I have been really impressed with @beaconbio.bsky.social, and find myself using it more and more. Excellent resource to replace ProMED.

I do want to fill this out, but "5 minutes" to consider all the options and decide on a ranking ... feels super optimistic ???

Screenshot of the announcement by Simon Urbanek regarding Tomáš Kalibera’s untimely passing. It is with great sadness that we announce the passing of Tomáš Kalibera on 1 April, a valuable member of the R core team for close to 10 years and a good friend, after a short but aggressive illness. Tomáš brought in fresh perspective and knowledge, enabling him to improve many aspects of R, including performance and reliability. He created many tools aimed at aiding package authors to make their packages more reliable, and was instrumental in modernizing the Windows build of R. He was an active member of CRAN and the R community, providing help to package authors, and he was the most prolific writer on the R core blog. He will be remembered for his profound contributions to R by millions of users. He is survived by his wife and 1 year old son. A full obituary will be posted in due time. Respectfully, Simon Urbanek

RIP Tomáš Kalibera. #rstats lost a huge contributor today. Condolences to his young family.

stat.ethz.ch/pipermail/r-...

Did it really translate "cwt" phonetically as "coot" ... ?? welsh-dictionary.ac.uk/gpc/gpc.html... (translators are offering me "hut"; is this idiomatically some kind of roadside stand ... ?) @q.pheevr.ca ?

OK, the `add_ranef()` + `mutate()` combo does the trick. `lme4::simulate()` and `glmmTMB_simulate_new()` fns are nice bec they use the same syntax as the model, but (1) make it easier to simulate *without* knowing what you're doing (pace McElreath) and (2) parameterization can be confusing

(2/?) clearly you can simulate 'by hand' if you're up to it (e.g. use rnorm() to generate random effects [or MASS::mvrnorm() for random-slopes models], then put together the pieces, or use any of the packages in the 'Power Analysis & Simulation' section of cran.r-project.org/web/views/Mi... ...

(1/?) @debruine.bsky.social will have more to say about this, but: a *quick* look at `faux` suggests that it's mostly for simulating *designs*. Once you've simulated the design, don't you need lme4::simulate() or glmmTMB::simulate_new() to generate the responses for a specified effect size/etc.?

same picture as before, but with overseas territories circled in red (two in the Pacific, at left edge; four in South Atlantic; two in north Atlantic [Bermuda, Gibraltar])

library(maps) library(mapdata) m <- map("worldHires", "uk") m2 <- map("worldHires", c("UK:Great Britain", "UK:Northern Island", "UK:Scotland", "UK:Guernsey", "UK:Jersey")) others <- data.frame(x = m$x, y=m$y) |> subset(x<m2$range[1] | x>m2$range[2] | y<m2$range[3] | y > m2$range[4]) png("uk.png") map("worldHires", "uk") par(xpd=NA) points(y~x, data = others, col = "red", cex = 2) dev.off()

PS:

a map of the UK that includes a bunch of overseas possession in the Atlantic and Pacific, so Great Britain is a very small blob in the upper right hand corner.

reminds me of this `library(maps); library(mapdata); map("worldHires", "uk")` (hint: in this old database "UK" regions included Gibraltar, Henderson, Pitcairn, Ascension, Gough, Tristan da Cunha, Saint Helena, and Bermuda, in addition to GB/NI/Channel Islands ...)

@cstross.bsky.social ?

I admit I haven't thought too carefully, but: any system that involves extensions/added-on bits (TeX, R, Julia, Python, Quarto ...) has the potential to hit dependency issues eventually (what if a colleague wants to build your project in 5 years in a then-current software ecosystem ...?)

A few times

I realize they're not packages, but ... they seem to be hosted on github. If see they're versioned/quarto-versioned (quarto.org/docs/extensi...), but I can still imagine dependency/conflict issues arising over time ... can a Quarto document specify/pin extension versions ... ??

The quarto extensions quarto.org/docs/extensi... are impressive but I also get a little bit nervous about whether this part of the ecosystem is eventually going to lead to a new version of package rot/dependency hell ...

OK, this definitely sounds like a "quantitative becomes qualitative" thing - i.e., incremental improvements to usability push past a threshold. I don't do much with the book/website/reporting etc. ecosystem; for me the new bells&whistles are nice but not must-haves ...

Post Doc position alert! Come work with us on the evolution of cannibalism. Full job description in link below. Salary starts at $58k plus benefits. (Please help spread the word)
emdz.fa.us2.oraclecloud.com/hcmUI/Candid...

Also FWIW this is only likely to be a problem if lots of people share a meal at (e.g.) ceremonial feasts. 1-on-1 cannibalism doesn't amplify disease well ... Rudolf & Antonovics. 2007. “Disease Transmission by Cannibalism: Rare Event or Common Occurrence?” Proc Roy Soc B doi:10.1098/rspb.2006.0449

Cooking is a good idea but as pointed out up-thread won't help with prions; autoclaving + sodium hypochlorite is recommended). BSE/vCJD prions might only be in CNS tissues but chronic wasting disease prions show up in muscle ...
Quirk 2006. Lancet Inf Dis doi.org/10.1016/S147....

PS this is not a faithful representation of what was in the recording ...

A page of complicated sheet music that is more or less gibberish

transcribe AMR (sound recorder) file to ABC automatically they said ... it'll be easy, they said ... (AMR →WAV [ffmpeg] → midi [ waon kichiki.github.io/waon/ ] → abc [abc2midi] → PostScript [abcm2ps] → PDF [ps2pdf]. Unsurprisingly, WAV → midi is the weak link. (Should try a simpler round trip next)

(2/2) Also easier to use/better tooling. Is this an example of "incremental improvements cross a tipping point"/network effect (quantity becomes quality)? Or is there something else I'm missing? (Serious questions, not trolling)

(1/?) What's revolutionary about quarto that doesn't apply to its ancestors (rmarkdown, Sweave, etc.)? quarto is certainly slicker/cooler, & people are making amazing extensions for it, & it's more polyglot, but for me those don't count as "transformational" ...

It failed for me just now. Maybe server is overloaded? Will try back later. "Error code: PR_CONNECT_RESET_ERROR; The page you are trying to view cannot be shown because the authenticity of the received data could not be verified. Please contact the website owners to inform them of this problem."

silly question, did you check the queue (`foghorn::cran_incoming()`) ?

I'm ignorant of Claude skills (still in 'have only dipped toes into LLM-coding' camp), but ... does this skill cover the kind of stuff that @mattansb.msbstats.info has posted about recently (model troubleshooting/simplification), or only interpretation à la `marginaleffects`?

Promotional graphic for a "Contributing to base R" event with Ella Kaye on Tuesday, March 31 @ 12 pm ET at pos.it/dslab. The background features a large blue R logo over faint grey statistical plots, including a scatter plot, histogram, and line graph.

Did you know you (yes, YOU!) can contribute to base R?

Base R is the foundation of everything #RStats. If you’ve ever found a bug or thought of an improvement, you can help shape the language.

Join @ellakaye.co.uk on the Posit Data Science Lab tomorrow to learn how to get started! pos.it/dslab

Glad to hear it. (Not all that surprising that horror stories are more salient/overrepresented in CRAN folklore ...)

Are you lumping together convergence failures (A [calc.derivs=FALSE], B, C) and singularity issues (A [complex random intercepts], F)? Are you considering the (new to lme4, available for a while in glmmTMB) diag()/cs()/ar1() options, or factor-analytic/reduced-rank (rr() in glmmTMB) models?

The deadly case of the Therac-25: a warning for medical AI AI systems have a black box problem, according to Omdia's Felix Beacher.

This immediately reminded me of Therac-25 (I had to look it up): en.wikipedia.org/wiki/Therac-25. Someone has already made the analogy to AI: www.worldhealthexpo.com/insights/ai-... (also, 🇨🇦 content, although not the kind you want ...)

Seems unfortunate if results come out on April 1 (which happened in 2015 and 2025 ...) [caveat: these are dates of internal notifications to candidates from university research offices, I think these might differ by a day or so across universities ...]