HN user

tomkinstinch

2,465 karma

[ my public key: https://keybase.io/tomkinsc; my proof: https://keybase.io/tomkinsc/sigs/RwgGjNSMa0i-voAs_CTTYphBfTO4YpIa72xcrDm_nek ]

Chris Tomkins-Tinch, Ph.D.

--

I do bioinformatics, full stack dev, instrument electronics/embedded systems/firmware, simple optics, CAD/CAM/rapid prototyping, some wet bench biotech, and technical photography.

open to interesting side projects

e-mail: tomkinsc@[google's consumer email service].com

linkedin: https://www.linkedin.com/in/christomkinstinch/

Posts37
Comments554
View on HN
bdi-pathogens.shinyapps.io 6y ago

Quantifying the importance/timing of different routes of SARS-COV-2 transmission

tomkinstinch
1pts0
myriscope.com 6y ago

Myriscope for scanning (bio|a)rxiv articles

tomkinstinch
2pts0
www.topic.com 7y ago

There's no such thing as a free watch (2017)

tomkinstinch
162pts88
up.csail.mit.edu 7y ago

Cook My Meat: heat diffusion through meat over time

tomkinstinch
226pts66
www.jennyodell.com 7y ago

There's no such thing as a free watch [pdf]

tomkinstinch
2pts0
up.csail.mit.edu 7y ago

Cook My Meat: heat diffusion through meat over time

tomkinstinch
2pts0
www.washingtonpost.com 9y ago

America’s ‘Miracle Machine’ is in desperate need of, well, a miracle

tomkinstinch
9pts0
github.com 9y ago

Whereami – Using WiFi signals and machine learning to predict where you are

tomkinstinch
127pts19
www.nytimes.com 10y ago

How Much Warmer Was Your City in 2015?

tomkinstinch
1pts0
www.eater.com 10y ago

How ChefSteps and Gabe Newell Plan to Transform Sous Vide Cooking

tomkinstinch
4pts0
www.pbs.org 10y ago

The new tool colleges are using in admissions decisions: big data

tomkinstinch
1pts0
github.com 10y ago

github-wide CSS style

tomkinstinch
1pts0
itunes.apple.com 10y ago

Show HN: MadBlocker — iOS 8 Ad and Tracking Script Blocker

tomkinstinch
44pts49
flowcharts.llnl.gov 10y ago

Charting the Complex Relationships among Energy, Water, and Carbon

tomkinstinch
1pts0
www.patreon.com 11y ago

Astronaut Chris Hadfield Creating Science Videos: “It's Not Rocket Science”

tomkinstinch
1pts0
www.macrumors.com 11y ago

Apple Introduces Revamped Two-Factor Auth for iOS 9 and OS X El Capitan

tomkinstinch
4pts1
rain.works 11y ago

Invisible art activated by rain

tomkinstinch
3pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
8pts9
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
news.ycombinator.com 11y ago

Ask HN: Where do you turn for good data on market salaries?

tomkinstinch
0pts0
en.m.wikipedia.org 11y ago

AMPRNet: computer networks over amateur radio

tomkinstinch
2pts1
thermal.com 11y ago

Seek: A thermal camera for your smartphone

tomkinstinch
4pts0
artfulmac.com 11y ago

Show HN: Artful for Mac

tomkinstinch
229pts145

Looking forward to the day when FreeCAD is a viable and stable option for free parametric CAD. There are a few free options for direct modeling, but not for parametric design.

As far as commercial software goes, my current favorite CAD software for hobby use is Rhino[1]. It's not parametric[2], but it's stable, fast[3], can import and export a wide variety of 3D file types, and it's pay-once-per-major-release. It's not cloud-based. The marketing around it seems to emphasize design/architecture/artistic use cases, but it also works well for dimensionally-accurate mechanical parts.

For those eligible for a student license, the pricing is reasonable (cheaper still if you shop around among third-party edu software vendors). Surprisingly, the student license also allows commercial use.

1. https://www.rhino3d.com

2. Well, Rhino is not parametric in the usual sketch-based way. People do wild things with the Grasshopper plugin.

3. Rhino also runs on macOS, w/ hardware acceleration of graphics via Metal

The Apple Store sells both and will let you compare the two models if you ask nicely (with new/clean AirPods). I found AirPods have noticeably better noise cancellation (for me) than the QC35s, but YMMV. Subtle but better. It's hard to say if my impression was due to better passive noise isolation or improved active noise cancellation (does the QC35 have an error-correction microphone downstream?). One thing that is absolutely true is that the QC35s do not work as well mechanically for those of us who wear eyeglasses: the cushions can't seal around the temple pieces of the frames, so some sound will always leak under the eyeglasses. For that reason alone—even if the AirPods have equivalent noise cancellation tech—the AirPods can work better for people with glasses as long as the tips seal properly. My wife, an eyeglass wearer with long hair, also found the AirPods more effective at reducing noise when we tried them both at the Apple Store. I certainly find them effective for commuter rail travel.

As an aside, one slightly annoying usability issue with the AirPods is that in comparison to the QC35 they behave differently when part of a multi-output audio device under MacOS. My wife and I often watch video using two pairs of bluetooth headphones paired and synced to the same Macbook Pro[1,2]. It's great for late-night movies in city apartments. Using the QC35s and her headphones, the aggregate device remains selected when the QC35s are paused or turned off (meaning, she can still listen if I go out of range or turn off my headphones). With the AirPods, if I take one out of my ear (to scratch an itch, say), the AirPods suddenly become the only audio output device when I put it back in, and the second pair of headphones goes silent. A small annoyance, but curious.

1. Set up your own bluetooth audio "splitter" to allow multiple people to listen togther: https://support.apple.com/guide/audio-midi-setup/play-audio-...

2. https://support.apple.com/en-us/HT202000

Indeed; a good library is a long-tail institution. The Harvard Library system has a large collection of books at the various on-campus libraries, but an immense catalog of infrequently accessed books at a "depository" warehouse off-site. Scholars can request items from the Depository and receive them within a day.

This split-storage model may a sensible solution for other universities going forward: prime and expensive library buildings on-campus can be reserved for quiet study areas and a few commonly-used books, with the main collection of books retrievable from a nearby location where land is less expensive. It also has the advantage of allowing books to be stored in the ideal climate for their preservation.

Fun fact about the Harvard Depository: books are organized not by topic, publication date, or anything resembling the Dewey decimal system, but rather by a metric that makes sense for high-density: physical size. Books are stored in barcoded boxes by height. Here is an artsy documentary, "Cold Storage," about the Depository:

https://vimeo.com/116603551

or in interactive form:

http://librarybeyondthebook.org/cold_storage/

Helpful tip, thanks. Here's the Safety Data Sheet (SDS) in case anyone is curious what's in 3M 38983 General Purpose Adhesive Remover[1]: Mostly methyl acetate, with some naptha, xylene, and a small amount of ethylbenzene. Definitely worth heeding the precautionary statements...

For contrast, "Goo Gone" is mostly naptha-like light petroleum distillate with small fractions of limonene and orange extract[2].

1. https://multimedia.3m.com/mws/mediawebserver?mwsId=SSSSSuUn_...

2. https://googone.com/mr_sds/data/2018-06-14%2013:44:51/GG_Ori...

Hot water "hydronic" heating with radiators is very common in the US as well, at least among homes in the Northeastern US constructed before the 1960s. It's a shame more new construction does not use hot water heating in conjunction with central air for cooling because radiators are much quieter, more efficient, and better for those with asthma.

It's worth mentioning that the package manager component of anaconda is released as a separate (small) install: miniconda[1]. It includes only Python and the package manager, and not all of the 700+ packages installed as part of a full anaconda installation.

With its ability to install Python and non-python packages (including binaries), conda is my go-to for managing project environments and dependencies. Between the bioconda[2] and conda-forge[3] channels, it meets the needs of many on the computational side of the biological sciences. Being able to describe a full execution environment with a yaml file is huge win for replicable science.

1. https://conda.io/miniconda.html https://conda.io/docs/user-guide/install/index.html

2. http://bioconda.github.io/

3. https://conda-forge.org/

They may be withholding physical samples, but some sequence data appears to have been released[1(eight segments of the virus from one patient),2(sequences from the last ten years)]. Sequence data like this is helpful for genomic epidemiology—modeling how the virus evolves over time and across geographic regions, as well as for investigating the potential functional impact of mutations.

1. https://www.ncbi.nlm.nih.gov/nuccore/?term=Influenza+A+virus...)

2. https://www.ncbi.nlm.nih.gov/nuccore/?term=(%22H7N9%20subtyp...)

Rochester, NY. Very affordable real estate, many of the prime areas of the city and suburbs are served by gigabit fiber, and between local universities and industry a surprising amound of technical talent is present in the area: some recently graduated, some the legacy of Kodak, Xerox, and other older companies, and some from defense firms like Harris, ITT, etc. The quality of life is very high with good restaurants, many local parks, festivals, and strong cultural institutions (great indie cinema, and a world-class orchestra). The costs are low, and it's still reasonably close to NYC and Boston in distance and culture. Sure it has winter, but it also has <$200k Victorian homes and tree-lined streets.

If you're interested in atomic history and the Manhattan Project, a great book is Richard Rhodes' The making of the atomic bomb (ISBN: 1451677618). It's ~900 pages of history, from the personalities of the scientists involved, to the sites of Oak Ridge, Los Alamos, and Hanford, and the first uses. I enjoyed reading it while interning at ORNL and Sandia in the summers during undergrad.

The weapons are horrifying in their destructive power, but the Project is a testament to what we can accomplish in science and engineering if we apply virtually unlimited resources toward R&D to solve a problem. Beyond the obvious product of the Project, it led to the creation of the DOE national laboratory system, which has since served as a means of maintaining an able technical workforce in the US, with physicists, mathematicians, chemists, biologists, and computer scientists, among others, all working in the public interest (and on things beyond weapons). Investing in science can pay dividends well into the future.

For some perspective, the cost of the Manhattan Project was on the same order of magnitude as the Apollo Program in terms of inflation-adjusted cost, about $22-26 billion in 2016 dollars. Of course that amount does not include the environmental cost still being felt in Hanford and elsewhere, or the human cost.

Knowing the cost of those two programs really makes you question some other efforts, like the program to develop the F-35 aircraft, estimated to cost $380 billion before even considering the additional $1.1 trillion in operational cost. For comparison there, the Interstate Highway system cost about $500 billion to construct.

Python Environment 8 years ago

Python environment headaches have been largely solved for me thanks to Miniconda[1]. Not only do the environments isolate dependencies, they can easily use different versions of Python, and can include arbitrary binary packages too. It helps that for my work in biology an extensive number of packages are available from the bioconda[2] channel (with many non-bio packages from conda-forge[3]). Environments can be described via environment files, allowing them to be transferred to collaborators, managed with source control, or included with publications to support reproducible science. If you like virtualenvs, you may want to give Miniconda a try. One current limitation is that environment files cannot specify a source channel for specific packages--they're installed from channels based on the global channel precedence.

1. https://conda.io/miniconda.html

2. https://bioconda.github.io

3. https://conda-forge.org

Yes, like us, fungi are opisthokonts, but where collagen is the mammalian structural protein used for making leather, fungal hyphae are principally comprised of chitin—structurally closer to the polysaccharide cellulose of the mentioned "kombucha leather" than to the collagen protein. Collagen is present in fungi in certain edge cases, but bulk mycelium is mostly chitin. This is speculation, but leather made from chitin may actually be a bit stronger than animal leather made from collagen. Maybe closer to keratin in strength? (Interestingly, keratin is chemically cleaved in hides during leather production--not entirely sure why, to enhance flexibility and allow access to the collagen fibers?). Agreed about fungi being easier to engineer. As you point out, endogenous expression of dyes in hyphae for different colors of mycelial leather is compelling. GFP would be cool too for the rave and nightclub crowd.

The Quorn meat analogue[1] is another product made from compressed mycelium. I've had "chicken" nuggets made from it. They're...not bad, but as far as mycelial structures go I prefer mushrooms.

1. http://www.misac.org.uk/PDFs/MiSAC_Briefings_1.pdf

The community of microbes resident on skin (or rather, their metabolites) has also been shown to influence attractiveness to mosquitos[1]. Interestingly, some taxa were associated with decreased attraction, which the authors of the linked study hypothesize is due to a masking effect. Skin probiotics may be on the horizon for mosquito deterrence and other applications (think acne, eczema, etc.). The composition of the skin microbiome is influenced by diet, host genetics, environment, and other factors (sun exposure, hygeine, etc.).

1. http://journals.plos.org/plosone/article?id=10.1371/journal....

A group upstairs at the Broad Institute built out a system to use Jupyter notebooks for analysis of genomic data, with backend computation happening on a Spark cluster[1]. Science on large datasets can happen via interactive notebook. In a connection with a recent GWAS on a massive dataset from the UK Biobank, the researchers involved decided not to write a traditional scientific journal article (at least for now) since their analyses will continue to mature. Instead, they've been posting insights online in blog form, with associated code on GitHub[2]. It's a daring move toward publishing at the speed of research. Once their conclusions mature, traditional journal articles may follow to distill and preserve the key findings. In the mean time, those in the field can apply the same code to their data, replicate the analyses, and get an early look at the output of the research. This works partially because the methods (univariate GWAS) are understood in the field and the interpretation and rendering of a particular dataset is the science in this case, rather than a new method (which would still likely warrant a paper).

1. https://hail.is

2. http://www.nealelab.is/blog/2017/7/19/rapid-gwas-of-thousand...

While it will certainly be hard to contain, there are historical precedents for containing disease outbreaks where small mammals are the vectors. A notable case is the Bolivian hemorrhagic fever, caused by the Machupo virus, where an outbreak in 1966 was largely extinguished after two weeks by extensive trapping of the rodent vector[1]. Of course the effectiveness of trapping depends on many factors, and it is not always feasible. Studies like the ones being conducted by the CDC are essential for understanding these diseases and inform strategies on how to contain them.

If you’re interested in reading about the initial cases of various emerging diseases (or rather, emerging diagnoses), Laurie Garrett‘s book, The Coming Plague, is a good read.

1. https://wwwnc.cdc.gov/eid/article/1/3/95-0308_article

I prefer physical buttons, or better yet knobs I can remove and soak to clean. My range has capacitive buttons for its oven controls, and every time I wipe it down with a damp sponge every button is triggered along with a cacophony of beeps. It could disregard inputs when multiple are registered in quick succession, but then each button would have to have a delay before action to see if it was pressed individually, and of course rapid legitimate use would pose a problem (as in the iOS calculator issue).

The Cathedral of Learning is such a beautiful building. I wonder though: did they leave any dirty patches as a small reminder? When they cleaned the ceiling of the great hall of Grand Central Terminal of the tobacco stains and train soot, a small patch was left dirty to show how filthy it was before the restoration [1]. As a kid, seeing that patch was a powerful argument to never smoke tobacco.

1. http://www.atlasobscura.com/places/grand-central-ceiling-dar...