My 7″ 1280x800 monitor got a bit flaky during the data center move, and while it survived, the screen quality will never be quite the same, so I picked up an Eyoyo 10″ 1920x1200. The blindingly bright blue power LED needed covering with gaffer tape, but the image quality is excellent, and it even has speakers. Works great with my Mac, but even though my Surface Pro 2 detects it and shows its resolution, for some reason it doesn’t send a video signal out to it via HDMI. I tried two different mini-Displayport adapters and a brand new high-end cable, so I suspect HDCP handshake issues.
[Update: there’s something a bit odd about the Eyoyo monitor; when I first plugged it into my new Raspberry Pi 3, it worked perfectly, including showing the splash screen, but ever since, it only works if I force-select a video mode. Still works fine with the Mac, but it simply cannot auto-sync with the Pi, or show anything from the Surface Pro 2, which makes it “less than versatile”]
While setting up my cube in the new building (which included swapping the desk pieces around and adding a partial roof), I decided to ditch the crufty old USB2 hub I was using in favor of an Anker 10-port USB3 with fast charging support. Mostly because the right-hand USB port on my MacBook Pro went out a while back. Maybe after the holidays I’ll send it in for service; I think the AppleCare runs until February.
Those were easily justified purchases, but in the nearly-pure toy
department, I picked up a TrueRNG3 USB
hardware random number generator. Automatically supported in Linux,
works on a Mac if you manually send the output to /dev/random, and
they have a Windows driver. Sadly, while OpenBSD 6.2
supports several USB HRNG, it looks
like the TrueRNG isn’t one of them. Yet.
I don’t have dieharder test output from it at the moment, because it only delivers 50 KB/s of random bits, but ent really likes its output. The vendor claims to pass the dieharder tests, but doesn’t say how long it took to generate enough samples (some of the tests consume a lot of bytes).
Why am I trying to improve the quality of my random numbers? Because I’ve been playing with random passphrase generation again, using a rule-based generator that lets me feed in a variety of patterns and dictionaries, for Diceware-style generation with calculated entropy. For instance, one script generates an improved romanized Japanese diceware list, based on the word-frequency rankings in JMdict, and I can do side-by-side comparisons between different lists to see how easy the results are to remember. Side note: the EFF wordlists are a nice improvement on the original English 5-dice lists, and they’ve got two nice 4-dice lists as well, with interesting properties.
For real fun, I set my generator up to allow mixing multiple rulesets, which is a relatively minor increase in entropy, but definitely more entertaining (especially if you include the Lord Of The Rings ruleset…).
And it also let me prototype an updated syllable-based system using polyhedral dice. Cut out and save!
+--------------------------------------+--------------------------------------+
| Random Passphrase Generator | Digits, Specials, Uppercase |
| | |
| 1 2 3 4 5 6 7 8 9 10 11 12 | 1 2 3 4 5 6 7 8 9 0 |
| ----------------------------------- | @ # % $ ! | & _ , : |
| s v p h f k r t j m w b | A B C K N Q T X Y Z |
| | + - * / |
| 1 2 3 4 5 6 7 8 | |
| +-------------------------------- | d12 * d8 * d6 = 9.17 bits |
| 1 | eth id ol eg od az it el | var/op/digit = 8.64 bits |
| 2 | ug og an as on ar is al | d10 = 3.32 bits |
| 3 | ath ay ot iz us os a e | d4 = 2 bits |
| 4 | et il u ed ig uz oth o | |
| 5 | oy ul ag en un oz or es | 10 syl ~ 14-char random ASCII |
| 6 | ad ez ith ut i ud in at | 12 syl > 16-char random ASCII |
+--------------------------------------+--------------------------------------+
(this table shouldn’t wrap on most screens, but for some reason it does in Chromium on my shiny new Raspberry Pi at 1920x1200, unless I zoom the font size to 80%, 110%, or 200%; some odd scaling in the version of Bootstrap I’m using, I guess)
Most people never find the old Diceware tables for generating syllables, special characters, etc. They’re buried in the original mailing-list post for Diceware, and aren’t referenced in the current HTML pages. Most of them aren’t particularly useful, but if your password policies require some combination of upper-case, special characters, and digits, they’re better than just appending “A1!” to every password. My var/op/digit rules above exist for the same reason; I find it pretty easy to remember “Z/4” or “X+2”.
Markdown formatting and simple HTML accepted.
Sometimes you have to double-click to enter text in the form (interaction between Isso and Bootstrap?). Tab is more reliable.