You can have debug symbols even in release builds. With gcc, `-ggdb` for debug symbols and `-O2` for optimization can be selected in dependently of each other. Optimized binaries are a bit harder to debug, but locating the source of this crash should be easy even in optimized versions.
As additional complication, debug symbols are often removed from the binary post-build to reduce binary size. The `strip` utility either discards debug symbols entirely, or it puts them in a separate folder as .dbgsym filses. See the gdb `debug-file-directory` option and the `add-symbol-file` command.
Development feedback—intended to help team members learn and improve—needs an environment where the recipient can absorb information. This requires psychological safety: no blame, no pressure, no personal attacks, no consequences, no reason to get defensive.
Performance evaluation—informing employees how happy the company is with their output—is a high-stakes and high-stress situation that precludes learning.
The author speaks with an authoritative voice and presents the content as facts. Sadly, the article also contains factual errors. For example:
> This capacity could also be served by a fleet of just 40 737s [...], of which Boeing makes more than 500 per year. Bought new, this fleet would cost $3.6b, and with a lead time of, at most, a few months.
Boeing has a backlog of 4000 planes.[1] Current delivery lead times are 5-10 years, so getting 40 planes within months is ludicrous. Aircraft might be available on shorter timelines from aircraft lenders, but probably not in that timeframe either. It's also not what the article argues.
This puts a question mark over the content: which parts are actually correct and which are merely presented as fact without any checking?
As of January 2021, SHA256 repositories are supported, but experimental. They can be created with `git init --object-format sha256`. If I understand correctly, they don't mix at all with SHA1 repositories (i.e. you can't pull/push from/to between SHA1 and SHA256 repos).
Organic synthesis is about making new molecules that haven't existed before. That's important for developing new drugs, improved batteries, better plastics, etc.
Organic molecules are networks of atoms, with both a distinct connectivity pattern, and a specific 3D orientation [1] of the atoms to each other. See e.g. the Wikipedia page of Lipitor[2] a picture of the connectivity pattern.
We build these molecules through chemical reactions. Over time, we have become pretty good at creating the connectivity patterns we want. However, achieving the correct 3D arrangement is still challenging.
List and MacMillan developed new chemical reactions that enable us to get both the connectivity, and the 3D aspect right. Such new methods are frequent Nobel contenders, and won e.g. in 2001 with Knowles/Noyori/Sharpless.
As for how these reactions work: it is true that they are catalyst-based and that catalysts speed up reactions, but that perspective is a bit misleading. The key point is that without catalysts, these reactions would not happen at all. So the catalysts List&MacMillan found accelerate some desirable reactions so much that they turn from "practically doesn't happen at all" to "done in an hour".
Congratulations to the outstanding work, and to the Nobel price!
I've been wondering a few times about how to best search the medical literature given a set of symptoms. How did you search for and find your particular disease?
Honestly, it is a lot of work. I searched for years, whenever I had "free time" in engineering school, because my problems never completely added up and I could never put my finger on it. I very well felt like I was on borrowed time and I knew I was in a lot of trouble, health-wise. It took me about 4 years to figure it out.
You have to keep an open mind, and PUBMED [1] is your friend. Also, if you do not have institutional access to journal articles, Sci-Hub [2] is your friend. Libgen [3] can also be your friend too.
Searching Google Scholar with advanced key terms also helps, but I find it important to keep a paper log of my queries, so I can also follow my thought process.
Ultimately, you have to learn how to be able to "play" with bits and pieces of information that you get from the articles, and make it (relate it) into something meaningful to you. It requires a lot of intuition, and if you find yourself bored, then you are not doing it right. Some people just have this sort of intuition, but anyone honestly can do it. Just focus on never getting bored with the information, and you will eventually learn this skill.
Anyways, the rare disease I have is called autoimmune autonomic ganglionopathy [4]. I wrote a story [5] on how I figured it out awhile back. People seem to enjoy reading it.
I guess the best way to answer your question: I basically searched everything I could about autonomic neuropathy (autoimmune autonomic ganglionopathy is a form of autonomic neuropathy--and I had already been diagnosed with "diabetes-realted autonomic neuropathy"), to see if there was anything that I could possibly relate to my situation. When the situation really got blown out of proportion, I started looking hardcore at really rare stuff.
If you are looking in terms of software-based solutions for an issue like this, the best resource is likely via Stanford University's SNAP group, which publishes BioSNAP datasets [1], which can be used for scaling.
For example, in the US, a lot of people are on a bunch of prescription drugs. This is called polypharmacy. Using AI, the SNAP group created AI to identify side-effects when on several drugs [2]. There is excellent sample code for this via the link I provided, that can be viewed on GitHub. Generally this is the case for all of the BioSNAP repositories.
There are also AI tools which Stanford has created which can help augment a deep search through the literature for a rare disease. For example, this "Disease-function association network" [3] can give useful outputs to help one direct a search for finding a certain rare disease.
> "This is a disease-function association network that contains information on relationships between diseases and cellular functions. Cellular functions capture biological processes (e.g., pathways made up of the activities of multiple proteins such as cell communication), cellular components (e.g., components where gene products are active such as mitochondria), and molecular functions (e.g., molecular activities of gene products such as drug binding). Nodes represent diseases and functions, and edges indicate associations between them."
The problem with AI is that it is intellectually bankrupt: it will tell you what it thinks, but it will not tell you why. So, it is critical to develop excellent intuition as an individual.
Cooley-Tukey Fast Fourier Transform can be complex to implement, if you want to optimize it. A basic version is surprisingly simple, I implemented it for fun in Python a couple of years ago:
def fft(x):
N = len(x)
if N == 1:
return x
assert N % 2 == 0
E = fft(x[::2])
O = fft(x[1::2])
X = [None] * N
for k in range(N // 2):
tf = np.exp(-2j * np.pi * k / N)
X[k] = E[k] + tf * O[k]
X[k + N//2] = E[k] - tf * O[k]
return X
As additional complication, debug symbols are often removed from the binary post-build to reduce binary size. The `strip` utility either discards debug symbols entirely, or it puts them in a separate folder as .dbgsym filses. See the gdb `debug-file-directory` option and the `add-symbol-file` command.