I actually think these fees are an easy problem to solve, but they can’t be solved by academics themselves due to the incentives (i.e. publishing in these journals plays a big role in promotions and possible future grants being awarded).
If funding agencies capped grant publishing fees at say $250 per article, while still requiring open access, that would be the end of these insane page charges. The journals would have to adapt. No one is paying $11,000 out of pocket to publish in these journals…
This indeed! In the US the fed government wastes massive amounts of grant money covering excessive publishing and open access fees. This could easily be restricted to a small max per article that must be accepted by any journal grant funded results will be published in. The journals would adapt.
Governments shouldn’t be paying significant amounts of tax payer money in page fees or open access fees. There should be a small cap on what will be paid, a requirement work is open access and deposited in public databases, and that’s it. It’s outrageous that tax payers are spending $5000 so someone can publish in a Nature type journal, and a waste compared to the additional postdoc and researcher salaries all these page charges could instead go to pay for. Unfortunately, the incentives in academia are such that as long as grants will provide such funds researchers will publish in these journals, so I think this can really only be fixed by funding bodies reducing the money they provide for publications. At the end of the day, I don’t think any journal should be getting more than ~$200-$300 for publishing an article. They simply don’t add much value at this point beyond the prestige of their name.
Most of my biology collaborators have switched in the past 5 years to putting all their new submissions on biorxiv. It seems like medrxiv is now getting popular since COVID started too. But, as in math and physics, I guess a big problem is articles that predate this change are still paywalled in many cases.
Keep in mind, in math and other hard sciences it is not uncommon to cite 50 year old papers which are still highly relevant today (which is less common in biology). Such papers are often behind publisher paywalls since they predate arXiv. So closed access is very much a problem in these fields too.
And this whole discussion is because they’ve announced that will be dropped with version 8. No more standalone licenses, local vaults or syncing via Dropbox/iCloud. Even this (potential) “self-hosted” option sounds ridiculous; users are going to have to run a version of their cloud server locally and use that for syncing instead of just having a vault they can sync locally or through the service of their choice.
How are you supposed to assess whether a student has learned quantum mechanics, or chemistry, or calculus if any questions you ask them they can simply contact another student for help, or get someone to do it for them on sites like Chegg? Short of giving an individualized oral exam, which is not feasible for anything but tiny classes, proctored exams are necessary for assessment in such classes. This has nothing to do with memorization or time pressures, and everything to do with making sure the student is the one who actually did the work. In an ideal world one could trust students to do take-home exams on their own, but I’ve yet to be at a university where undergraduates did not take advantage of such trust…
Then you’re lucky to have students with a lot of integrity at your college/university. Even with a great unique exam where nothing can be looked up online, this doesn’t prevent students from copying each other, posting problems on sites like Chegg and paying for help, or working together…
Julia has some young projects that look promising as future broad FEM/PDE discretization libraries, like Gripap.jl, but they are certainly not at the level of the more mature C++-based libraries you mention.
For my own work I would never say C++ is a nicer choice; I hate writing and dealing with C++ code. It is however a necessary choice since that is where many of best the libraries live.
If you are going to use Julia in one of the absolute worst workflows for how it is designed, you shouldn’t be surprised it doesn’t work well... That said, have you tried using PackageCompiler to add your needed libraries to the system image? This seems to show a factor of 100 speed up for the time to first plot: https://julialang.github.io/PackageCompiler.jl/dev/examples/...
Seconded for macOS. I usually update when the next major release is about to be announced. By 11 months they usually finally have the bugs worked out. It isn’t always necessary for every yearly release, but once you’ve been burned a few times you learn it’s better to wait for several point releases...
I would say that is most likely due to a massively higher salary and no teaching responsibilities. Academia can’t compete on salary with industry in AI / data science.
Industry is not going to fund the overwhelming majority of research areas in biology, physics, chemistry, mathematics, etc. Data science and AI are an exception, where people in industry are much better paid, and can get access to much better resources that would be hard to afford in academia... It’s not surprising this type of advance came from an industry funded group. On the other hand, it is academia and its structure that has enabled so many other discoveries, for example, Crispr DNA tech, our understanding of gravitational waves, or the proof of the Poincaré conjecture.
As an alternative, don't import all of DifferentialEquations, just import the sublibraries you need. For ODEs usually just DiffEqBase and OrdinaryDiffEq are sufficient.
The US Supreme Court ruled in 1980 that faculty at private universities can be classified as managers, and hence denied the right to unionize. This probably isn’t relevant for GATech as it is a public school, but it does mean at many universities faculty are at the relative mercy of the administration.
I’ve never heard of a reviewer writing 15K lines of code. That would be a serious undertaking for someone who is reviewing as an unpaid community service in their free time outside of doing their own research/teaching/etc.
Maybe a very dedicated reviewer might check a short simulation code that can be written in a few hours max, but even that is exceptionally rare in my experience. Reviewers usually look at the results and validation simulations in the manuscript, and if they look reasonable trust the code doesn’t have major bugs affecting accuracy.
I think he was an Associate Professor of the Practice -- it's not a tenured position, but more a longer-term teaching-type appointment. It generally offers more job security and benefits than an adjunct might receive, but at many universities such a position is not (and was never) on a tenure-track line.
Agreed, no meat substitute I've had tastes the same or even good, including Beyond Burgers and such. A good veggie burger, that is not trying to be faux meat, but instead is highlighting the veggies, can be great. And unlike every faux-meat substitute it has the added benefit of not making me sick to my stomach...
When I search Julia questions on Google I always restrict the search to the past year or six months. You are right that there is an enormous amount of irrelevant info out there from earlier versions of the language.
This is something that was just heavily discussed on the Julia Discourse, with a big back and forth between individuals used to the computer integer arithmetic rules vs people wanting Python like functionality.
Based on the confusion there, it definitely seems that it could be made clearer in the docs that if one wants floating point results one needs to use 10.0^17 instead of 10^17.
As a rule of thumb, it is probably a good idea to ignore Julia articles that are more than a year or so old. Anything that is discussing pre-1.0 may be completely irrelevant at this point. The language has evolved a lot, even the past few years, and it is only with the 1.0 release last fall that the devs promised to start maintaining backwards compatibility.
So there are lots of older articles discussing drawbacks to Julia, or ways to do things in Julia, that simply aren’t relevant anymore. Unfortunately they still often are the first links that come up when searching google for Julia questions.
The Julia tooling is not currently designed for repeatedly creating new REPL/interpreter instances to quickly solve one small problem.
What it excels at is rapidly solving the same problem many times within one REPL/interpreter (since after the first time the code will be compiled), or for longer computations where the overhead of the initial compilation does not matter.
It would be great if it could also get fast at the use case you are interested in, but this isn’t currently what it excels at.