Philippa Cowderoy @flippac@types.pl

Thing I'll add: while a whiteboard will reflect sound somewhat, a shotgun mic still has the "talking in the wrong direction" problem in exactly the way a lavalier mic or adequately-used vocal mic doesn't - the lav sits on you and comes with, the vocal mostly just needs to stay about the same distance in front of your mouth.

Less of a problem for slides, but you're still going to point, right?

(do remember to take off the lav before going to the lav, mind)

Thinking things about hybrid and/or recorded-but-intended-as-local talks and mic setups.

I know some people who really do need both hands free or at least a marker in one hand and another to point with.

I definitely know people who won't be used to/comfortable with holding a mic in their off-hand while talking, especially while using the other hand to write and/or manage slides.

Managing to point with the mic hand while talking into it definitely requires a certain level of coordination! Me and my dyspraxia might have got away with dual-wielding a mic and a stick to point with, but my dyspraxia is tempered by having done some martial arts including just enough with a pole to make it look easy.

Sometimes I can't help wondering if universities should've given groups a bigger AV budget during the pandemic though. Omnidirectional mics "to help everyone participate" end up doing bad things to the mix by default, and you can do a lot with more than one mic: for starters you can make the speaker sound a lot clearer!

At the same time, I know why I know a damn thing about audio engineering and it's pretty difficult to ask for budget when none of you really know yet...

Thought experiment: you've got an interface with four inputs (let's assume XLR by default) and money for enough low-to-mid-end prosumer/professional mics to populate it plus any cabling you need. You don't have to use all of the mics in a given talk: what would you buy and why? How does that change if you're allowed acceptable wireless interfaces for them?

(things I'd seriously consider: one each of lavalier+wireless interface, shotgun which someone's probably going to have to keep pointed at the speaker, one standardish vocal mic, then some thinking about one for room participation. Speaker gets first pick of what's handling them with lavalier and shotgun as the "just make it work" options, organiser gets whichever of vocal or lavalier's spare...

...and then I start figuring out how to get a decent signal chain in front of Zoom, but a basic software mixer will do in a pinch)

Does "the devil's arse mints" (or "devil's ass mints" as appropriate) mean a specific drug to you?

I really need to figure out better storage for buttons (arcade, 30mm and 24mm of varying height) and microswitches (whether arcade button or keyboard).

Interviewer: Can you explain these presences in your resume?

Me: I might not have worked a job in my life but that doesn't mean I've done nothing

New stick, old mods

Nice to have the headset jack. In that position the touchpad is well worth unbinding if you're playing at all seriously.

(It's a Qanba Drone 2 - the stock buttons aren't bad, the lever's awful but also pretty easy to replace with a couple of screwdrivers and a better one to hand)

Today I implemented overcommitment. Then I toned it down a bit and had the feature I actually wanted.

Well, the one I'll put up with for now, anyway. The layer I'm working on's still a compilation target.

I'd like prompts/named choice points and commitment to a named one instead of the most recent though. And not just because I've been meaning to write a small parsing lib with that as the major change!

Technically it's missing a minor-variant case that I'll sort out tomorrow, but I have successfully implemented semi-unification in terms of stuff that's supposed to be directly in my source language!

Need a few more bits and bobs, but full (first-order) unification to follow at some point.

Just don't ask about all the failure-to-unchain bugs?...

I'm still having to ask for fresh variables explicitly, but I'm writing code that smells a lot like Prolog-without-unification and so far I'm still doing it directly in Haskell without a notion of rule...

Got a ways to go, but it's sure coming along

reminds self that having to rebuild Hakyll and dependencies does not make it okay to go fix a minor thing in the blog source while it's building

TFW you've gone mechanical-keyboard-nut enough that one of your keyboards now has two different types of switch in it to make sure you learn the minor oddities of its layout vs your current laptop and the keyboard you've been using since the 80s...