Apr 23 4 2023


A fun place in Washington’s refurbished market space “La Cosecha”, which has been refashioned into a sort of upscale food court or, rather, a set of small restaurants that share a big room (and, I presume, some back-of-the-house facilities for greater efficiency). A very small but well-considered menu with some of the best Mexico City plates I’ve seen. Duck breast with avocado cream and heirloom frijoles blancos. Braised short ribs with leeks. Good cocktails, too.

The place is near Gaulladet, and several tables near us were engaged in energetic ASL discussions. So local color, too.

Apr 23 3 2023

Into The Woods

We saw Into The Woods at the Kennedy Center last week, and it was a terrific show.

I heard the original cast album quite a bit in the early days of Eastgate, but I’d never really understood how Into The Woods is a meditation on being a Jew in a world with an empty sky.

I wish my son were not a fool
I wish my house was not a mess
I wish the cow was full of milk
I wish the walls were full of gold—
I wish a lot of things...

It starts with folk wisdom in a characteristic slant vein: “And though scary is exciting/Nice is different than good.” Everyone, pulling together, wins through.

But then the giants show up and wreck everything: the house, the palace, the village, all gone. And we ourselves are not are selfish, short-sighted, greedy, vengeful, careless — and it’s all your fault. Still:

Witches can be right, Giants can be good.
You decide what's right. You decide what's good.

In the end, standing in the wreckage:

Hard to see the light now
Just don't let it go
Things will come out right now
We can make it so.
Mar 23 15 2023

Ocean State

by Stewart O’Nan

“When I was in eighth grade my sister helped kill another girl.” That’s quite a first sentence. Our sister is Angel, and the girl she eventually helps to murder was Birdy. A and B. This is a schematic novel, one with interesting ways to explore shifting points of view. It’s got a nifty sense of place, too, and makes a fun pair with Ready, Set, Oh!, Diane Josefowicz’s fine and spooky 2022 novel.

Hi! I’m Jessica Brown, from Miskatonic University (MU). I use Tinderbox every day to allocate time on the various time machines on campus. I don’t know how we could manage without it!

Remember: you know who you are, but the software developer doesn’t. Your email address might be sufficient, but often it's not. It also helps to give some sense of what you’re doing, and how familiar you are with the product.

It is also a good idea to communicate a sense of your purpose here. Are you proposing something that might be intellectually interesting, or something you need next Thursday? Is it something you need just once in order to finish your dissertation, or is it something a hundred people will use every day?

We currently have no fewer than 17 time machines, but each of them has its own constraints. The machine in the basement of Severance Hall, for example, is only available during the day, while the machine in Performing Arts is available 24/7 but reserved on weekend evenings by the Theater Department. The machine in Somerville was built for the Women’s College and is not very comfortable for people taller than 175cm. The machine on the roof of Holmes Hall must be primed with burning tobacco, which can cause problems with allergies. When a professor wants to reserve a machine, they fill out a web form ( This sends me an email which Tinderbox parses into a note with the request. An agent, which is linked to that note, then finds time machines that might be suitable.

People are often tempted to abstract away all the detail of their work, or to assume those details are self-evident. In my experience, it can help a lot to have some perspective and some sense of the details.

The agent works really well! I’ve attached a simplified version, one with just three time machines and a single reservation request, so you can see the details. I know this is a very simple-minded use of Tinderbox!

Everyone says that they’re a novice, that they’ve scarcely scratched the surface. There are many different ways of being a novice. One person is “only using the easy parts” but knows them thoroughly. Another is a professional developer, but only started with Tinderbox yesterday. Having a small example lets the developer see what you’re actually doing.

The agent is very good at saying, for example, that Professor Higgins could use Severance on Tuesday or Holmes on Sunday night. My problem is that Professor Higgins inevitably calls me to ask why he cannot use the machine in University Hall instead. Can an agent — or something — tell me what part of the request rules out that particular machine? Is the Attribute Browser pertinent?

Clearly identify your problem. This is perhaps the very best use of bold fonts.

Context helps. Here, we understand that the information for which we are looking is likely to be requested by phone, at an arbitrary time.

As it is, I have to fiddle with copies of the agent, changing queries until I find the issue. This is slow and prone to error, and mistakes make people suspect that we are arbitrary or unfair in assigning time machines. In the sample file, you’ll see that University Hall is not suitable to Jane Doe’s project; how could one display why each unsuitable machines was excluded? (In this case, the answer is that this machine, when not in use by the administration, is reserved for digital humanities research and Professor Doe is in Physiology.)

A small sample file lets the developer focus on your problem without wading through extraneous detail.

Have other ideas about feature requests? Email me.