Slashdot Mirror
Newly Discovered Star Has an Almost Pure Oxygen Atmosphere (popularmechanics.com)
William Herkewitz, reports for Popular Mechanics: A newly discovered star is unlike any ever found. With an outermost layer of 99.9 percent pure oxygen, its atmosphere is the most oxygen-rich in the known universe. Heck, it makes Earth's meager 21 percent look downright suffocating. The strange stellar oddity is a radically new type of white dwarf star, and was discovered by a team of Brazilian astronomers led by Kepler de Souza Oliveira at the Federal University of Rio Grande do Sul in Brazil. The star is unique in the known pool of 32,000 white dwarf stars, and is the only known star of any kind with an almost pure oxygen atmosphere. The new white dwarf has a mouthful of a name -- SDSSJ124043.01+671034.68 -- but has been nicknamed 'Dox' (pronounced Dee-Awks) by Kepler's team. The discovery was reported today in a paper in the journal Science.
No smoking on that planet...
If you want news from today, you have to come back tomorrow.
Who the hell thought this "slashvertisement" thing would be funny?
No worries - he'll only visit at night (cue sad trombone noise...)
Quo usque tandem abutere, Nimbus, patientia nostra?
I wouldn't buy it. It is in a bad neighborhood. Neighbors regularly have supernova parties.
It's a real article, as are all the others on the front page. Get your panties out of a bunch. What do need to stop doing exactly? Posting articles on April 1?
Lighten up Francis.
Better yet, pop 20 Xanax and sleep the rest of your life out.
Apparently (and I get this from reading the article), it's not just the surface that's 99% oxygen, it's the whole star that is mainly oxygen.
Also the crazy thing is how he found it.......he had data from 300,000 stars printed out, on 300,000 pages, and just started reading through them one by one to see if there was anything interesting in the data.
"First they came for the slanderers and i said nothing."
It is a pleasant change from most sites, where the jokes are things people assume are bugs.
That's the Brazilian way to make science. By the way, Kepler de Souza Oliveira sounds like a very appropriate name for a Brazilian astronomer.
Linux is for people who don't mind RTFM.
Ok see ya
whiplash, sometimes the best response to criticism is no response at all. Besides, bitching about April Fool's stories is every bit as much of a Slashdot tradition as the April Fool's stories themselves. It just isn't the same without them. I come just for the schadenfreude.
I actually think the binary user numbers are cool
That just seems weird (the paper thing). I mean, seriously, has nobody just set loose a program to flag every observed star based on its statistical departure from the norm? I mean sure, you'd might need to go through a few iterations to get the data clustered well to spot the outliers falling between clusters, but still. That seems like the sort of thing that should be done on day one of looking for interesting things.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
You're one of those 10 kinds of people
My understanding is that, actually no - they're not so good at that. We humans still excel at pattern recognition.
Now, this is a layman's understanding (though I understand it a bit better than most laymen, I should assume) so take it at face value and with that caveat.
But, we're just now getting to the point where computers are approaching child-like recognition. Remember the recent outrage when black people were recognized as gorillas? How they have issues with Asian eyes not being open?
See, as a kid, we can point out a tree to you and, sure enough, you then will recognize other trees - even if they are deciduous and the tree you were shown was coniferous. You'll recognize a ball even if it is a baseball, billiard ball, basketball, etc... You'll see money and know that it is money - you'll probably even know it's money if it's from a county you've never seen before.
Computers, well... They're good at filtering things we know about and how to describe. That's it, really. We're still working on machine learning (sometimes called AI) and it's only in the infant stages. Humans, for now, will spot flaws much better (and quicker) than a computer will unless there's something we know to look for.
As this was certainly not something we'd know to look for, we'd not have been able to (yet) program it to spot such anomalies. Now that we've spotted it, know some characteristics, and can do some deterministic interpretations of the data then we might be able to program spotting this in the future with varied levels of accuracy. For the time being, we're still going to need a human in the loop.
I'm not sure if that makes sense or not but that's the best way I can think of to describe it.
"So long and thanks for all the fish."
Now that's what I like to see. I seem to recall that my very first post to you, on your very first day, was mentioning the need to say things pretty much just like that.
You can't please 'em all and it's futile to try. There is, literally, not one thing you can do that will make everyone happy. Sometimes, you just gotta tell 'em to pound sand. Which, well, you just did.
I, for one, appreciate our new realistic and mostly down-to-Earth overlords. I also second your sentiment but I do worry that we're having a bad influence on you. ;-)
"You were such a nice young man, until you started hanging around *that* crowd." Say the old ladies who amass around the village well. "Well, I'll never let MY daughter anywhere near him." Exclaims one of them, as the rest titter and nod in affirmation.
"So long and thanks for all the fish."
Glad to see you're actually doing to Slashdot what needs to be done. So far I've liked the changes.
No problem, credit due where it's deserved. I also posted a submission follow-up on that oil scandal from Wed (https://slashdot.org/submission/5731939/monaco-based-oil-company-unaoil-raided-by-police). I really like that Slashdot is the only news agency in North America that is reporting on this.
But the patterns aren't random. Most stars spectra will will fall into one of a handful of standard classifications, and all those should be immediately removed from the potential "interesting" set as they are only relevant for determining the standard deviation within their cluster for purposes of deciding just how strange other stars are.
If you're specifically looking for interesting compositions, you could do things like categorizing the dominant elements in a star based upon it's emission lines, and then look for anything with an abnormal composition. This isn't rocket surgery, I've got little background in data analysis, and even I can feel the shape of the software I'd need to write to find odd-composition stars, might take me an afternoon or two without using any special tools. And that's before even considering things like the the mathematical packages designed specifically to perform clustering of N-dimensional data sets.
Now granted, there's still a lot of weirdness a human might spot that I wouldn't know how to begin to program for, but an oxygen star? That should have been flagged as unusual within minutes of recording its spectrum.
--- Most topics have many sides worth arguing, allow me to take one opposite you.