Kurt Schwehr

@kurtschwehr

he/him; Geospatial, Geology, Geophysics, Software Engineering, Open Source, Open Data, Proj, GDAL, Ocean Mapping, etc. Cofounder of Global Fishing Watch, Whale Alert, NOAA's Environmental Response Management Application (ERMA), NASA Ames Viz&StereoPipe

279 Folgt457 Follower446 Posts

YouTube Bluesky

Ist das dein Profil?Übernimm es — dann gehört dir kurtschwehr.dirty.social.Profil übernehmen

Kurt Schwehr@kurtschwehr· 21d

Ugh. I know we are all overworked open source maintainers, but when you say "I don't mean to be rude"... I'm trying not to waste peoples' time. Security is a royal pain in the ass, but attackers have lots of resources and people get freaked out about SHA1 & MD5. github.com/pyca/cryptog...

Kurt Schwehr@kurtschwehr· Jun 24

A great horned owl staring down at me from the roof of our house this evening

Kurt Schwehr@kurtschwehr· Jun 7

Our apricot tree is doing better this year. Turns out we were putting softened water on it until this year. Sigh. The fruit are just starting to get ripe, but are pretty tasty when they have been sitting in sugar for a while.

Kurt Schwehr@kurtschwehr· Jun 6

I've been exploring Rust using Bevy + Ratatui for some simple text based games. It's interesting, but the time required to build and test the games is painful for even small stuff. I'm assuming this is mostly because bevy is huge. A view of my yahtzee game...

A Rust + Bevy + Ratatui yahtzee text game in a terminal

Kurt Schwehr@kurtschwehr· May 1

I take issue with this definition of "terrain elevation:" Emphasis is mine. Terrain elevation is only relative to MSL in some instances. There are many vertical datums. Sigh. gcmd.earthdata.nasa.gov/KeywordViewe...

NASA's GCMD Keyword Viewer entry for "TERRAIN ELEVATION." They have "The vertical distance from mean sea level to a point or object on the Earth's surface." Apparently this is from: Dictionary of Geological Terms, American Geological Institute.

Kurt Schwehr@kurtschwehr· Apr 27

Alice hiding under the sheets on our bed

Kurt Schwehr@kurtschwehr· Apr 27

Today I made a clone of the 1972 game Pong using Rust and Bevy. I don't know the language of the game engine, so I used Google Jules to do 99% of the coding. It was about 10 hours from start to finish, but most of the time I was doing other stuff. Baby steps...

A picture of the screen showing a simple version of the game Pong at the point where one player has won and the game is waiting for someone to press the space bar to start a new game.

Kurt Schwehr@kurtschwehr· Apr 7

Falcon 9 launch - Starlink 17-35 mission. Was scheduled to lift off at 7:50 p.m. PDT from SLC-4E at Vandenberg Space Force Base.

Kurt Schwehr@kurtschwehr· Apr 5

#caturday

A tortoise shell cat Alice passed out on a bed

Kurt Schwehr@kurtschwehr· Mar 26

I am pretty sure this isn't a sensical sentence: "The band fires grayscale ultrasound at tendons". Isn't it a grayscale visualization, not the specific waveforms from the transducer. Sigh. The original does not have that. From the summary in robotnews.therundown.ai/p/amazon-now...

Kurt Schwehr@kurtschwehr· Mar 12

Shame on me for not following the PROJ mailing list policy :)

Kurt Schwehr@kurtschwehr· Mar 12

The COPERNICUS/DEM/GLO30 Digital Surface Model (DSM) now has a much more detailed example script in Earth Engine. Give it a try! developers.google.com/earth-engine... code.earthengine.google.com?scriptPath=E... #earthengine

Kurt Schwehr@kurtschwehr· Mar 8

At the park today. I think Jay is scoping my kid's apple

Kurt Schwehr@kurtschwehr· Feb 26

Sigh. Vertical datums are super important, but seriously confusing. Mean Sea Level (MSL) is not what most people think it is. Sigh. vdatum.noaa.gov/docs/datums....

Kurt Schwehr@kurtschwehr· Feb 26

Slowly working through cleanup of geemap... a long ways to go.

Kurt Schwehr@kurtschwehr· Feb 19

Doh! @globalfishingwatch.org

Kurt Schwehr@kurtschwehr· Feb 11

A new article for my reading list: "A framework for assessing the trustworthiness of scientific research findings" www.pnas.org/doi/10.1073/...

Gemini gen checklist:

1. Is it Accountable? Integrity
* Oversight: Approved by an ethical board (IRB/IACUC)?
* Disclosures: Funding sources and conflicts of interest clearly stated?
* Corrections: Clear policy for correcting errors rather than hiding them?

2. Is it Evaluable? Transparency

* Openness: Raw data, code, and detailed methods available for scrutiny?
* Reproducibility: Is there enough detail to repeat the experiment exactly?
* Access: Open Access?

3. Is it Well-Formulated? Design
* Pre-registration: Hypothesis defined before data collection/p-hacking?
* Fit: Study design appropriate for the specific question asked?
* Context: Boundaries of where findings apply clear?

4. Has it been Evaluated? Review
* Peer Review: Has it undergone independent expert review?
* Community Feedback: Shared as a preprint or at conferences?
* Replication: Have independent labs successfully replicated these findings?

5. Does it Control Bias? (Objectivity)
* Safeguards: Did the study use randomization and blinding/masking to prevent bias?
* Reporting: Did they report *all* results, including negative or "boring" findings?
* Credibility: Are qualitative interpretations checked against participant perspectives?

6. Does it Reduce Error? Rigor
* Power: Was the sample size large enough to support the conclusions?
* **Calibration:** Were the instruments and methods valid and reliable?
* **Self-Correction:** Did the authors actively try to find and root out their own errors?

7. Are Claims Warranted? (Evidence)
* Alignment: Do the conclusions match the data without exaggeration?
* Humility: Are limitations and uncertainties explicitly acknowledged?
* Generalization: Do they avoid applying findings to populations not represented in the sample?

The "Proxy" Trap

* Journal Reputation: High "Impact Factor" does not guarantee truth.
* Author Fame: Prestige does not prevent error or bias.
* Novelty: "Groundbreaking" results are often riskier and less reliable than incremental science.

Kurt Schwehr@kurtschwehr· Jan 23

Santa Cruz, CA sunset from two days ago.

Kurt Schwehr@kurtschwehr· Jan 15

GitHub message search doesn't seem to do what it is saying. I ask for "EarthEngine" and get a whole lot of not Earth Engine :(

Kurt Schwehr@kurtschwehr· Jan 13

Well, this sucks! "Santa Cruz County is ranked worst in the state for victims killed and injured in bike crashes." www.ots.ca.gov/media-and-re...

Kurt Schwehr@kurtschwehr· Jan 9

TIL more than just gdal_translate can use all the CPU cores for decompression. So, pass `--config GDAL_NUM_THREADS=ALL_CPUS` to gdalinfo for massive compressed files! 🚀 github.com/OSGeo/gdal/i...

crschmidt:

When running gdalinfo -stats on a large GeoTIFF, it uses only one core to compute stats. It would be nice if (for thread-safe datasets) it could run with more than one core.

(Seems like this might be more practical now that https://gdal.org/en/stable/development/rfc/rfc101_raster_dataset_threadsafety.html is in place?)

In particular, running gdalinfo -stats on CAFA_2020 from https://drive.google.com/corp/drive/folders/1exF1KpGr0EJqZ10yqPIkzqLfh6gGj3Sa took 8 minutes on my Macbook with an M3 chip, which seems like it could have gone faster with multi-threading.

<snip>

rouault

@crschmidt Wouldn't https://gdal.org/en/latest/programs/gdal_dataset_check.html added in master a couple days ago do the job as far as your goal is concerned? (i.e. it reads all blocks ensuring they don't result in errrors). It is not multi-threaded per se. That said if your GeoTIFF is compressed and you specify "--config GDAL_NUM_THREADS=ALL_CPUS" then you'll get multi-threaded decompression, which should reach close to optimal performance.

Kurt Schwehr@kurtschwehr· Jan 6

Asking NASA's GDMC keyword database for AIS, NMEA, and the like be added. forum.earthdata.nasa.gov/viewtopic.ph...

My request for additional keywords at https://forum.earthdata.nasa.gov/viewtopic.php?t=7493

The body:

I'm working on some documentation for Automatic Identification System (AIS) and I was wondering if it's possible to get entries key ship tracking systems? Tags for AIS (instrument and format), NMEA(format), NMEA TAG BLOCK (format), Vessel Monitoring System (VMS; instrument), ORBCOMM (platform), SpaceQuest's AprizeSat (platform) would be helpful. I'm sure I'm missing a bunch more critical ones. I see that GCMD has a keyword for RINEX for GNSS data which is more raw than NMEA, Spire who has AIS receivers on their satellites, etc., so it looks like these tags would fit into the space of keywords.

Kurt Schwehr@kurtschwehr· Jan 4

Did a touch of work starting on a book called "AIS Open Manual" in github.com/schwehr/ais-...

Gemini generated logos for the AIS Open Manual

Kurt Schwehr@kurtschwehr· Dec 27

I asked Gemini to "Turn this drawing into a 3D futuristic cartoon" from a pencil drawing I did when I was 7-8 years old. I got this back.

Kurt Schwehr@kurtschwehr· Dec 27

GitHub isn't my primary code space, but I had a bunch in 2025. Thank you to all who contributed to open source at any scale... from a single commit/pr, issue, or comment to folks who churn out volumes of contributions. Open source needs everyone. We all stand on each others shoulders Escher style.

Kurt Schwehr@kurtschwehr· Dec 22

Skittles and warm water for fun experiment with the kids. It was hard holding the phone up so long to get this tinelapse

Kurt Schwehr@kurtschwehr· Dec 17

We have had some nice sunsets in Santa Cruz in the last week.

Kurt Schwehr@kurtschwehr· Dec 11

I asked nano banana to do this. Kinda fun results. The prompt: "Create an image that shows who Kurt Schwehr is professionally and how his work all interrelates."

A graph with pictures show Kurt Schwehr's work on oceanography & marine science, robotics & space exploration, and software engineering & data visualization.

Kurt Schwehr@kurtschwehr· Dec 2

So, I'm doing some reading on prompt engineering and I'm to the section on role prompting. So here is what I tried... I want you to act as an RF receiver.

Kurt Schwehr@kurtschwehr· Nov 29

A window from the cottage we stayed in last week