"We demonstrate that statistical software is used widely but rarely cited in political science, and we highlight a partial solution to this problem: software bibliographies. To facilitate their creation, we introduce softbib, an R package that scans analysis scripts, detects the software used in those scripts, and automatically creates bibliographies."
Bioinformaticians: If you see an R package in bioinformatics that is in CRAN instead of Bioconductor, does this raise a bit of suspicion? Why or why not? #bioconductor#bioinformatics#rstats
@lambdamoses
Having myself some packages on CRAN I guess just because it easier to have them published on GitHub. The quality control at CRAN is hard but worth it.
@lambdamoses@brodriguesco@Mehrad Can agree on several parts with that. Most of my packages are S3 one is S4. I never really saw a benefit for me regarding S4. On CRAN one can choose een R6, whcih is a profoundly different OO system from S3 and S4. On BioC S4 is the rule as far as I know.
@lambdamoses It does not make me suspicious. For me it is easier to deploy a package via CRAN. Just compare
install.package("Foo")
vs.
if (!require("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install(version = "3.17")
BiocManager::install(c("Foo"))
My tip for #rstats user new to the game. Take your time and browse through the functions available in the stats, base and utils packages that come with each #rstats installation. Just look at names that sound 'odd'.
You will find some things that are otherwise promoted by auxiliary packages or for things you start to code by yourself.
My example is trimws() from base to remove leading/trailing whitespace. I have seen rather complicated proposals over at #stackoverflow with #regex and alike.
Help page of the trimws() function of the base #rstats package, shown in #RKWard
@elinwaring True. I've been using #rstats since 2002 and keep discovering new things that are actually quite old. The advent of #LLM has shown me how centered solutions seem to be. Some of my students have used it for their assignments and generated code that works, but unfortunately, is not elegant. Many solutions in LLMs circulate around packages, which bring more dependencies and thus also make projects a bit harder and less reproducible. Base packages are often a remedy.
Delta Chat 1.38 releases: WebXDC apps can now exchange data with other tools and run on other messengers! We've implemented a little Calendar and Draw app to showcase the two new APIs and besides have other good news for the nascent privacy-guaranteeing Next-Generation-Internet ... just a "start" button for chat-shared apps with no GDPR or cookie-consents, no coins, no app or account hosting and thus no logins ;) https://delta.chat/en/2023-06-28-webxdc-import-export
Recently, I switched from Dell (XPS 13) to a #Tuxedo laptop. This machine is just great. I also use the Tuxedo OS, which severs me well while working with #rstats.
Tuxedo avoids #snap, which caused some mild trouble since I use #LanguageTool as snap. There is no #flatpak as far as, I know. However, one can get a #java jar her: https://languagetool.org/download/
Works nicely with #openjdk. No need to use snap at all.
Occasionally, I think about how to work effectively with #rstats. Currently, I am teaching my #bioinformatics courses with #RKWard again. I try to do most of it with packages from the base installation. #datatable is an exception. But otherwise, I like to use #within (very fast) instead of #mutate.
But there are more approaches, which are often simpler/faster/stable:
@magljo
I always try to keep it simple. This comes often with speed. And the dependency hell? Well, some of my packages have many dependencies. Unfortunately. Often not really needed. At least from my more experienced perspective. Guess they could be faster.
Lot's of my R work/knowledge dates back to 2003. Maybe, because of this I considere more often the base approach.
They write “Veraltetes Android-System” as a disadvantage. This means outdated Android system. It seems to me that this is not entirely true for all but some of your offered devices.
Anyway, nice to see that.
I have received an unsolicited message (aka SPAM) for a statistical software package called SEM-PLS. Looks like a GUI for R to me. Some base plots, some #ggplot2, some statistics.
"You can purchase SEM-PLS 2023 with an 80 percent discount(Only 66 euros)"
I didn't see this coming. #DeltaChat is discussing to implement a feature that that other messengers already have: #RFC9078
I never expected this to happen. 🥳
danke für alles ❤️ nicht zuletzt deine gelebte Toleranz gegen meine Emacs-Affinität. #Schnitzelhaus (und vieles andere) will never be the same without you.
#moolenaar#vim#SvenGuckes
It is sad to hear that Bram #Moolenaar now left us too. Would :wq! be an appropriate description. I mean he left behind a lot.