tested recipe takes time. Writing 720 of them — with accurate measurements, ingredient notes, process steps, and the kind of detail that actually helps a beginner succeed — is a serious body of work. He chose not to redo it, and that is a completely reasonable call.
Instead, he shifted his approach. Rather than rebuilding the full catalog, he started posting recipes on request. Someone would email him asking for a specific fruit wine or mead, and he would write it up. The result was a “requested recipes” section that grew organically based on what people actually needed, rather than an exhaustive list that tried to cover everything upfront.
Why Recipe Gaps Don’t Have to Stop You
Here is something worth knowing: you do not need 720 recipes to make great wine at home. You need to understand a handful of core principles, and then any recipe — even an incomplete one — becomes a starting point rather than a strict set of rules.
Most fruit wine recipes follow the same basic logic. You start with fruit, add water to dilute it to a workable volume, adjust the sugar content (called the Brix or specific gravity, measured with a hydrometer) to hit a target alcohol level, add acid and nutrients to keep the yeast healthy, pitch yeast, ferment, and then clarify and bottle. The ingredients change. The process barely does.
When a specific recipe is missing, you can often reverse-engineer it from that framework. What does this fruit taste like? Is it high acid or low? Is it naturally sweet or tart? Does it have enough fermentable sugar on its own, or will you need to add some? Answering those four questions gets you 80 percent of the way to a working recipe even with no reference at all.
The Real Value of a Recipe Collection
So if recipes are not strictly necessary, why did Jack spend years building one? Because experience is the fastest teacher, and a well-written recipe is someone else’s experience compressed into a readable format.
When you read a recipe that says “use 4 pounds of elderberries per gallon,” that number did not come from a chemistry textbook. It came from someone who tried 3 pounds and got a thin wine, tried 5 pounds and got something too tannic, and landed on 4 as the sweet spot. That is a lot of fermentation cycles compressed into a single line of text.
The loss of those 720 recipes was not just a loss of instructions. It was a loss of accumulated trial and error — the kind of knowledge that takes years to rebuild. That is why it mattered, and why Jack’s apology was genuine rather than just a formality.
How to Move Forward Without a Complete Archive
If you are looking for a recipe that no longer exists online, here are some practical strategies that work.
Search by ingredient, not by name. A recipe filed under “Serviceberry Wine” might also appear under “Juneberry Wine” or “Saskatoon Wine” because the same fruit has a dozen regional names. Search all of them.
Use a base recipe and adjust. Most fruit wines work on a ratio of roughly 2–4 pounds of fruit per gallon of must (the unfermented juice-and-water mixture). Light, delicate fruits like peaches sit toward the lower end. Dark, tannic fruits like blackberries or elderberries sit toward the higher end.
Ask in a community. Winemaking forums and clubs have members who have been at this for decades. Someone has almost certainly made what you want to make and will share their notes.
Document your own process. Every batch you make is a data point. Write down what you did, what it tasted like at racking (the process of moving wine off its sediment), and how it changed in the bottle. After ten batches, you will have built something more useful than any recipe: a working mental model of how wine actually behaves.
Why This Works
Think of a winemaking recipe the way Alton Brown thinks about cooking techniques: the recipe is the map, but understanding the terrain is what gets you home. Maps burn. Terrain knowledge doesn’t.
Once you know that yeast needs sugar to produce alcohol, that acid keeps bacteria in check, that tannins give wine structure and help it age, and that clarifying agents like bentonite (a type of clay) attract and drop out suspended particles — you can adapt to missing information on the fly. The recipe tells you what someone else did. The principles tell you what you should do given what you have in front of you right now.
Jack Keller’s lost recipes were a real loss. But the winemaking knowledge that produced them — the hands-on understanding of how fruit, sugar, water, and yeast interact — that knowledge is still out there, and it is still being passed along, one batch at a time.