In stat method life and then in R pkg life, I had to intentionally cultivate a neutral-to-positive attitude re: what others were doing — especially if the work is adjacent to mine. We’re embarked on a collective effort! This makes me much happier and more productive too.
In which I have reckons about why people write packages, and also data https://notstatschat.rbind.io/2018/12/17/what-are-packages-for/ …
-
-
-
Yes, that too.
ಸಂವಾದದ ಮುಕ್ತಾಯ
ಹೊಸ ಸಂವಾದ -
-
-
FYI
@Sophos blocks this website—labeled as gambling! I submitted a reclassification request.
pic.twitter.com/cDKIFzEpfu
-
Huh. That's weird. I don't think there are even any posts about the lottery.
ಸಂವಾದದ ಮುಕ್ತಾಯ
ಹೊಸ ಸಂವಾದ -
-
-
can you point me in the direction of people who have used it for rare-variant genetic epidemiology?
-
Here's the paper: https://www.ncbi.nlm.nih.gov/m/pubmed/29932245/ … And it has been/is being incorporated in GENESIS, on bioconductor
-
Ta muchly. Will add it to my break reading list.
ಸಂವಾದದ ಮುಕ್ತಾಯ
ಹೊಸ ಸಂವಾದ -
-
-
Question: If only five people will ever use a package, what's the probability that those five people will just be learning R for the first time…
pic.twitter.com/VEISBPbe90
-
When the major exemplar in your argument is SPSS you need to walk away slowly and take the L
ಸಂವಾದದ ಮುಕ್ತಾಯ
ಹೊಸ ಸಂವಾದ -
-
-
I haven't written any
#Rstats packages (yet), but whatever Stata packages I wrote were mostly to the extent of simplifying my own life... some while working on a tedious, voluminous homework in grad school; some while facing a recurrent task in real life. -
I don't think I hoped to improve anybody's analysis -- I would have written -polychoric- in a less capricious way; it's awful. Writing documentation is more like hoping that somebody will use your package *correctly*... and that's why -ipfraking- is close to 100 pages now.
-
Also I may be in a relatively unique position per
@sTeamTraen's reply to your post (which is golden, and I hope he reproduces it here). Of the 15K packages on CRAN, I would trust maybe a good couple hundred. -
Most of time I try something, it breaks with cryptic or no error messages, or crashes R... and as I usually just need to get stuff done, I go back and code all I need in Stata.
ಸಂವಾದದ ಮುಕ್ತಾಯ
ಹೊಸ ಸಂವಾದ -
-
-
This so very thoughtful and well written!
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
Another good (and for some, primary) reason why people write packages: for their own future self to read, understand, and reuse their code, months/years later (think for a paper revision or report update).
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
Here's another reason why people (me) write packages: The more people use the code I originally wrote for myself, the more errors and bugs are discovered. Where no errors are discovered, I can have more confidence in my own code.
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
Loving it! Thanks for sharing this analysis
@tslumleyಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
Some great thoughts on why people write R packages and how long they stay working.
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
-
-
I'm not doing R pkg now but use pypi, npm, and even an unofficial C pkg manager. The main purpose is reusability and modularisation. Packaging means getting your code base organised. You need a good boilerplate repo to wrap things up quickly (and not waste a lot of time)
ಧನ್ಯವಾದಗಳು. Twitter ಇದನ್ನು ನಿಮ್ಮ ಕಾಲರೇಖೆಯನ್ನು ಉತ್ತಮಗೊಳಿಸಲು ಬಳಸುತ್ತದೆ. ರದ್ದುಗೊಳಿಸುರದ್ದುಗೊಳಿಸು
-
ಲೋಡಿಂಗ್ ಸಮಯ ಸ್ವಲ್ಪ ತೆಗೆದುಕೊಳ್ಳುತ್ತಿರುವಂತೆನಿಸುತ್ತದೆ.
Twitter ಸಾಮರ್ಥ್ಯ ಮೀರಿರಬಹುದು ಅಥವಾ ಕ್ಷಣಿಕವಾದ ತೊಂದರೆಯನ್ನು ಅನುಭವಿಸುತ್ತಿರಬಹುದು. ಮತ್ತೆ ಪ್ರಯತ್ನಿಸಿ ಅಥವಾ ಹೆಚ್ಚಿನ ಮಾಹಿತಿಗೆ Twitter ಸ್ಥಿತಿಗೆ ಭೇಟಿ ನೀಡಿ.