cbkeller

If that’s your workflow it would probably be highly advisable to compile plots.jl into your sysimg

Totally fair, though note that these slow times are only for _first_ plot - I leave my repl open all day, and all but the first plot are sub-second

I’m always amused to see these “I’m sure caching wouldn’t even be a hard fix, the community must be resisting it” takes.

Feel free to show me the PR’s that have have been rejected that would have solved the problem. Or if you think it’s easy, feel free to make that PR yourself.

As the quote goes, “There are two hard problems in programming: cache invalidation, naming things, and off-by-one errors”. This is cache invalidation.

I’ve never tried it myself, but can’t think of any reason why it wouldn’t work in principle if you do want to go that route

Technically yes, but then IIRC you’d have to recompile the whole sysimg any time you update any single package, which could get to be a pain. The usual compromise seems to be to include just a handful of your most used packages in the sysimg (say Plots + Revise)

It’s not trivial to do, but compiling plots into your sysimg might actually fix that

Yeah, this is probably actually the closest equivalent to the Rust Book, but I’ll also second the suggestion from the other comment of “Think Julia” (beginner) and “Design Patterns and Best Practices with Julia” (intermediate /advanced).

For me, the most important thing to grasp when coming from another language was that Julia’s multiple dispatch brings with it effectively a whole paradigm of “dispatch-centric programming” that you have to embrace to really get the most out of Julia, including the c-like speed (have to strictly avoid type-instability for that) and the composability that everyone talks about.

Linear algebra is the most practically consequential one - and suffice it to say, it’s not a coincidence that BLAS, LINPACK, and LAPACK were all written in one-based languages. Matrix indexing is one-based, and beyond the issue of translating notation from equations, IIRC there is even a slight performance difference in some cases.

I believe what it really comes down to is that, just as zero-based really is more natural for offsets (zero offset means the first position, great for pointers), one based really is more natural for counting [and _sets_] (when your index is n that means you have counted n elements up until now).

I use this to format all the assignment handouts for my intro-to-programming-in-Earth-sciences class. Has worked great so far.

That hidden Julia s-expression syntax helped me start to understand metaprogramming in Julia. Along the way I found that it takes surprisingly few lines of code to turn the normal Julia repl into a more-or-less working Julia-lisp repl: https://gist.github.com/brenhinkeller/44051118c2f9d18b26dc76...

Vera Rubin would approve. Despite making a lot of the critical rotation rate measurements on spiral galaxies [1,2,3], it seems as though she was never particularly happy with the interpretation of this result as dark matter:

> "If I could have my pick, I would like to learn that Newton's laws must be modified in order to correctly describe gravitational interactions at large distances. That's more appealing than a universe filled with a new kind of sub-nuclear particle." [4]

[1] http://articles.adsabs.harvard.edu/full/1970ApJ...159..379R/...

[2] http://adsabs.harvard.edu/full/1980ApJ...238..471R7

[3] https://www.annualreviews.org/doi/abs/10.1146/annurev.astro....

[4] https://scientificwomen.net/women/rubin-vera-86

There's still a working donation-supported command-line-focused version of it here [1] for anyone who doesn't know!

[1] https://github.com/timvisee/ffsend

That is what I would have interpreted that to mean as well, but the numbers in the far right column don't really appear to be inflation-adjusted relative to the actual primary sources.

Here, for instance, is the actual appropriations bill for 1998 [1]. As you can see,

> "There are authorized to be appropriated to the Foundation $3,505,630,000 for fiscal year 1998."

For comparison, the value in the aforementioned table for FY1998 is $3,430.63 M. This is a bit off, presumably because in contrast to the bill the values in the table are " after supplemental appropriations, transfers, and reprogrammings".

I had missed, though, that the appropriations for the last two years were actually higher than the request, so some signs of recent growth.

[1] https://www.congress.gov/bill/105th-congress/house-bill/1273...

The NSF budget increased dramatically between 1951-2008 -- but when accounting for inflation in the figures you linked, has declined by about 10% in real terms since then.

(That is, the 2010 NSF budget of $7.045 B in 2010 is in real terms equivalent to about $8.413 B today, about a billion dollars more than the real 2020 budget of $7.066 B, using the inflation calculator at [1].)

[1] https://www.in2013dollars.com

>AWS makes you into an owner and that includes owning your career.

This sort of corporate jargon does not exactly instill confidence. I think I'm more concerned about Amazon's engineering culture now than I was before.

That's fair enough from a political perspective (after all, as they say in the article, >70% of NIF shots are going to [nuclear] stockpile stewardship now) -- but just to give NIF a bit of credit on the "basic science" front, it has turned out to be quite good for studying the properties of matter under extreme compression more generally.

Some mineral physicists I know have used this to write some cool papers about the properties of natural materials at conditions replicating those expected in the cores of giant exoplanets [1-3], for example.

[1] https://www.nature.com/articles/s41550-018-0437-9

[2] https://doi.org/10.3389/feart.2019.00023

[3] https://advances.sciencemag.org/content/4/4/eaao5864

Well I think the catch is that, to use the US as an example, "big science" projects such as the NIF, particle accelerators, synchrotron light sources, and supercomputers are generally all funded by the DOE, while small individual research labs are generally funded by NSF or (if biomedical) NIH.

Critically, the only reason that DOE can support the cost of these "big science" projects is that they are generally "dual use" so to speak, with both basic science and national security applications. A similar dynamic is arguably at play in the funding of the largest projects in other countries as well.

So it is not clear that cutting such "big science" projects would in practice actually open up more funding for small basic science grants, lacking the same national security applications.

This is a pet peeve of mine as well; significant figures can be handy, but fundamentally they originated as a quick and dirty method of keeping track of uncertainty that was quite handy when doing calculations by hand, but is more or less obsolete today any time you have an actual quantitative uncertainty bound.

Sure, it's still convenient when writing things out in text to not continue adding digits far beyond the uncertainty of your value, but even so there is nothing "conservative" or in any way accurate about chopping off a bunch of digits (however imprecise) and replacing them with zeros. This is especially clear if we consider that we generally only ever consider sig figs in base 10, but in any computational context that number is definitely not stored in base 10.

Here are the current grad student salaries at UC Davis [1].

A "Teaching Assistant" (AKA a grad student is TAing) is paid $45,138.00 per year, and a grad student who is currently paid out of grants earns between $42,729.00 for "Graduate Student Researcher - Step I" and $83,727.00 for "Graduate Student Researcher - Step X". I was slightly shocked, but this is the official salary scale. Granted, very few grad students are likely compensated at the upper end of that range, but still.

UC Berkeley [2] and UC San Diego [3] appear to be quite similar, so this is would seem to be somewhat representative of the UC system overall.

My own institution would seem to be actually lagging behind somewhat compared to this.

In any case, graduate student compensation has increased significantly in absolute terms over the past decade or two.

[1] https://grad.ucdavis.edu/resources/student-employment/salary...

[2] https://www.ucop.edu/academic-personnel-programs/_files/1819...

[3] https://grad.ucsd.edu/financial/employment/student-pay-rates...

You're definitely right about it being structured to assume you followed the traditional path. That said, top schools will pay grad students about the US median salary ($35k/yr). Other schools potentially much less. Whether the US median salary is itself a starvation wage is, I suppose, another question.

By noun, I mean as opposed to phrase; i.e., unless I'm missing something this does not seem to touch particularly on why "man of science" [/ "men of science"] was not thought to be acceptable, given that it was apparently used about twice as frequently as "natural philosopher" in the surviving books of the time.

I also just noticed that in the 1831 letter, Whewell seems to have been sufficiently impressed by Somerville's book as to write her a poem about it. Such things were presumably more common at the time given that this was the Romantic era, but nonetheless would seem to sugesst that it made an impression on him.

Well, this is where it seems to get complicated.

There is an 1831 letter from Whewell to Somervilles' father, asking him to thank her for the advance copy of her book which she apparently sent him, and also on behalf of the Philosophical Society for the copy she sent them [1] -- so they would have at the very least been aware of her and her book by the time of the 1833 discussion.

It is definitely not clear to me that the assertion in the article is correct, but one thing they do appear to be correct about is that "man of science" appears to have attained substantial currency (or indeed dominance) as an alternative to "natural philosopher" by 1830 [2,3]. Why not keep using that? Perhaps just because it is unwieldy? On the other hand, this was 1830 -- people were not so adverse to verbosity back then.

In any case, the deeper questions of why the profession was thought to need a noun at all, and of Whewell's internal motivations, seem to me still rather unanswered.

[1] https://archive.org/details/b21960239/page/170/mode/2up?q=wh...

[2] https://plato.stanford.edu/entries/whewell/

[3] https://books.google.com/ngrams/graph?content=man+of+science...

Yeah, this is consistent with what I have seen.

At the end of the day, unemployment numbers for holders of PhDs [1] are starkly different from those of the general population.

[1] https://www.nsf.gov/statistics/infbrief/nsf14310/

Academic here -- this brings up something that I always like to spread the word about: Contrary to some apparently common perceptions, grad school, especially in the sciences, generally pays you rather than the other way around.

Things like professional masters degrees (incl. some engineering degrees) are likely to cost money, but in the sciences both MS and PhD should be fully funded at reputable universities -- either through TAing, your advisers grants, or if you're lucky, fellowships.

I suppose it depends whether we want to give more weight to the first use in conversation (1833 or before), or the first use in writing (the 1834 article).

The context of the first use in writing is relevant in either case.

I know.

The point is that you seem to have left out some rather relevant context. If that was an accident, no problem.

Some of your other comments in this thread indicate that it may be a good idea for you to review the hn guidelines

> Please don't use Hacker News for political or ideological battle. It tramples curiosity.

> Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that."

https://news.ycombinator.com/newsguidelines.html

Here is the full text [1] of Whewell's review of this book, by the way, in which the term scientist appears to have first appeared in print [2,3]

[1] https://blogs.kent.ac.uk/sciencecomma/files/2015/04/whewell_...

[2] https://blogs.kent.ac.uk/sciencecomma/2015/04/09/whewell-and...

[3] https://www.oed.com/view/Entry/172698 (note: paywall, but generally the authoritative source)

The word does indeed seem to be attributable to Whewell, with the OED attributing the first printed use of the word to an 1834 article by Whewell in the Quarterly Review, in which the discussion you allude to regarding "philosophy" being too broad of a term, appears on page 59.

However, you seem to have neglected to mention that this 1934 article by Whewell was a review of Mary Somerville's book (On the Connexion of the Physical Sciences).

The full text of Whewell's 1834 article appears to be available [1] from a University of Kent blog post on the topic [2]

[1] https://blogs.kent.ac.uk/sciencecomma/files/2015/04/whewell_...

[2] https://blogs.kent.ac.uk/sciencecomma/2015/04/09/whewell-and...

Her most famous book On the Connexion of the Physical Sciences, which is mentioned in the article, is available in full on Archive.org and looks to be a rather interesting read [1].

[1] https://archive.org/details/onconnexionphys01somegoog

Haven't found anything yet myself, but would love to know -- that looks quite interesting!