closed

> When I wrote anti-chinese, I meant at the time

Ah, okay. I read your original post as critical of scmp as anti Chinese today, but am not sure if that's what you meant now.

How is Tse Tsan-tai a western, anti-chinese propagandist? He was a collaborator w/ Sun Yat-sen. I realize the two drew influence from western ideals (and christianity), but wonder what makes them anti-chinese.

That the Chinese government praised Sun Yat-sen, and printed a commemorative coin for his 150th birthday, seems at odds with your sentiment.

http://www.chinadaily.com.cn/china/2016-10/25/content_271713...

We ran into this problem often teaching R when I was in grad school. However, in the years since, the tidyverse has put strong focus on getting users manipulating data even if they don't know how to define a function, etc..!

Here's a useful post, comparing the classic approach you mention to an alternative

http://varianceexplained.org/r/teach-tidyverse/

What do you hate about it?

You might be interested in the design principals of the tidyverse, a group of R packages built to support data analysis.

https://design.tidyverse.org/unifying-principles.html

One emphasis of theirs that I find really powerful is the "pit of success", where the least effort action moves toward a positive outcome.

Broader overview of tidyverse here:

https://tidyverse.tidyverse.org/articles/paper.html

> Who among us couldn't guess what the NYT's take is going to be in a mormon leader's obituary before they write it?

Their obit for the previous president, Gordan B Hinckley is quite positive...

https://www.nytimes.com/2008/01/28/us/28hinckley.html

Edit: here are more if people are curious. One before, and one after the Mormon church allowed black people similar position as non-black people. They're still pretty upbeat.

https://www.nytimes.com/1970/01/19/archives/david-o-mckay-mo...

https://www.nytimes.com/1985/11/07/us/spencer-kimball-mormon...

> single-handedly

There are many people in the R community working on this together (e.g. Jenny Bryan, Charlotte, etc).

> lapply(), sapply(), tapply(), vapply() each does something different.

The apply situation has been standardized through the purrr lib and dplyr for a long time. They are base library functions that aren't mandatory.

> two kinds of assignment operators even

Consider the custom of using <-. It reduces the kinds of assignment operators to 1. Similar to avoiding from lib import * in python. You can do it, but there are community standards against it.

I strongly disagree. And in general, it seems like poor quality code in science is often not because of the language, but because scientists rarely lose their jobs when code breaks.

There are many books that cover how to develop in R in detail, and they are no less thorough than treatments of the subject in other languages (e.g. Hadley's books are as good as any I've read for python).

Many issues around inconsistency, etc, in language design (mostly how base functions / data types behave) have very clean, consistent implementations in libraries like rlang.

The main differences I see when comparing R vs python package code, that affect style are...

1. Most R operations are immutable.

2. R often uses single dispatch, rather than putting methods on a class object.

3. In R, vectorised behavior is often the norm.

4. R functions can choose to use lazy evaluation (it usually very clear when this happens in e.g. tidyverse packages).

These issues are covered in detail in books like Hadley's Advanced R.

> How is it possible that two opposing theses could be maintained in something like neuroscience

It is very common in neuroscience (and also from experience psychology; but I'd bet also econ and Phil). Popular science books often paint results as ironclad, and gloss over small, or convenience samples, methodological issues that might cause researchers to disagree, etc..

Another big one in psychology and neuroscience (and many areas) is that even if you have a clear result, the interpretation often requires buying into certain assumptions. (E.g. interpreting activity in fMRI when no data from methods with better spatial or temporal resolution; intepreting non human animal studies)

Although, beware that you may need to change `find_packages()` to `find_namespace_packages()` in your setup.py.

I get burned by that one a lot!

The style of this series reminds me of the book, "Excercises in Style", which is the same short (2 page) story told in 99 different writing styles!

(Not sure about the style of this individual story)

https://en.m.wikipedia.org/wiki/Exercises_in_Style

What? Can you put in probabilistic terms what "this is not the case" is?

There are an infinite number of models where p(HHTHT | model) != 1, or where p(HHTHT | model) = 0. We need to know which one you're referring to, in order to calculate a p-value.

I think you have made a serious error by believing you can simply "reverse" the model p(HHTHT | conspiracy model) = 1, p(everything else | conspiracy model) = 0.

If the null hypothesis is a fair flip, then the alternative can't be a conspiracy, because the null and alternative need to be complementary statements. So if the null is fair flip, then the alternative is "not fair flip".

edit: whoops, changed mutually exclusive to complementary. see http://www.its.caltech.edu/~mshum/stats/lect8.pdf

Wait, in frequentist statistics getting, say, a p-value of 1 is not a bad thing--unless you erroneously assume that value is evidence for your null hypothesis.

Consider that if your data generating process really is a fair coin, then the conspiracy outcome you mention only occurs 1 our of 16 times, so 15 out of 16 times you observe a likelihood of 0. 15 out of 16 times your reject the conspiracy case.

There is also a tricky component here, because the notion of sample size is not clearly defined (can we generate multiple 4-tuples of flips, and consider each one a sample? Is your example really just a funky way of discussing type II power?)

I'm really enjoying Types and Programming Languages right now. Imagine that rather than explaining typing using metaphors (a useful activity), people started at a much lower level and built up from there. That's this book!

https://www.cis.upenn.edu/~bcpierce/tapl/

This reminds me of a lot of Matlab code I had to read in grad school :/.

It seems worth noting that Anders Ericcson's handbook on expertise has chapters studying business expertise (though I get the feeling C level management gurus aren't reading it; there's an interesting chapter on entrepreneurs though!).

https://www.amazon.com/Cambridge-Expertise-Performance-Handb...

> the literal opposite of democracy since it allows people with no understanding to vote.

I would need to know the definition of democracy you're using to fully appreciate the meaning of this comment.

I think there's an important, main point in OP's comment that you are missing.

Hong Kong has never really had much democratic process, in terms of direct elections, and universal suffrage. There are only some roles where this happens.

Hong Kong has a long history of protesting this, and of failed attempts of implementing more democratic process.

To say it was democratic under British rule is a reach.

https://en.m.wikipedia.org/wiki/Democratic_development_in_Ho...

I don't think this lines up with memory research. It's true that people commonly make source monitoring errors, but generally feeling negative affect toward something enhances memory. People might forget the specific company, but I doubt they would feel more trust toward travel companies in general.

https://en.m.wikipedia.org/wiki/Source-monitoring_error

https://www.apa.org/science/about/psa/2011/10/positive-negat...

nbsphinx is super handy! A cool tool to combine with it is jupytext, so you can keep your notebooks as rmarkdown files, which is a bit more human readable / GitHub editable.

I did this with a recent library, siuba, and have not regretted it!

https://github.com/machow/siuba/tree/master/docs

This is super interesting! I've never seen codecs used to introduce language features like this... Do you have any good references for getting started with using codecs?

Also, I wonder if there's an easy way to implement this in a jupyter notebook, without fiddling with the kernel...

(issues of whether it's a good idea aside, definitely seems useful to know about.. )

> If you're struggling, you're learning.

I think this is close, but reversed. Learning requires a struggle (eg effort), but you can struggle without learning (eg you aren't getting good feedback).

Thanks! I've been studying Cantonese a couple years, but just learned that when there's a choir + orchestra, I have 0% comprehension of even the sounds of the words they're making :o.

I've thought a lot about the speed, and am very curious from an organizational perspective how say, a person managing an engineering org might not feel the same frustration from slow page loads.

Do they only open it occasionally? Do the pages they use open more quickly?

Anyone managing teams of engineers that can chime in? :)

One important caveat is that the non-random strategy would need to be invertible. For e.g. a ml model doing binary classification this is easy (switch the classification), for most other things it's not clear what invert means.

Highly recommend first reading the wikipedia entry on working memory capacity. Nelson Cowan's work showing that a limit of 4 +- 1 pops up quite a bit, and subsequent investigations of why earlier research got 7 is key here. 7 shows up when people are allowed to subvocalize, and are being tested on "verbalizable" material.

https://en.wikipedia.org/wiki/Working_memory#Capacity

https://en.wikipedia.org/wiki/Working_memory#cite_note-23

> ... ask about arrests and not convictions

Expungements are for clearing arrests[1]

> or an employer fires you for a conviction found on a third party database what are you going to do?

Ask for a copy of the report (they are required by law to provide it to you; you will know they have it because they had to get your legal consent beforehand; have to inform you decision is based on report), then dispute the report.

A company can do dishonest work and pretend their decision is not based on a report, but they can't legally use expunged records to make a decision in PA (or afaik anywhere; sealed records are another story), just because they got them from an outdated consumer reporting agency.

1: http://plsephilly.org/get-help/expungement/

They failed to replicate the study over multiple attempts, 1 with four times the sample size.

Using the time reversal heuristic, if we flip the order, so the replication came before the original, how would you feel about Science publishing a paper with a positive result, when multiple studies--with larger samples and the same materials--had not obtained the result?

If you think the original should still be published in Science in that case, why? Since the original is published and highly visible, it's dangerous that Science won't correct the highly publicized, unlikely-true result.

Is this true? Couldn't we run a process on top of a lot of humans, which, assumming they are weak learners, garuantees strong learning (i.e. boosting)?

These days, I often use Rmarkdown for both python and R work. On the R side, Rstudio makes it a pure delight. On the python side, jupytext makes it easy to open / edit Rmds as notebooks.

Hats off to the many people in these communities working to give us the best of both worlds :).