For my own convenience I note the Wikipedia entry on moonrise and moonset, which describes something I had seen but had not observed (as Holmes remarked to Watson in “A Scandal in Bohemia”): the full moon always rises at sunset. The opposite is also true: the new moon (though you can’t see it) rises at dawn. And midway, the first quarter moon rises at midday and the last quarter moon rises in the middle of the night.

The Moon rises earlier each day, but it’s not regular: it can vary from half an hour to over an hour earlier (some of this depends on how close you are to the equator). This Astronomy magazine answer explains why.

.bt-video-container iframe,.bt-video-container object,.bt-video-container embed{position:absolute;top:0;left:0;width:100%;height:100%;margin-top:0}

Rise/Set Times with Moon Phases (YouTube) wonderfully explains moonrise and moonset and the phases of the Moon.

Earlier this week I updated Fictional Footnotes and Indexes and fixed all the broken links, including to the archives of the wonderful journal The Indexer, which has a rolling paywall but decades of past issues are openly available. I saw some pieces by Paula Clarke Bain and remembered I wanted a copy of Index, A History of the: A Bookish Adventure from Medieval Manuscripts to the Digital Age by Dennis Duncan, for which she did the index.

I emailed Book City about getting it—it’s just out in paperback—and they had a copy and put it on hold. Then a staffer followed up to say the book had been recalled because the print was faint. It looked fine to her, and to me when I saw it, so I bought it. When I got home I saw the real problem:

All through the book, “hn” has been replaced with “Ù” (a capital U with a grave accent, Unicode U+00D9). The pages about Samuel Johnson are the worst:

By affixing illustrative quotations to his definitions, JoÙson turned the dictionary into the par excellence resource for the index-scholar—the “apotheosis of index-learning,” as Robin Valenza has put it. Nice to know, then, that JoÙson was not above a little index-hunting himself.

Aside from John and Johnson, the “hn” pair doesn’t appear at lot elsewhere, but whenever “tecÙology” is mentioned things go weird.

I can’t imagine how this mistake was introduced. You have to go to a fair bit of work to replace every “hn” with “Ù” in an manuscript written in English. It’s incredible it made it through all the publishing process and was printed and distributed.

But there’s more. Bain said in Index, A History of the: Conference Adventures of Author and Indexer (The Indexer 40, no. 3):

I was told that the US edition text was “exactly the same” as the UK version, and so there was no need for an updated index. Sadly, this was not entirely the case, as the list of figures was moved from the prelims in the UK book to the back of the US book, so bumping on the page numbers of the computer and human indexes. Normally this would not matter, as an index would not refer to itself. However, mine did refer both to itself and to the computer index, and so now those page numbers were wrong in the US edition, a fact which only came to light after publication. At the time of writing, this is getting fixed for future editions, but it was a very annoying and embarrassing situation with how exposed I am in this index.

That error is fixed in this paperback edition—except in one case I saw. One of the index entries under “Bain, Paula Clarke” is “non-robotic, superior index 309–40.” Her index runs from pages 313 to 343! I think they fixed the index by subtracting 4 from all the page numbers, to accommodate the movement of the list of figures from the front to the back, but the index is after that list, so its self-reflexive page references should have stayed the same.

I won’t be taking this back to get the replacement printing. This is a keeper. Bain and Duncan must be very dismayed with W.W. Norton, but those mistakes don’t distract from quality of the work, which seems delightful and informative. I look forward to reading it.

A blurb about Going Zero, a new thriller by Anthony McCarten, caught my eye: a librarian is competing to stay hidden for thirty days from a massive surveillance system run by a huge Facebook-like company that wants an enormous contract with US security agencies and has set up a contest to show how good the system is.

It starts off well but turns into a pretty standard high-tech cyberthriller, and I’m afraid I don’t recommend it.

It was not copy edited well. For example: “Three minutes later, the night-vision images start to come in from a pursuit helicopter crossing Lake Michigan from Buffalo.” That should be Lake Ontario. Later, in a meeting, someone says something memorable, and a page or two later the line is quoted but misattributed.

The Zuckerberg-like head of the Facebook-like company reads like Roddie Ho from Mick Herron’s Slough House series, but without the irony or humour.

What especially caught my eye are lines like this:

Although never a hacker per se, he knows about digital back doors, about security patches, about anonymity and self-destruct safeguards, about how to turn lines of code such as exploit/admin/smb/ into a crucial key to unlock a trove, a library, a universe.

exploit/admin/smb is not a line of code, though it could be a file path to an executable. I think this is a mangled reference to Metasploit.

But then, while typing scp-r /path/to/local/data/—the actual command to steal/move data—a new thought strikes.

scp -r /path/to/local/data server:/remote/path/ would work as an example of how to use scp to recursively copy files from this machine to a remote one, with placeholder file paths. But you need a space before the -r switch and a location to copy the files to.

That file copy exfiltrates exabytes of data in under an hour. One exabyte is 1,000 petabytes, one petabyte is 1,000 terabytes, and one terabyte is 1,000 gigabytes. Let’s say 2 EB are copied in one hour. That is a lot of data … at about 550 TB per second. That’s just not possible. (The Internet Archive’s collection is about 100 PB, so that’s like copying the entire Internet Archive in two seconds.) Plus, this guy is copying from Utah to Manila.

Later, someone else “[wipes] the hard drives of all the megadata until, in under ten minutes” they are wiped. Maybe rm could delete exabytes of files in ten minutes, but that’s not good enough. Using shred to securely wipe (by overwriting each byte on the disk multiple times) would take days.

There isn’t a single mention of either Signal or Tor. The closest is this:

On the edge of town, where there’s still a decent signal, she fires up the phone. Two bars. Anonymizes the browser, then enters [a URL] into the bar.

“Anonymizes the browser”?

It may seem petty to cavil about technical points like this, but those “commands” should have been caught in copy editing. If you’re going to put them in, they should be right. Or use the trick Cory Doctorow learned about guns and say something about how the computer or disk or system had been “modified” to wave it all away.

Beyond all that, and worse, the denouement doesn’t ring true.