I highly recommend the book, "Speed brewing" by Mary Izett, particularly her short mead recipes.

Short meads are pretty straightforward, 1. 0.5 lb honey + 2L of warm water in a 2L bottle that can withstand pressure 2. Add ~1.25g of wine yeast (I usually use ec-1118, but just about any brewer's yeast would work) 3. optionally, add yeast nutrient and acid tannins 4. ferment for about 3 days with the cap loosely tightened so air can escape (ie. poor man's air lock). 5. depending on how dry you like it, screw the cap on ~1 day before you want to drink it. 6. When ready, open the cap slowly to depressurize 7. enjoy! The above is a basic recipe, but you can add adjuncts and have fun with it: • fruit • whiskey soaked oak chips • replace honey with agave • ginger • brew a 100mL of water with habenero 3 minutes

I'm usually a lazy brewer. Ciders are also really easy: • Buy 1 gallon of cider • toss in ~1.25g of cider/dry wine yeast • optionally add yeast nutrient and acid tannins • put on an airlock • wait 14 days • enjoy still or bottle and carbonate

ferment for about 3 days with the cap loosely tightened so air can escape (ie. poor man's air lock). I do this with an elastic band, elastic band over the metal bit that should clip and hold it shut, that way its closed properly, but if it tries to "burp" or blow the elastic band will have give

neat. what kind of stuff do you make?

I have a couple that I like, one with pineapples, cinnamon and a bit of chilli and one with pears, apricots, cinnamon and coconut. Add water and yeast, sugar, honey etc. Do you think temperature has a big effect, some times after a couple of days it's like this:

which is fine as I can filter it and pass it through cheese cloth and it comes out nice, but other times there's almost nothing happens

^^ that one filtered to

I think temperature makes a really big difference.

I tried an experiment with coconut milk, which was a disaster. A complete disaster.

i did actually try filtering it, and tried to drink some with some ice, but it was gross

I tried to start a "guess the secret ingredient" series last year right before things shutdown. I fermented some cocoa puffs before deciding to postpone the series. It wasn't undrinkable, but it definitely wasn't good.

My dad used to make elderberry wine and he found that putting the demijohn full of fermenting elderberry gloop in the airing cupboard/on top of the boiler really helped speed up the process.

Cautionary note: don't allow the airlock to become clogged up, as my dad did one time, or you'll need to re paint the ceiling.

Interresting thread. I do bake sourdough for over a year now every couple or days. My bread machine has a program to ripe the sourdough, I think it keeps it at about 27 Celsius. Something like that would be nice for consistent results I guess.

Whey soda is still one of my favourite easy fermented drinks to make, it only went wrong for me once. It does contain whey, perhaps unsurprisingly.

It seems like a compiler is an example that isn't temporally decomposed (hopefully). As long as each step is data in, data out, there's no requirement about how or when the data was produced as long as the data is valid for the current step. For example, you could skip the optimizer step before emitting code. Or rather than parsing source code, you could programmatically generate whatever AST/data is required by the optimizer.

In fact, I think moving away from the coupling between stages was one of the design goals for LLVM vs the dominant compiler at the time, gcc.

Right, thank you. I fully agree.

I agree with the way these parts can be decoupled, but the whole thing (the compiler) still has a notion of sequentiality. It cannot emit code before it was parsed.

The quoted book above has a weak discussion on what temporal decomposition means I think. Or I don’t understand how to decompose a compiler in parts that are not stages in a process.

> It cannot emit code before it was parsed I think the key difference is that it can emit code that was never parsed

Right, similarly read and eval are things we can use in Clojure independently.

and in theory, you could have these stages running concurrently where the compiler starts emitting code before the full program is fully parsed (eg. web assembly)

Another quote from the book, I had to laugh: > “Most modules have more users than developers, so it is better for the developers to suffer than the users.” > > Excerpt From: John Ousterhout. “A Philosophy of Software Design”