I think that all kinds of jargon and style canons (like using latex for academic papers or the special unwritten rules for formatting movie scripts) are primarily an in-group/out-group mechanism.
By communicating in the “expected” way, you are communicating that you are part of the in group.
As a side note, writing for a broad audience is harder than writing stylistically. You have to not only understand all the concepts involved, but you have to be able to accurately convey those concepts in simple sentences without the use of jargon. I believe this is a rare skill.
In academia, for the last few years there has been a push for Plain Language Summaries (PLS) as an accompaniment to traditional abstracts. This is a step in the right direction IMO, because many people don’t even bother reading the paper, or give up quickly, if it’s overly obtuse.
Most of the time, obscure language can call out to C. Therefore you should almost certainly just use FFI and leverage a C implementation of whatever crypto algorithm you want to leverage.
True, but encryption should generally be an exception to that rule. Not that C is good for writing encryption, but because there are so many weird issues with encryption which can result in an implementation that passes all the test to still be severally broken. At least the C version has had a lot of experts looking at it and preventing those issues.
You shouldn't be rolling your own crypto primitives. You can completely implement the algorithm 100% 'correct' according to the research paper but introduce a side channel that could cause key extraction by an attacker. For instance, if it doesn't always take the exact same amount of time to process something, a timing attack can be used to figure out what the private key is.
Always use the battle tested implementation.
Power analysis, timing attacks, acoustic cryptanalysis, etc... there's many forms of side channel attacks that can be used to defeat a theoretically sound cryptosystem.
Related: in tptacek’s Cryptopals/Matasano security challenge, there are two kinds of problems:
A) Implement this off-the-shelf cryptosystem based on the public documentation about it.
B) Given this cryptosystem and these hints, find and exploit a vulnerability.
Surprisingly, I found the type A problems harder — because the documentation was always missing some critical knowledge you were just supposed to know.
Have you, like, tried to write an academic paper without LaTeX? It’s only in the past few years that viable alternatives (Sile, Typst) have been available, and they all owe a lot of their design to LaTeX.
LaTeX made quality typesetting readily available to non-typographers. It’s the opposite of gatekeeping.
TeX/LaTex has revolutionized technical publishing, enabling the easy (and free) typesetting of very complex documents. Its strange to cast it as "jargon and style canon" and does not help the rest of your argument.
And you are 100% right but I do not think the point is against LaTex. But I am willing to bet money that after seeing thousands of pdfs formatted for icml/cvpr/nips, a reviewer would have an unconscious bias towards a pdf printed from msword or markdown. That's just a group thing and not that unexpected.
Yes it has. And documents created with it have a distinct recognisable look which is instantly recognisable to others that also use it. This goes to the in-group/out-group argument of the GP.
I agree about the recognizable look and its subtle second-order effects, but using LaTex as the typical in-group/out-group example is problematic when its use is a precondition for achieving a workable outcome.
Its like saying that carpenters are using their toolkit to merely signal professionalism.
In fact the same ambiguity applies to in certain cases to the original post (see my other comment). If people are forced to use a certain communication technology / form by technical or legal reasons then this is not a good example of in-group/out-group. Such examples are much better served by discretionary choices.
> Have you, like, tried to write an academic paper without LaTeX?
It's only the core hard sciences that use LaTeX. Mathematics, computer science, physics, part of statistics, part of economics, part of engineering. When you move away from this core to e.g. chemistry, biology, applied physics, then it's all MS Word.
This is one of the reasons why it's so important to be able to explain your specific field in simple language.
Not only does it mean you understand your field, but it also means that you have developed your personality to look at things from an outsider's perspective.
You have matured to be empathetic.
This is also why having kids is a major step in being mature. You have to explain things from their perspective.
I disagree. If I'm reading a guide on how to set up say on on-prem enterprise password manager, I don't want it to explain what "files", "network paths", and "LDAP" is. If you can't understand the document, it means you are missing prerequisite knowledge and it's not everyone's job to get you up to speed on that. Assumptions have to be made for documentation to be usable for its intended audience.
By communicating in the “expected” way, you are communicating that you are part of the in group.
As a side note, writing for a broad audience is harder than writing stylistically. You have to not only understand all the concepts involved, but you have to be able to accurately convey those concepts in simple sentences without the use of jargon. I believe this is a rare skill.
In academia, for the last few years there has been a push for Plain Language Summaries (PLS) as an accompaniment to traditional abstracts. This is a step in the right direction IMO, because many people don’t even bother reading the paper, or give up quickly, if it’s overly obtuse.
Law could take a lesson from this.