wdkrnls

Last I checked China was doing pretty well. Soon they will have double the nuclear power plants of everyone else and already have a monopoly on rare earth mineral processing.

Implement a statistical software suite that ubiquitously uses this framework instead of the usual hierarchical mixed modeling tools whose assumptions often don't match what experiments were actually done.

From what I've read, DS9 was heavily based on the Babylon 5 Bible which was pitched by JMS to Paramount years before. You might charitably say DS9 was the Guix to the Nix of Babylon 5: Same core ideas mapped onto different story universes. The earlier B5 Bible apparently even had a changeling security officer which evolved into the "changeling net" plot shown in the pilot episode.

I heard the original story with O'Hare was for Babylon 5 to blow up after an alien attack and for the Babylon 4 to be sent forward from the past to replace it. We saw hints for that in two different premonitions in season 1. That's a pretty big departure from the story we actually got.

When I was in high school I learned AutoCAD and I remember that back then it was scripted in LISP. I'm not sure if that is still true.

Computer scientists had this idea that some things should be public and some things private. Java takes this to the nth degree with it's public and private typing keywords. R just forces you to know the lib:::priv_fun versus lib::pub_fun trick. At best it's a signal for package end users to tell which functions they can rely on to have stable interfaces and which they can't. Unfortunately, with R's heavy use of generics it gets confusing for unwary users how developers work with the feature as some methods (e.g. different ways to summarize various kinds of standard data sets as you get with the summary generic or even the print generic) get exported and some don't with seemingly no rhyme or reason.

Is there a way to trace an attribute to a function? I couldn't find one, but curious if it exists. I seemed blocked by the fact that trace seemed to expect a name as a character string. Some functions in base R have functions in their attributes which modify their behavior (e.g. selfStart). I ended up just copying the whole code locally and then naming it, but for a better interactive experience I really wish there was a way to pass a function object as I can with debug.

R works exactly as you describe. You can type `+`(1, 2) and get 3 because in R everything that happens is a function call even if a few binary functions get special sugar so you can type 1 + 2 for them as well. The user can of course make their own of these if they wrap them in precents. For example: `%plus%` = function(a, b) { `+`(a, b)}. A few computer algebra systems languages provide even more expressivity like yacas and fricas. The later even has a type system.

You have a very different definition of free than I do. Free to me means that people enter into agreements voluntarily. It's hard to claim a market is free when it's participants have no other choice...

Nope. R's advantage is that the language is extremely expressive and makes many things about it's implementation extremely transparent to it's users. The huge package count for a niche language is a direct result of that.

Speak for yourself. Programming in R is amazingly expressive for prototyping. Its semantics are extremely lispy, yet it provides excellent support for fast numerics. It keeps me focusing on the problems I want to solve and provides me the tools to abstract them quickly to handle related problems. Meanwhile, python keeps forcing me to care about pesky implementation details which I don't care about on a first or even second or third pass. I really can't understand people who like python over R. Did you guys not read SICP?

You must hate lisp/scheme then too, which has similar semantics as R. In that case books such as SICP would be lost on you.

This is so far from my experience. For me, R codes do tend to skimp on polish so it takes longer to get to the initial figure, but that is made up for by enabling me to see the data from a much richer perspective (to some extent because I had to think harder about what the output meant) such that I can find all the bugs in the data and in the underlying experimental plan: the stuff which makes it clear all the commercial reports are mostly useless anyway because Garbage in -> Garbage out

On the contrary, I find base R less arcane than the current de jour python libraries which copied it

Tyler is actually using R for exactly what R and it's predecessor S were designed to do since the beginning. You can read more about it's history by googling John Chambers who helped develop S at Bell Labs.

Typesetting and typography are very different tasks as far as I can tell. Scheme typesetting of documents is best done with TeXmacs.

TeXmacs can execute code too. Honestly, if it had 1/10 the community of Emacs, I would be using it for everything from running my window manager to driving my statistical simulations. It's already what Stallman keeps asking Emacs to become.

Sometimes knowing how to do something isn't nearly as important as badly wanting to do something

In other words, EM makes more sense. All this imputation stuff seems to me more like an effort to keep using obsolete modeling techniques.

Or it could be a problem of seeking statistical detection of any difference whatsoever versus detecting a practically meaningful difference... a type III error (answering the wrong question).

This looks quite promising to me. I've loved using TXR in the past for such functionality. However, I would really rather have the TXR pattern language as a library embedded inside the main programming language I am using rather than having to remember both TXR and TXR lisp. At first glance, Rosie seems to offer this approach. Although, I feel like the elegance of TXR syntax and semantics is hard to beat.

I've used this at work. It's very polished. I wish Linux had something this clean and simple to use. Honestly, I would prefer less UI, though. I just want something that works without fuss for the common use case.

As someone who is a data analyst, I am constantly reminded of the huge amount of necessary context needed to trust results from any analysis. I know that I lack some of that context, yet I also know that so do my collaborators on mathematical or computational aspects. A Shiny app makes it easy to get something in the face of everyone so we all can come to a consensus about what is useful and actionable.

How does this compare with ugrep? I know that does many of these things while sticking with C++.

Has anyone looked at writing a script to transform those rules for yacas? Making yacas 94% as good as Mathematica at integration sounds like a worthy goal considering how easy it is to deploy it.

As a tiling window manager user, I am really frustrated with mainstream web browsers. They are an ergonomic disaster for me, forcing me to use the mouse and waste so much precious screen real estate with gray space and redundant tabs. I want to disable tabs and just use windows managed by my window manager. That's what window managers are for! Separation of concerns is a pretty good notion in my book. Emacs can have a server and multiple "frames", so why can't Firefox or Chrome? Someone please tell me this is possible and I just missed it.

Your comment about Git LFS is intriguing. I have been put off from using git lfs because of the perception I needed a separate server to backup the binary files. I had heard of tools like lfs-folderstore (1). That no longer seems maintained and seems to be an extra service. Did git lfs integrate the lfs-folderstore ideas so it doesn't need them?

[1] https://github.com/sinbad/lfs-folderstore

Getting git to handle large data sets in regular repositories seems like a great way to go for many reasons. What is preventing the ideas of Xet from landing into git directly?

I really like that it tries to be extremely minimal, but was surprised to read it didn't support eval/apply at the lowest level.

I'm curious if anyone has a setup for emacs for either of these that supports audio/video calls. Emacs seems to have basic support for text chats on both protocols, but I wish I could use XMPP JMP.chat as my phone from Emacs and initiate video calls from Emacs as well. Is there some kind of audio/video server protocol analog to the language server protocol that let's plaintext and command line enthusiasts talk to each other without having to open that giant insecure web browser thing or a monolithic chat client?