Stephen StrowesInterview
Network measurements researcher
Adapted from Stephen Strowes's interview on Uses This, published 10 July 2020. The kit is drawn from their answers and each reason is an excerpt, lightly edited. Shared under CC BY-SA 4.0. Read the full interview ↗
Is this you? Claim this kit →Hardware
My main workhorse machine -- for actual work -- is currently a 15" mid-2015 MacBook Pro.
I have an HP ProLiant N40L that's been chugging along since forever, with more terabytes of storage than I'll need any time soon.
In the more Googley ecosystem, I've been using a OnePlus 6 as my phone for the last 18 months, and it's extremely nice.
I have two Chromecasts: one audio, plugged right into the back of my amplifier, and one video, plugged right into the back of my TV.
For audio, my amplifier is a Cambridge Audio Azur 540A, which isn't super new but it's been with me for over a decade and, aside from soldering new switches into it a couple years ago, just keeps on going.
RHA MA650 wired earbuds, which I typically use when on travel (so currently they're gathering dust).
AKG Y500 Bluetooth on-ears which I throw in my bag between home and the office (so currently ...
Audio Technica ATH-M50x over-ears, which I can wear for hours on end.
A small Pway KVM to help reduce cable switching.
And, a cheap FIFINE USB desk mic, an easy step-up from a built-in mic.
For something more, I have a Canon EOS 80D and a handful of lenses.
I archive my old floppy disks with a KryoFlux device that reads the magnetic data from the disk itself.
Productivity
I feel like I ought to double-down on Jupyter notebooks or somesuch for more open analysis and collaboration.
For public speaking, Keynote, mainly because it's the default.
I've built Beamer slides with tikz diagrams in the past and, with enough time, they can be great.
Notes
Academic conferences push almost all of the typesetting work onto the authors, and for this LaTeX is still definitely the common tool.
Coding & Development
Certain types of work need me to throw blobs of JSON around, and jq gets me through a lot.
For drawing plots, Gnuplot is still my go-to, and the main reason it stays that way is I can rattle off Gnuplot commands without reference.
Other
If I want to burn some cycles on a few local cores, or across some locally-managed machines, GNU Parallel comes in useful.
The TeX Live distribution is gigantic but it'll build a document that looks identical regardless of the system you build it on.
Recently we've been throwing data into Google's BigQuery, so I've learned a bunch about how to work in that ecosystem.
In the past, I've run ongoing analyses on massive Hadoop clusters, and ad-hoc analyses on small clusters, and in all cases I find working with Hadoop to be way more work than I truly want.