It might all be fine again in a year's time, it was fine a year back, but plumping for Ubuntu just as they go their own way with Unity doesn't strike me as the best way of getting a slick Linux. (The same would go for a GNOME 3 distro right now, like Fedora.) I'd have been tempted to put on Linux Mint, Mandriva or openSuSE -- something accessible and slick enough, but likely to have a more stable user experience for a while...
I feel a bit I'm talking to a brick wall here. Probably my fault getting into a more extended conversation than the original point, and not making that clear enough.
Quoting my reply to the original poster, "The point was that *THAT PLOT THAT THEY PUT OUT* is, in contradiction to your statements "To construct the map, the standard assumption that red shift is proportional to distance was made" and "taking redshift and using it as a spatial dimension for a map implies a distance interpretation"."
That was my only point. He'd picked up on "distance" in the press release and then made some incorrect comments about the plot. Thing is that the plot they gave out is just in redshift space, which involves no assumptions, unlike when you employ a model -- the standard one or his (unquantified) alternative -- to convert those redshifts to something else. I'm sorry if I interpreted "to construct the map" in his post as meaning "to construct the map". Maybe I should have interpreted it as "When they say in the press release and on their website about distances that...", but I made the mistake of actually reading what he wrote.
Just to clarify - to construct the map, they did a large galaxy survey. They noted the right ascension and declination of each galaxy. They measured the magnitude in a variety of colours. This then lets you use what are known as "photometric redshifts" to get a reasonable approximation to the actual redshift of the galaxy. It's a phenomenological technique which is used because the prospect of finding spectroscopic redshifts on a good few hundred thousand galaxy clusters is a horrific one that would slow the survey enormously, while photometric redshifts can be found automatically. A certain sample of the redshifts are then checked against spectroscopic redshifts to test the accuracy of the algorithms. That's all you need to make this map: the redshift, the right ascension and the declination. Anything else is model dependent.
About the other point, the one that wasn't my original point but which I seem to have got involved in... Obviously the analysis will be done using a particular model, which is the standard model of cosmology, which attributes the redshift of high-redshift (z>~0.001 or so, or whatever it comes to) objects to the expansion of the universe. If they didn't do this they would have no quantitative results, so they have to use a model. To choose the model, they use the model that is currently the simplest that explains all the data so far, which is the FLRW model, involving an expanding universe. Given this expansion you will get a redshifting of the light. Given a background model -- which is judged from a compilation of this kind of data along with CMB data, the Hubble key project and possibly supernovae 1a data if you choose to include it -- you can then convert these redshifts to distance measures. This is obviously highly dependent on the model, which is dependent on all the assumptions underlying it
And there are *loads* of assumptions underlying the Robertson-Walker model, way more than things about redshifts. Indeed, it's a pretty old-fashioned approach to start deriving cosmology by looking at a Hubble diagram and redshifts -- those are predictions of it, ultimately. These days you tend to use the existence of an isotropic CMB, along with the assumption that the Earth is in no special place int he universe, to motivate the universe being homogeneous and isotropic "on average". If you then assume gravity is metric you're more or less lead to it being described on large scales by the Robertson-Walker metric. If you then assume GR you're lead to the standard model of cosmology. Is it right? No. A lot of observers may even not relise that, but I can say that with certainty. Is it the best model we've got? Yes, currently. Does anything that replaces it have to closely resemble it? Ultimately, yes, at least in observable results.
Ultimately, if someone doesn't like it, the data is publicly available for them to take and run it against their own f
When they interpret the map, they're going to interpret the redshift as a measure of distance. This is because they're interpreting it through the standard (big bang/Robertson-Walker) model of cosmology, which has that the universe expands. This imparts a gravitational redshift on propagating light rays -- basically the wavelength is stretched as the universe expands, which makes sense.
The analysis of the 2MASS data will be done (by the team themselves) within the bounds of Robertson-Walker cosmology. That's perfectly natural; it's the dominant model of cosmology, and it's that for a reason (it's passed most tests we've thrown at it and nothing else can actually fit all the data as well with as few parameters so of course we'll use it). Other people can get hold of all the data and run it through any other model they like -- and I'd encourage them to do so, absolutely! So long as the science isn't wonky, that's fine. For instance, I can guarantee that this data is going to be checked against models where the Earth sits near the centre of a massive void, which absolutely contradicts the first assumption of cosmology, that we're in no unique spot.
So yes, 2MASS will convert those redshifts into a distance measure. The reason they'll also have chopped off low redshifts is to avoid contaminating the data too much with galaxies where the redshift is *definitely* significantly affected by its own motion. That data will still be there, they just haven't presented it in this plot and they won't use it in the analysis. Anyone with an alternative model can keep it in if they want to.
The plot that they put out, though, is in redshift space. That means that it's not converted to distances. That was my only point -- that the plot that accompanies the press release is in redshift space and as such is as unpolluted by model-specific assumptions as possible.
(I also didn't like the bolshy nature of the original post. As a rule, anyone who posts on a forum stating something controversial as absolute truth, the way this guy did, is either trolling, a crank, or harbouring some bitterness against the establishment. Standard models are often talked about like that -- and that was one of his points that I *do* agree with; the standard model is accepted dumbly by far too many people -- but that's because they're received wisdom. The same way people are taught Mendeleven genetics even though it's a gross oversimplification and according to my dad (who's a geneticist) doesn't even work. Because it's the simple received wisdom. If you're saying something controversial and against the received wisdom you absolutely have to provide evidence to back it up... very strong evidence. I'll always be happy to recommend people take data and run it through their own model to support themselves *quantitatively*. They have to be able to explain everything LCDM can. Alas, it never actually seems to happen. And that's a big shame because there is no way on Earth or elsewhere that LCDM is the whole truth. Absolutely no way. It's a phenomenological model with parameters characterising things we don't understand which are derived from an implicit averaging process that has never been shown to work and, thus far, cannot possibly work.)
you wouldn't believe it reading the drivel that the popular press have been writing about it. you'd have thought the words "electron density" would have given it away that we're not talking dark matter here, but no. STUDENT FINDS MISSING MATTER scream the headlines. "ok, fair enough," you think. then the article is filled with things about dark matter. pah.
under the assumption that that's an honest question, no, not really, unless you're seriously going to complain that north is "up". that plot is in galactic coordinates, which are based on the sun's position in the galaxy. http://en.wikipedia.org/wiki/Galactic_coordinate_system
That's why I emphasised "the plot we are talking about", and commented that in the press release they'd interpret it in terms of a simple model for the general public.
"But *that plot that they put out*, I mean *that one we're talking about* is in redshift space."
"in the press release obviously they're going to simplify down for a general audience"
They're going to use the standard model. And like it or fucking lump it, but the standard model interprets those redshifts as evidence for a universal expansion, which can then be converted to distances. What are they going to do, drown a press release with technical details?
The point was that *THAT PLOT THAT THEY PUT OUT* is, in contradiction to your statements "To construct the map, the standard assumption that red shift is proportional to distance was made" and "taking redshift and using it as a spatial dimension for a map implies a distance interpretation". I'm not arguing with "the redshift is proportional to distance dogma is so ingrained in astronomical teaching and literature that authors assume distance to be implied by a mere mention of a redshift measurement" necessarily; I'd say that actually the people doing this kind of work are well aware of the assumptions that underly it, but they're not often made clear to a general audience. What I'm pointing out is blindingly simple. This plot is a 2D plot with redshift information. What you choose to make of that is your own, well, choice. Don't want to interpret the redshifts as a distance measure? Go right ahead, no-one's stopping you. The press release concerned the simplest model, as it's obviously going to do, because that's the standard cosmological model.
Don't like it? Construct an alternative, fit it to data, make predictions, and do it properly. That means getting a model that fits the observed abundances of hydrogen, deuterium, tritium, helium, lithium and the like, that fits the CMB angular power spectrum, that even fits the CMB blackbody spectrum, that fits the matter power spectrum, that fits the baryon acoustic oscillations in the matter power spectrum, and fits the supernova 1a data too. And make it so that given parts of that, it can predict the rest of them, and make it uncontrived. Like it or not (and I'm not that big a fan myself), Lambda CDM isn't actually contrived -- you assume general relativity to be the theory of gravity, and then assume homogeneity and isotropy on large scales and the rest of it falls straight out. The BAOs were predicted, and their angular scale predicted, well before they were observed. The agreement is striking, even if LCDM seems to predict a slightly lower amplitude than we observe. That's not a bad thing, it's a good thing since it means we've still got things to learn. But the angular scale is there... and believe me, it's really quite hard to fit that angular scale if you play with the model much.
Oh I totally agree with you. At the least they could have released a 3D chart with galactic coordinates and redshift as the axes, even if they didn't convert to distances -- or converted to distances and printed the caveats about that conversion being model-dependent. It woudl really help visualisation. I imagine it'll happen soon though; it didn't take long for someone to import the SDSS data (release 1 or release 3, last time I checked, unless it was actually 2dF) into Celestia and someone will do the same with this. Then we can go flying around through it -- and it really helps emphasise the filamentary nature of the cosmos when you do that, too.
What you're doing is interpreting redshift as distance and then attacking that. In the science that will be done on this dataset then, yes, the cosmologists will employ a background model and convert the redshifts to "distances" (if you like) to do an analysis. Of course they will; that's what you have to do. If you want to apply your own pet model to it the data is freely available. Go ahead and knock yourself out. No-one's going to stop you.
But *that plot that they put out*, I mean *that one we're talking about* is in redshift space. What's your argument with that? They've looked at the distribution on the celestial sphere and plotted up everything with redshifts above some cut-off, whether that cut-off is z=0 or z=0.01 or wherever they put it. It's just a plot of redshifts. No assumptions made about anything.
As for this "taking redshift and using it as a spatial dimension for a map implies a distance interpretation or, if you wish, a uni-variate spatial interpretation of redshift", that's simply wrong. I can use it as a dimension if I want. I could convert it to velocities and use that quite happily -- although as you rightly point out that assumes that the entire redshift is given by a velocity. In a standard cosmology at least the vast bulk of the redshift is given by a velocity. You're free to reinterpret the data through your own model, of course. I could even sit there and examine each and every galaxy there and find the colour of it and plot that as a dimension instead. I could convert that colour to an approximate temperature and use that as a third dimension. It's you that's then interpreting that as a distance right now -- and I suspect you're doing that so you can then attack the standard model.
Good on you, the standard model is riddled with holes and I work on attacking some of those holes myself. But don't set up an immediate strawman. That plot is in redshift, and that's a perfectly valid variable to use. Furthermore, it's hte *only* variable they can use that doesn't involve any extra assumptions. As soon as you swap it to *anything else*, whether it's from your own little pet model or from standard Lambda CDM or anything else, it's model-dependent. The redshifts are what are observed, ergo that's what they've put on this plot.
I seriously don't see an issue with that.
For everything else it's pretty irrelevant to the plot, and in the press release obviously they're going to simplify down for a general audience. And as I said, you can feel free to take the raw data, or take the filtered data (where the redshifts have been calculated, doubtless using photometric redshifts which certainly contaminate the data but at a controllable rate given the sheer size of 2MASS which will beat down the statistical error), and do your own analysis. Indeed, please do. Lambda CDM is a startlingly succesful model and has beaten off anything else simply because it *FITS*, but I don't think anyone is seriously sold on it, other than it works. All of us would love something more interesting. Fewer people would love the ditching of a FLRW background entirely, chiefly because they don't have the background necessary to work with something more complicated, but there are those of us that want to drop that assumption too. But you have to be able to fit the data at least as well as Lambda CDM does otherwise you're going to be ignored just as pretty much every other alternative is ignored.
Well, not really. The press release talks about distances (hence that 380m light years bit) but the plot itself is marked up in redshifts. Then they're just colour-coding for those.
that's the view as seen from earth. so we can't be represented on that map at all -- that's what we see surrounding us, except that the celestial sphere has been mapped onto a 2d plane. you can transform that map back and put it on a sphere, and then yes, we'd be in the centre of the sphere.
that map is in redshift space. did you not notice the colour-coding at the bottom is in redshift? they've not converted it to distances at all. the most they might have done is removed objects with redshifts 0.01 or 0.05 or something like that.
try and look at what's going on before you sit there banging some tired old drum.
....it's a tenet of cosmology (which seems more or less justified) that if you look on large enough scales it *is* a homogeneous mixture of galaxies. If it wasn't, our cosmological models would be pure baloney. (The interesting part comes when you try and actually prove that this tangled network of strings and voids and filaments and massive clumps of superclusters averages out to a homogeneous universe, given that we can't define a meaningful average in general relativity. It seems very likely, especially given that the standard cosmological model works so damned well, but there are enough odd features (dark energy, for example) that suggest something slightly odd might be going on with this implicit averaging. Or of course the averaging could work fine and it's actually a scalar field slowly rolling down its potential, or a manifestation of two branes moving slowly closer to one-another, or a reduction in gravity's effects at low curvatures or.....)
Of course there are objects moving towards us, quite a lot of them when you look towards the Virgo supercluster (because we're falling into it). However they tend to be extremely nearby, as far as cosmological distance scales go. The universal expansion is known as the "Hubble flow", and it grows rapidly as you look further and further away from us, and quickly swamps any local motion (known as "peculiar motion", although the terminology is imprecise and sometimes that only refers to the motion as projected onto the celestial sphere, perpendicular to the line of sight).
These very low redshift (and blueshift) objects will be removed -- my guess is that they'll have done a cut and removed everything below z=0.05 or z=0.1 or so. That will remove all the local structure which would otherwise contaminate the cosmological plot. (Yes, you could argue it would also remove any odd outliers that are far away *and* flying towards us, but such weird objects would be flagged and basically there aren't any.)
But I'm not involved in 2MASS and it's literally four or five years since I went to a seminar given by anyone involved, so maybe they're doing something very different - this is all just my guess.
It's a lot more useful to be in redshift space than it would be in "real" space. To swap it to real space you have to make a lot of assumptions about the nature of the universe, and you get out something that may or may not actually correspond to the actual distribution. Redshift space is distorted by other effects, particularly for galaxies nearby where the Hubble flow is weak and peculiar motions strong (the extreme example being the Andromeda galaxy with its strong blue-shift given that it's flying straight at us looking for a barney) but it's still what we actually see.
It's always possible anyway to draw a 3D plot with galactic coordinates and then redshift as your third axis. That's how I'd recommend it.
That plot is absolutely rubbish. I do know what redshift is, I do know the coordinates they're using on the plot, and I even understand why they might want to do a 2D plot (because that's basically we see from Earth; everything is projected onto the celestial sphere; and if you converted to distances you'd be making a lot of model-specific assumptions). But that plot is still absolutely rubbish. Surely they can do better than that? At the very least they could do a 3D plot in redshift-space. That's not hard. Sloan and 2dF did big wedges in redshift space ten years ago.
Also, "a mess of colors on a 2d image then doesn't give us an indication of what color the planets are". Planets? These aren't even *stars* they're looking at, they're entire galaxies. Each individual point will be a galaxy at least as big as our own, or perhaps quite a few galaxies as big as our own.
And can you imagine how messy the plot would be if they gave us some clue of the galaxy motions? It's a horrible mess as it is. If they added a bunch of vectors showing the peculiar velocity (a conversion, by the way, that *also* involves a lot of model-specific assumptions, since the bulk of the galaxy motion is coming from the universal expansion, which has to be subtracted) then it would look like a child's scribble.
In which case either Red Hat evolved significantly from Red Hat 6, or I'm a bigger idiot than I thought...:) (Or our Solaris wasn't running the GNU utilities. Given that I had no clue where anything was I suspect this to be so...)
Being in academia and spending time in a lot of departments I can at least confirm that a large number of departments are running Scientific. I've worked in Britain, the USA, Canada, Norway and Germany and while Germany (predictably enough) has a hankering for SuSE, the others have a tendency to run Scientific.
I did type in a long and boring anecdote about my experiences administering things running SGI Irix and Solaris back in the day, but wiped it when it began to look a bit incriminating and for all I know my ex-boss reads Slashdot. So I'll summarise as "don't administer SGI Irix or Solaris if you can avoid it". I'm no computer scientist, so maybe people who are better at it have no problems, but as a vaguely-competent scientist with an interest in computers but little more (like the original poster) I didn't get on with either of them. Red Hat was fine, and we hung Fedora machines off our central network and that was OK even though it was Fedora Core 1 with all its teething problems. And Scientific is very widely used in academia on big networks.
Yeah I know but I last used Linux much in 2006, and that was Gentoo. I was exaggerating for emphasis or something. I've got an Ubuntu installation dual-booting a Vista machine but I never really use it.
Try playing Hero Quest on a Sinclair Spectrum or C64. About eight minutes to load from tape... and then another five minutes per level, and another five minutes to save your progress.
Slashdot Mirror
Yeah, I've not been on Slashdot long but it's long enough to get out of the habit of even reading the summary, let alone the article...
Ignore me. Reading which version they put on is useful before jumping in and commenting like an idiot.
It might all be fine again in a year's time, it was fine a year back, but plumping for Ubuntu just as they go their own way with Unity doesn't strike me as the best way of getting a slick Linux. (The same would go for a GNOME 3 distro right now, like Fedora.) I'd have been tempted to put on Linux Mint, Mandriva or openSuSE -- something accessible and slick enough, but likely to have a more stable user experience for a while...
"It's not in redshift space in the sense of a spatial dimension"
that's what i've been trying to say, too... they just plotted RA vs declination and then redshift on a colour scale.
I feel a bit I'm talking to a brick wall here. Probably my fault getting into a more extended conversation than the original point, and not making that clear enough.
Quoting my reply to the original poster,
"The point was that *THAT PLOT THAT THEY PUT OUT* is, in contradiction to your statements "To construct the map, the standard assumption that red shift is proportional to distance was made" and "taking redshift and using it as a spatial dimension for a map implies a distance interpretation"."
That was my only point. He'd picked up on "distance" in the press release and then made some incorrect comments about the plot. Thing is that the plot they gave out is just in redshift space, which involves no assumptions, unlike when you employ a model -- the standard one or his (unquantified) alternative -- to convert those redshifts to something else. I'm sorry if I interpreted "to construct the map" in his post as meaning "to construct the map". Maybe I should have interpreted it as "When they say in the press release and on their website about distances that...", but I made the mistake of actually reading what he wrote.
Just to clarify - to construct the map, they did a large galaxy survey. They noted the right ascension and declination of each galaxy. They measured the magnitude in a variety of colours. This then lets you use what are known as "photometric redshifts" to get a reasonable approximation to the actual redshift of the galaxy. It's a phenomenological technique which is used because the prospect of finding spectroscopic redshifts on a good few hundred thousand galaxy clusters is a horrific one that would slow the survey enormously, while photometric redshifts can be found automatically. A certain sample of the redshifts are then checked against spectroscopic redshifts to test the accuracy of the algorithms. That's all you need to make this map: the redshift, the right ascension and the declination. Anything else is model dependent.
About the other point, the one that wasn't my original point but which I seem to have got involved in... Obviously the analysis will be done using a particular model, which is the standard model of cosmology, which attributes the redshift of high-redshift (z>~0.001 or so, or whatever it comes to) objects to the expansion of the universe. If they didn't do this they would have no quantitative results, so they have to use a model. To choose the model, they use the model that is currently the simplest that explains all the data so far, which is the FLRW model, involving an expanding universe. Given this expansion you will get a redshifting of the light. Given a background model -- which is judged from a compilation of this kind of data along with CMB data, the Hubble key project and possibly supernovae 1a data if you choose to include it -- you can then convert these redshifts to distance measures. This is obviously highly dependent on the model, which is dependent on all the assumptions underlying it
And there are *loads* of assumptions underlying the Robertson-Walker model, way more than things about redshifts. Indeed, it's a pretty old-fashioned approach to start deriving cosmology by looking at a Hubble diagram and redshifts -- those are predictions of it, ultimately. These days you tend to use the existence of an isotropic CMB, along with the assumption that the Earth is in no special place int he universe, to motivate the universe being homogeneous and isotropic "on average". If you then assume gravity is metric you're more or less lead to it being described on large scales by the Robertson-Walker metric. If you then assume GR you're lead to the standard model of cosmology. Is it right? No. A lot of observers may even not relise that, but I can say that with certainty. Is it the best model we've got? Yes, currently. Does anything that replaces it have to closely resemble it? Ultimately, yes, at least in observable results.
Ultimately, if someone doesn't like it, the data is publicly available for them to take and run it against their own f
When they interpret the map, they're going to interpret the redshift as a measure of distance. This is because they're interpreting it through the standard (big bang/Robertson-Walker) model of cosmology, which has that the universe expands. This imparts a gravitational redshift on propagating light rays -- basically the wavelength is stretched as the universe expands, which makes sense.
The analysis of the 2MASS data will be done (by the team themselves) within the bounds of Robertson-Walker cosmology. That's perfectly natural; it's the dominant model of cosmology, and it's that for a reason (it's passed most tests we've thrown at it and nothing else can actually fit all the data as well with as few parameters so of course we'll use it). Other people can get hold of all the data and run it through any other model they like -- and I'd encourage them to do so, absolutely! So long as the science isn't wonky, that's fine. For instance, I can guarantee that this data is going to be checked against models where the Earth sits near the centre of a massive void, which absolutely contradicts the first assumption of cosmology, that we're in no unique spot.
So yes, 2MASS will convert those redshifts into a distance measure. The reason they'll also have chopped off low redshifts is to avoid contaminating the data too much with galaxies where the redshift is *definitely* significantly affected by its own motion. That data will still be there, they just haven't presented it in this plot and they won't use it in the analysis. Anyone with an alternative model can keep it in if they want to.
The plot that they put out, though, is in redshift space. That means that it's not converted to distances. That was my only point -- that the plot that accompanies the press release is in redshift space and as such is as unpolluted by model-specific assumptions as possible.
(I also didn't like the bolshy nature of the original post. As a rule, anyone who posts on a forum stating something controversial as absolute truth, the way this guy did, is either trolling, a crank, or harbouring some bitterness against the establishment. Standard models are often talked about like that -- and that was one of his points that I *do* agree with; the standard model is accepted dumbly by far too many people -- but that's because they're received wisdom. The same way people are taught Mendeleven genetics even though it's a gross oversimplification and according to my dad (who's a geneticist) doesn't even work. Because it's the simple received wisdom. If you're saying something controversial and against the received wisdom you absolutely have to provide evidence to back it up... very strong evidence. I'll always be happy to recommend people take data and run it through their own model to support themselves *quantitatively*. They have to be able to explain everything LCDM can. Alas, it never actually seems to happen. And that's a big shame because there is no way on Earth or elsewhere that LCDM is the whole truth. Absolutely no way. It's a phenomenological model with parameters characterising things we don't understand which are derived from an implicit averaging process that has never been shown to work and, thus far, cannot possibly work.)
no. this is not about dark matter. this is about "missing" normal matter.
you wouldn't believe it reading the drivel that the popular press have been writing about it. you'd have thought the words "electron density" would have given it away that we're not talking dark matter here, but no. STUDENT FINDS MISSING MATTER scream the headlines. "ok, fair enough," you think. then the article is filled with things about dark matter. pah.
under the assumption that that's an honest question, no, not really, unless you're seriously going to complain that north is "up". that plot is in galactic coordinates, which are based on the sun's position in the galaxy. http://en.wikipedia.org/wiki/Galactic_coordinate_system
That's why I emphasised "the plot we are talking about", and commented that in the press release they'd interpret it in terms of a simple model for the general public.
"But *that plot that they put out*, I mean *that one we're talking about* is in redshift space."
"in the press release obviously they're going to simplify down for a general audience"
They're going to use the standard model. And like it or fucking lump it, but the standard model interprets those redshifts as evidence for a universal expansion, which can then be converted to distances. What are they going to do, drown a press release with technical details?
The point was that *THAT PLOT THAT THEY PUT OUT* is, in contradiction to your statements "To construct the map, the standard assumption that red shift is proportional to distance was made" and "taking redshift and using it as a spatial dimension for a map implies a distance interpretation". I'm not arguing with "the redshift is proportional to distance dogma is so ingrained in astronomical teaching and literature that authors assume distance to be implied by a mere mention of a redshift measurement" necessarily; I'd say that actually the people doing this kind of work are well aware of the assumptions that underly it, but they're not often made clear to a general audience. What I'm pointing out is blindingly simple. This plot is a 2D plot with redshift information. What you choose to make of that is your own, well, choice. Don't want to interpret the redshifts as a distance measure? Go right ahead, no-one's stopping you. The press release concerned the simplest model, as it's obviously going to do, because that's the standard cosmological model.
Don't like it? Construct an alternative, fit it to data, make predictions, and do it properly. That means getting a model that fits the observed abundances of hydrogen, deuterium, tritium, helium, lithium and the like, that fits the CMB angular power spectrum, that even fits the CMB blackbody spectrum, that fits the matter power spectrum, that fits the baryon acoustic oscillations in the matter power spectrum, and fits the supernova 1a data too. And make it so that given parts of that, it can predict the rest of them, and make it uncontrived. Like it or not (and I'm not that big a fan myself), Lambda CDM isn't actually contrived -- you assume general relativity to be the theory of gravity, and then assume homogeneity and isotropy on large scales and the rest of it falls straight out. The BAOs were predicted, and their angular scale predicted, well before they were observed. The agreement is striking, even if LCDM seems to predict a slightly lower amplitude than we observe. That's not a bad thing, it's a good thing since it means we've still got things to learn. But the angular scale is there... and believe me, it's really quite hard to fit that angular scale if you play with the model much.
Oh I totally agree with you. At the least they could have released a 3D chart with galactic coordinates and redshift as the axes, even if they didn't convert to distances -- or converted to distances and printed the caveats about that conversion being model-dependent. It woudl really help visualisation. I imagine it'll happen soon though; it didn't take long for someone to import the SDSS data (release 1 or release 3, last time I checked, unless it was actually 2dF) into Celestia and someone will do the same with this. Then we can go flying around through it -- and it really helps emphasise the filamentary nature of the cosmos when you do that, too.
Err, thanks.
What you're doing is interpreting redshift as distance and then attacking that. In the science that will be done on this dataset then, yes, the cosmologists will employ a background model and convert the redshifts to "distances" (if you like) to do an analysis. Of course they will; that's what you have to do. If you want to apply your own pet model to it the data is freely available. Go ahead and knock yourself out. No-one's going to stop you.
But *that plot that they put out*, I mean *that one we're talking about* is in redshift space. What's your argument with that? They've looked at the distribution on the celestial sphere and plotted up everything with redshifts above some cut-off, whether that cut-off is z=0 or z=0.01 or wherever they put it. It's just a plot of redshifts. No assumptions made about anything.
As for this "taking redshift and using it as a spatial dimension for a map implies a distance interpretation or, if you wish, a uni-variate spatial interpretation of redshift", that's simply wrong. I can use it as a dimension if I want. I could convert it to velocities and use that quite happily -- although as you rightly point out that assumes that the entire redshift is given by a velocity. In a standard cosmology at least the vast bulk of the redshift is given by a velocity. You're free to reinterpret the data through your own model, of course. I could even sit there and examine each and every galaxy there and find the colour of it and plot that as a dimension instead. I could convert that colour to an approximate temperature and use that as a third dimension. It's you that's then interpreting that as a distance right now -- and I suspect you're doing that so you can then attack the standard model.
Good on you, the standard model is riddled with holes and I work on attacking some of those holes myself. But don't set up an immediate strawman. That plot is in redshift, and that's a perfectly valid variable to use. Furthermore, it's hte *only* variable they can use that doesn't involve any extra assumptions. As soon as you swap it to *anything else*, whether it's from your own little pet model or from standard Lambda CDM or anything else, it's model-dependent. The redshifts are what are observed, ergo that's what they've put on this plot.
I seriously don't see an issue with that.
For everything else it's pretty irrelevant to the plot, and in the press release obviously they're going to simplify down for a general audience. And as I said, you can feel free to take the raw data, or take the filtered data (where the redshifts have been calculated, doubtless using photometric redshifts which certainly contaminate the data but at a controllable rate given the sheer size of 2MASS which will beat down the statistical error), and do your own analysis. Indeed, please do. Lambda CDM is a startlingly succesful model and has beaten off anything else simply because it *FITS*, but I don't think anyone is seriously sold on it, other than it works. All of us would love something more interesting. Fewer people would love the ditching of a FLRW background entirely, chiefly because they don't have the background necessary to work with something more complicated, but there are those of us that want to drop that assumption too. But you have to be able to fit the data at least as well as Lambda CDM does otherwise you're going to be ignored just as pretty much every other alternative is ignored.
Well, not really. The press release talks about distances (hence that 380m light years bit) but the plot itself is marked up in redshifts. Then they're just colour-coding for those.
that's the view as seen from earth. so we can't be represented on that map at all -- that's what we see surrounding us, except that the celestial sphere has been mapped onto a 2d plane. you can transform that map back and put it on a sphere, and then yes, we'd be in the centre of the sphere.
..... no.
that map is in redshift space. did you not notice the colour-coding at the bottom is in redshift? they've not converted it to distances at all. the most they might have done is removed objects with redshifts 0.01 or 0.05 or something like that.
try and look at what's going on before you sit there banging some tired old drum.
that's the effect of this mess: http://apod.nasa.gov/apod/ap100709.html except at a different wavelength
....it's a tenet of cosmology (which seems more or less justified) that if you look on large enough scales it *is* a homogeneous mixture of galaxies. If it wasn't, our cosmological models would be pure baloney. (The interesting part comes when you try and actually prove that this tangled network of strings and voids and filaments and massive clumps of superclusters averages out to a homogeneous universe, given that we can't define a meaningful average in general relativity. It seems very likely, especially given that the standard cosmological model works so damned well, but there are enough odd features (dark energy, for example) that suggest something slightly odd might be going on with this implicit averaging. Or of course the averaging could work fine and it's actually a scalar field slowly rolling down its potential, or a manifestation of two branes moving slowly closer to one-another, or a reduction in gravity's effects at low curvatures or.....)
Of course there are objects moving towards us, quite a lot of them when you look towards the Virgo supercluster (because we're falling into it). However they tend to be extremely nearby, as far as cosmological distance scales go. The universal expansion is known as the "Hubble flow", and it grows rapidly as you look further and further away from us, and quickly swamps any local motion (known as "peculiar motion", although the terminology is imprecise and sometimes that only refers to the motion as projected onto the celestial sphere, perpendicular to the line of sight).
These very low redshift (and blueshift) objects will be removed -- my guess is that they'll have done a cut and removed everything below z=0.05 or z=0.1 or so. That will remove all the local structure which would otherwise contaminate the cosmological plot. (Yes, you could argue it would also remove any odd outliers that are far away *and* flying towards us, but such weird objects would be flagged and basically there aren't any.)
But I'm not involved in 2MASS and it's literally four or five years since I went to a seminar given by anyone involved, so maybe they're doing something very different - this is all just my guess.
It's a lot more useful to be in redshift space than it would be in "real" space. To swap it to real space you have to make a lot of assumptions about the nature of the universe, and you get out something that may or may not actually correspond to the actual distribution. Redshift space is distorted by other effects, particularly for galaxies nearby where the Hubble flow is weak and peculiar motions strong (the extreme example being the Andromeda galaxy with its strong blue-shift given that it's flying straight at us looking for a barney) but it's still what we actually see.
It's always possible anyway to draw a 3D plot with galactic coordinates and then redshift as your third axis. That's how I'd recommend it.
That plot is absolutely rubbish. I do know what redshift is, I do know the coordinates they're using on the plot, and I even understand why they might want to do a 2D plot (because that's basically we see from Earth; everything is projected onto the celestial sphere; and if you converted to distances you'd be making a lot of model-specific assumptions). But that plot is still absolutely rubbish. Surely they can do better than that? At the very least they could do a 3D plot in redshift-space. That's not hard. Sloan and 2dF did big wedges in redshift space ten years ago.
Also, "a mess of colors on a 2d image then doesn't give us an indication of what color the planets are". Planets? These aren't even *stars* they're looking at, they're entire galaxies. Each individual point will be a galaxy at least as big as our own, or perhaps quite a few galaxies as big as our own.
And can you imagine how messy the plot would be if they gave us some clue of the galaxy motions? It's a horrible mess as it is. If they added a bunch of vectors showing the peculiar velocity (a conversion, by the way, that *also* involves a lot of model-specific assumptions, since the bulk of the galaxy motion is coming from the universal expansion, which has to be subtracted) then it would look like a child's scribble.
In which case either Red Hat evolved significantly from Red Hat 6, or I'm a bigger idiot than I thought... :) (Or our Solaris wasn't running the GNU utilities. Given that I had no clue where anything was I suspect this to be so...)
Being in academia and spending time in a lot of departments I can at least confirm that a large number of departments are running Scientific. I've worked in Britain, the USA, Canada, Norway and Germany and while Germany (predictably enough) has a hankering for SuSE, the others have a tendency to run Scientific.
I did type in a long and boring anecdote about my experiences administering things running SGI Irix and Solaris back in the day, but wiped it when it began to look a bit incriminating and for all I know my ex-boss reads Slashdot. So I'll summarise as "don't administer SGI Irix or Solaris if you can avoid it". I'm no computer scientist, so maybe people who are better at it have no problems, but as a vaguely-competent scientist with an interest in computers but little more (like the original poster) I didn't get on with either of them. Red Hat was fine, and we hung Fedora machines off our central network and that was OK even though it was Fedora Core 1 with all its teething problems. And Scientific is very widely used in academia on big networks.
Yeah I know but I last used Linux much in 2006, and that was Gentoo. I was exaggerating for emphasis or something. I've got an Ubuntu installation dual-booting a Vista machine but I never really use it.
Try playing Hero Quest on a Sinclair Spectrum or C64. About eight minutes to load from tape... and then another five minutes per level, and another five minutes to save your progress.