Unisexual Ambystoma

Unisexual Ambystoma

Sunday, November 2, 2014

How We Work: Dr. Tyler Smith talks standing desks, cranberries, and coding to the Tron soundtrack


Our next guest for the How We Work series is Dr. Tyler Smith. Tyler is currently a research scientist for the Canadian government, but we first me while he was a faculty member in the department of biology at Eastern Kentucky University. 

Prepped and ready for sedge hunting in turkey season

Location: Agriculture and Agri-food Canada, Ottawa, Canada

Current Position: Research Scientist, Taxonomy of native crop wild relatives

One word that best describes how you work: Inertia would work. Reactive would be accurate too much of the time as well. I aspire to be a more mindful and deliberate in my approach. I struggle with the bad kind of perfectionism, which can be crippling when you get a manuscript 90% done. At that point it becomes irresistibly tempting to start something new, rather than complete something that won't live up to your aspirations. That's a real productivity killer.

Current mobile devices: Nexus 5

Current computer(s):
Personal Laptop: Debian GNU/Linux on a Thinkpad x201 with 4GB RAM and 320GB HD
Work Laptop: Debian GNU/Linux on a Standard Issue HP Elitebook

Both computers run the same operating system with the same suite of programs. I sync my important configuration and documents via BitBucket.

Digital Ocean Droplet with 512MB RAM, 20GB HD for my website and Owncloud storage
If you like tinkering with computers, Digital Ocean provides really good value. $5/month gets me my own virtual server. So far, other than serving a static Jekyll-generated website, I use it to host my OwnCloud instance. Owncloud is a Free Software replacement for Google Drive or Dropbox. It offers the same sort of features, but you control the whole system. Nobody is mining it for ad revenue, and I know the owner (me!) isn't in cahoots with the NSA or CSEC.

Another cool thing about Digital Ocean is that you can rent servers by the hour. So if you need to do some computationally intensive simulations and don't have access to a cluster, you can set one up for a few dollars a day, and shut it down as soon as you're done.

I expect I'll eventually install my own version control repositories, but until then I use both BitBucket and GitHub. Github is more popular, but BitBucket provides unlimited private repositories for free, so all of my manuscripts and coding projects go there.

High-performance clusters for simulation projects and next-gen sequencing projects.
My post doc was working on spatially-explicit community simulations, which introduced me to cluster computing. I'm gearing up to start working on next-generation sequencing data, so I'll be putting that experience to use as a taxonomist in future.

What apps/software/tools can't you live without?

Debian GNU/Linux
Debian provides a huge collection of Free Software, maintained by a chaotic but dedicated community of volunteers. Other than the programs I describe below, I use the i3 tiling window manager and spend a lot of time in Bash shells.

Emacs
It's not just a text editor, it's a lifestyle choice. The learning curve is steep, but in return you get an infinitely customizable workhorse.

R
It has its rough spots, but the enormous user/developer community means there's very little analysis I need to do that I can't either find a package for, or write my own scripts/packages to accomplish. Of course, my preferred interface is Emacs, via the ESS (Emacs Speaks Statistics) add-on. The knitr R package provides almost seamless integration of data analysis and manuscript preparation.

LaTeX
Writing manuscripts in LaTeX means I get to use the same tools I apply to coding (Emacs, version control with git or mercurial) to my manuscripts. And the integration of code and text in the same document with knitr is a game-changing approach to organizing manuscripts.

That said, Emacs and LaTeX are steep hills to climb. A more approachable version of the same workflow is using RStudio and Markdown. When I teach R workshops, this is the combination I recommend.

Distributed Version Control
git is the most popular option, but I find mercurial a little simpler to use. Either one provides a very powerful way to track changes in multi-file projects, share code with others, and sync files between computers. Having just finished collaborating on a manuscript with 'track changes', I wish all my peers were using one of these systems. 

Preparing a manuscript with Emacs, R, knitr, Inkscape and a PDF reader

What is your best time-saving shortcut/life hack?
For biologists, make your research Reproducible Research
"Reproducible research involves the careful, annotated preservation of data, analysis code, and associated files, such that statistical procedures, output, and published results can be directly and fully replicated" (from ropensci.org)

This can be a bit overwhelming, as there are a whole ecosystem of tools available, including text editors, programming languages, development tools and version control systems. But the pay-off is immense when it comes to sharing, revising and extending your work.

Other resources:
My brief overview
Software Carpentry
Roger Peng

How do you organize all the stuff you have to do?
I maintain my project outlines and to-do lists with the Emacs extension org mode. The data is stored as human-readable plain text files, but with lots of handy features for sorting and prioritizing tasks, tracking progress and scheduling. I also use it to clock my work on different tasks (inspired by tweets from Rob), to see if that helps keep my effort more focused and deliberate. So far I like it, although it takes a bit of concerted effort to stick with the habit.

Tyler's homemade standing desk
Besides your phone and computer, what gadget can't you live without and why?
In the field, my GPS. An important part of field trip prep is loading it up with all the maps and herbarium data I need ready access to.

In the office, I love my standing desk. I've had lots of neck, back and arm issues, and working upright is a big relief.

What do you listen to while you work?
Repetitive, atmospheric stuff. Nothing too melodic or lyrical, it needs to settle into the background. Other than that, I'm pretty eclectic: Persian sufi music, baroque, Arcade Fire. My current geeky pleasure for coding is the Tron soundtrack.

What are you currently reading?
So good they can't ignore you. This is a very contrary approach to building a career. The author argues strongly against the notion that you should follow your passion. His antidote is to build up a set of rare and valuable skills, and use them to carve out a rewarding niche for yourself. It's an interesting read, especially for a scientists considering a career outside academia.

Are you more of an introvert or an extrovert?
Introvert

What is your sleep routine like?
Horrible. I don't often get a restful sleep. I'm quite thankful I work in a job where I can set my own hours. Life was simpler when I was a professor -- I was so physically exhausted by the workload that I usually collapsed into sleep at the end of the day!

Tetraploid cranberry blooms

Fill in the blank: I'd love to see ___ answer these same questions.
Andrew Hipp (Morton Arboretum), Sally Otto (University of British Columbia)

What's the best advice you've ever received?
When you feel like you've run up against a brick wall, pause to consider if getting through it is really what you need to do. I get pretty attached to my ideas sometimes, and it can be hard to let go of the bad ones.

Also, not so much advice, but my wife is often responsible for getting me outside in my down time for civilian nature hikes. It's easy to get caught up in the abstract intellectual side of science. But just experiencing nature, unburdened by the need to find a target population and acquire more samples, that's an important way to reinvigorate my work and myself.



Thanks Tyler!

No comments:

Post a Comment