Baby basil

Building on my success last year of managing to not kill a basil seedling, this summer I’ll try having a garden. To that end, I started in March with basil, rosemary, peppers, and tomato seeds. So far, some 24 plants, mostly tomatoes and peppers, have survived to the hardening-off stage and are ready to go in the ground in a couple weeks’ time.

On top of that, I have a few basil plants that will remain indoors, and a couple of very poor looking rosemary sprouts that may never reach adulthood. This represents quite a few levels up in my plant-tending acumen, as until now the only things I have been good at growing are dandelions in my lawn.



The American Crossword Puzzle Tournament starts tomorrow. I’ve paid the entrance fee for the online “bragging rights only” track, so I’m looking forward to seeing some interesting puzzles and to get an idea of how my solving skills stack up. Considering that competitive solvers often finish well under five minutes on a Monday and my best is in the ten minute range, probably not that great, but we shall see.

I recently started a couple of the London Times cryptic crossword books; these are hard. As in NYT Saturdays are a cakewalk by comparison. Partly, because I’m not terribly knowledgeable of British geography and slang, partly because I haven’t learned to recognize the various clue forms, and partly because they are just tough. Example clue: [Notice boy going round vehicle after parking]. Answer: PLACARD. Because “PLACARD” is a notice, and in that word “LAD” (boy) is literally going around “CAR” (vehicle), which both come after “P” which you just have to know is an abbreviation of parking. So yeah, tricky.


Over the last few weeks I’ve been caring for a few billion microorganisms. And ate some of them.

Fig 1. My yeast making tons of bubbles overnight

Making a wild yeast starter turned out to be pretty easy. It took about ten days (in contrast to the five specified in the recipe I followed) for the colony to reach maturity in this winter kitchen, fed twice each day with equal parts flour and water, left at room temp overnight. Now, they are on a weekly maintenance schedule.

So far, I’ve used the starter in a hybrid (part commercial yeast) sourdough boule, two fully wild yeast loaves, pancakes, and, of course, pizza.

Compared to straight pizza dough, the sourdough version has a more complex taste, with basically the same texture. There is not much difference in the way the dough handles.

The wild yeast breads were both a little flatter than I’d have liked, owing to the less predictable rise; on the other hand they had a very nice sour taste. The hybrid dough is a nice compromise: less sour but airy crumb.

Pancakes of any stripe are always awesome.

State of the pizza

It has been a year since my last statement about home-made pizzas, so it may be a time to serve up (deliver?) an update. Here’s one from the most recent batch:

Ken Forkish’s straight-dough, low yeast, long-rise recipe continues to be my go-to, and there is little left to complain about. It has a great crumb and nicely developed taste, especially when the dough is a day or two old. I still do need some work on the technique side: getting an unbaked pizza off the peel — without using so much flour or cornmeal as to mar the taste — still eludes me. Thus, I have settled on using parchment paper for the first couple of minutes, then, after the crust firms up, lifting the pizza with tongs and yanking out the parchment paper before it incinerates in the 550 degree oven. This, it should be said, is a maneuver not without danger to one’s fingers.

Our grocery carries “pizza yeast” in addition to the normal stuff, so I decided to give this a try to see how it stacks up. Apparently, the packet consists of highly active yeast along with some conditioners that make a thirty minute dough handle like one with a longer rise. The resulting dough was as easily stretchable as advertised, but also had little to no strength. It was altogether too easy to tear a hole while trying to shape it into a round, but I persevered. Here is the side-by-side (pizza yeast on the right):

The resulting crust was rubbery and tasteless, so my advice is to not bother repeating this experiment.

I am considering branching out into a wild-yeast version next, depending on whether I can manage to get a viable starter going.

React reaction

An embarrassing admission to make: I wrote something in javascript.[1]

Actually, I wrote it in React, because I still think of javascript as that awful language that works differently in every browser, and so, you know, to stay relevant, I decided to learn the framework that web developers everywhere have already cast into yesterday’s wastebin for the new shiny. This, pretty much.[2]

Anyway, I replaced the third-party crossword app on my website with my own NIH version and made it somewhat responsive [3] so that it works on my phone. Also there are two new-ish puzzles since the last time I posted, dated 2016-11-25 and 2016-12-30. Each of them has its own set of problems, but I think I am at least learning where I need to improve.

To this neophyte, my impression is that using React (and ES6) is much nicer than that terrible language that they compile down to. On the other hand, the pattern of moving state up the object hierarchy away from an object’s view feels backwards and not very OO-like. Perhaps if I also used Flux then I might get enlightened on this point, but that is for another day.

[1] Here’s a nickel, kid. Buy yourself a real language.
[2] The day that my relevance depends on knowing the latest JS framework is the day that I will take up a completely different career.
[3] Anything worth doing is worth doing somewhat.

On placing letters

Work continues on my fork of the QXW crossword filler, which by now is about 50% rewritten by me and ultimately will end up looking nothing like its predecessor. Already it is a command-line, monte carlo automatic filler instead of a user-directed GTK program. Both have their uses, I think.

In a previous post, I wished that QXW supported scored dictionaries. It turns out that it does, though in my tests it gave unexpected results in many cases. I’ve reworked most of that code and am pretty happy with the results I get now. Just to cherry-pick a before-and-after example:







The latter grid looks a lot more like actual words, though it does still have some garbage entries.

QXW’s algorithm could be roughly stated as “find the hardest to fill cell, then fill it with the highest scoring letter, repeat.” It is a naturally recursive algorithm (implemented iteratively with stacks) that either terminates when the grid is filled, or unwinds and tries the next best letter.

I have some issues with this approach, the main one being that maximizing letter score seems unlikely to maximize score of whole words. I’ve been thinking about this problem some, including the form of optimality that I would like to achieve: maximize the sum of the scores of every word.

A simple brute force algorithm implementing this could be “fill each entry with a word, compute the score, and repeat until all words in all entries have been tried, taking the grid with maximum score.” It’s fairly easy to realize this exponential algorithm though you may be waiting a looooooooooong time for it to finish.

What I have now is something like this greedy algorithm:

  1. Base case (no more entries):
    1. if grid is filled, return the grid
    2. return None
  2. Choose the longest unfilled entry crossing a “critical” cell
  3. Fill with the next highest ranked word for that entry
  4. Recurse. Repeat step 2 until grid is filled or no more words

In practice, with a large enough dictionary, this usually terminates without searching forever but you get some junk like PXA. I have added an arbitrary upper limit at which point I just give up the search.

An important thing to note is that the problem does not exhibit optimal substructure, so dynamic programming is not applicable. For example, take this template:


We can fill it any number of ways. My algorithm will fill it like this:


…which is terrible (HPA — [Millibar alt.?]). Much better would be:


…which is what you get if you fill the child “EPA” first.

There are a couple of things at work here: first, I pick the longest entry so .IP and .PA don’t even get a chance. Does this make sense? Should I instead pick shortest entry? Entry with fewest completions? Any of these might make better puzzles or at least move the junk around.

Second, THE is really really really common. The commonest three letter word ever. The previous sentence had one and so does this one. That’s a lot of THEs. Also it is a terrible word for crosswords. [Well, looking back through my NYT database, it has been used over a hundred times: {It's definite} or {Genuine article?} being decent clues.] Unfortunately, THE is so very common and so highly ranked that practically every high scoring puzzle will have one. So, my optimality metric may be sub-optimal in that respect. Flattening the word probability curve somewhat could partially address this issue.

It is clear that we could pick worse words at some levels to get a better overall result. So I implemented a very simple monte carlo search that looks like this:

  1. Run the greedy algorithm, saving decision stack
  2. Pick a random location in the stack, and from there, pick a random word (instead of best word). Re-start algorithm from there.
  3. Repeat this zillions of times, keeping track of best N grids as seeds for future iterations

Lots of variations on the above remain to be explored. Also, this approach opens the door for taking a completed grid, manufacturing a search tree that generated it, and then optimizing it from there. Don’t like the fill in a published puzzle? Let the computer regenerate it for you.

Encrypted mesh PSA

I’m a bit late writing this up, as lately wifi stuff has taken a back seat to the day job. But, I’ve now seen the following issue reported more than once, so hopefully this post will save the other two mesh users some grief.

If you are running an encrypted mesh on wpa_supplicant or (ye gods) authsae, take note:

  • Encrypted mesh on kernel 4.8 will not interoperate with kernel < 4.8
  • wpa_supplicant 2.6 will not interoperate with 2.5
  • authsae as of commit 706a2cf will not interoperate with that of commit 813ec0e [1]
  • wpa_supplicant 2.6 and authsae commit 706a2cf require additional configuration to work with kernel < 4.8 (see below).

What is behind all of that? A while ago I noticed and mentioned to another networking developer that the 802.11 standard calls for certain group-addressed management frames in mesh networks to be encrypted with the mesh group key (MGTK). However, the implementation in the kernel only integrity-protected these frames, using the integrity group key (IGTK) [2]. I don’t happen to know the history here: it may be that this was something that was changed between the draft and the standard, but anyway that was the state until Masashi Honma fixed it, with patch 46f6b06050b that landed in kernel 4.8.

Meanwhile, Jouni Malinen fixed a bunch of related issues in the wpa_supplicant implementation, namely that IGTK was not being generated properly and instead the MGTK was used for that. I made the same kind of fixes for authsae.

Yes, it’s a multiple(!) flag day and we suck for that.

Now, on to how to fix your mesh.

Old mesh daemon + old kernel everywhere should still work.
Old mesh daemon + new kernel everywhere should still work.
New mesh daemon + new kernel everywhere should still work.

New mesh daemon + pre-4.8 kernel: you will most likely see that peering works but no data goes across because HWMP frames are dropped at the peer. The solution here is to enable PMF in the mesh daemon:

  • add “pmf=2″ or “ieee80211w=2″ to the relevant section of wpa_supplicant.conf
  • add “pmf=1;” to the meshd section of authsae.cfg

If you for some reason find yourself in the position of needing to run a mix of old/new wpa_s and/or old/new kernel, here are my (completely untested) suggestions:

  • patch the kernel to accept PMF-protected frames as well as encrypted frames (this allows older kernels to work with newer ones)
  • patch the new mesh daemon (wpa_supplicant or authsae) to copy the MGTK into the IGTK instead of generating a separate key (this allows older mesh daemon to work with the newer one since older daemon will use MGTK for IGTK).

Both of these will reduce your security and don’t follow the spec, so there’s little chance such changes will go upstream.

[1] Prior to 706a2cf, authsae doesn’t even interoperate with wpa_supplicant due to not filling in all of the elements in peering frames. My advice would be to switch to wpa_supplicant, which has things like a maintainer and a test suite.
[2] On encryption vs integrity protection: the latter is similar to PGP-signing of emails: it makes tampering evident but does not hide the content of the message itself.

More boring crossword stuff

Since last post, I made two more puzzles, both standard newspaper size, and I appropriated and modified someone else’s javascript so that you can fill it in on that page. Both puzzles went through two drafts where I completely redid the fill after Angeline test-solved them and found some problems.

I’m still using Qxw, now with a 130k+ entry dictionary I built from 10 years of NYT puzzles, which works OK. Still, I feel like it could help even more: for example, it could rank words by their commonality in the corpus so that you don’t use something hyper-obscure that happened to be in one Saturday puzzle when the editor was asleep; or, it could give partial fills even though it may not be able to complete the whole thing; or, it could give hints on places to place black squares to improve chances that a reasonable fill exists.

To get at the latter, I extracted the qxw filler into a command-line program that just fills templates passed to it on stdin, and, into that, piped a python script that generates random valid grids. This helped for the most recent puzzle where a few spots had only two or three words that would fit given the themers, and moving a couple of blocks around was the difference between no-fills and some-fills. A quick improvement there would be to score the resulting grids, and use a genetic algorithm to produce new templates based on their fitness.

Anyway, these various ugly hacks are now on my github.

Another puzzle

As I wrote before, I recently tried my hand with crossword construction.

This new puzzle is my sophomore effort. This one, slightly bigger at 13×13, came together much more easily for some reason (luck?). This time I did not have resort to using any unknown words, and nothing is really obscure, with the possible exception of 17-across (a common crossword answer).

As before, I mostly let the theme clues dictate the layout, while I left in a few longer stretches this time. The top left and bottom right corners were sufficiently isolated that I did them by hand before bed, then the next day I filled the longer answers and finished the bottom within about 30 minutes. At this point 3/4 of the puzzle was filled manually. Then, Qxw woke up and was able to help fill the rest (including 18-across, a nice gift which I happily accepted) — still using the standard Linux dictionary file. I reworked some of its suggestions using some of my own, but it was quite a time-saver.

In all I’m happier with this puzzle than the previous one: there’s more wordplay, better theme, nothing too difficult (I think). Comments welcome.

I guess this means the next one will be full-size.


I will admit to being an insufferable word nerd: the kind of person who will intentionally make a pun and then ruin it by acknowledgment. My wife is the sufferable sort who, presented with the same joke, groans outwardly and inwardly at the same time. Together, we have a shared appreciation for the crossword puzzle, and can be found on the porch filling out a grid on the rare occasion that a lazy summer afternoon presents itself. We can usually get through a Thursday or Sunday NYT puzzle, but we aren’t going to be entering any speed solving contests.

A few months ago, I heard about Qxw, a crossword construction software for Linux. As a puzzle solver, the art of creating the puzzle seemed mysterious and left to those far smarter than myself, but any old dumb person can run some software. I finally tried it out yesterday, and here’s the result.

This is an 11×11 puzzle; newspaper puzzles are usually 15×15 but this was my first try so wanted to keep it easy-ish and constructable in a few hours. For the uninitiated, there are a few rules that help guide one into picking a layout: generally there are a few (~3) longish answers that form a theme; the layout is diagonally symmetric; the smallest answers are 3 characters. Taken together, this means two of the theme answers are usually 3-4 rows from the top and bottom, of the same length and in the same relative location, while the third theme answer is in the middle. Additional black boxes are sprinkled about to break up extra-long words.

It turns out that Qxw didn’t actually help much besides automatically placing the mirrored black squares: once I filled in my theme answers and blacked off part of the grid, Qxw couldn’t fill any other spaces. I suspect expanding the dictionary beyond /usr/share/dict/words would help a lot in this respect. Instead I occasionally used egrep on the dictionary file with suitable regexes to find candidate words and resorted to some abbreviations and prefixes. I did cheat a few times by googling some combination of letters that I hoped would be a thing (1-down, 6-down, 35-down). I need to improve my 3-letter word vocab.

Things I did do that annoy me as a solver:

  • Fairly obscure names or characters (35-down, at least to me, a non-Trekkie)
  • Use of old stand-bys (10-, 14-across, I burned these quickly!)
  • Weak multi-word answers (9-down)
  • Foreign words (1-down)

Despite the flaws, Angeline solved it in about 10 minutes while watching TV; she also supplied a much better clue for 25-across which I have appropriated.

I’d say this effort gets a B-, but it was an interesting exercise to do the puzzle from the other side. 13×13 next.