Jump to content

Wikipedia:Reference desk/Archives/Computing/2013 May 31

From Wikipedia, the free encyclopedia
Computing desk
< May 30 << Apr | May | Jun >> June 1 >
Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.


May 31

[edit]

What does a Linux distribution mean?

[edit]

If each Linux distribution is a collection of packages, and these packages can be changed, what does it mean that you have this or that distribution? Why do you need distributions at all? Couldn't you just download from a repository Linux kernel, a set of GNU libraries, X Windows System and choose from a list of compatible libraries? OsmanRF34 (talk) 15:10, 31 May 2013 (UTC)[reply]

A Linux distro is not just a collection of packages, it is a system for managing, distributing, installing, and maintaining a collection of packages. Looie496 (talk) 15:15, 31 May 2013 (UTC)[reply]
Furthermore, in actual practice it's nearly impossible to set up a comprehensive list of compatible libraries and applications if you try to do it by hand. Looie496 (talk) 15:17, 31 May 2013 (UTC)[reply]
There is much more to a "Linux distribution" bundle than just the kernel - even more than Kernel + Gnu libraries + X. When I was using the OpenSuSE distro, there were 1600 programs installed in addition to the kernel, etc. The people putting the distro together make a lot of choices on you behalf - for example - which file manager should be the default? Which paint program? Which volume control/mixer tool? Which USB driver? They also have to get a harmonious set of those things that will all run with the same revision of the libraries they chose. Things like the way you pick preferences and install software varies dramatically from one distro to the next.
With an operating system like Windows - there is basically one single set of things that Microsoft pick for you - and that's that. If you buy a new computer with windows pre-installed, you'll probably get a handful of other programs (mostly junk or demo packages probably) pre-installed by the computer manufacturer. So in effect, each manufacturer produces their own "Windows distro". But their choices are typically rather limited. With Linux distro's, the difference from one to another can be dramatic! That's why personal preferences between the various distro's produces such wide-spread debate amongst Linux users - where Windows people hardly ever debate the relative merits of the Dell distro versus the Gateway distro for that OS.
SteveBaker (talk) 16:12, 31 May 2013 (UTC)[reply]

Aesthetics of Windows 8

[edit]

This might come across as soapboxing, but it is intended as a question. When I look at the "Metro" interface, my reaction is that it's ugly -- square icons with dull flat colors, it just looks crude and ungainly to me (in stark contrast to the beautiful interface of Windows 7). I've read a number of criticisms of Windows 8, but to my surprise I don't recall any that criticize the look of it. Am I simply out of touch with modern taste, or have I missed a set of reviews that express a similar reaction? Looie496 (talk) 17:39, 31 May 2013 (UTC)[reply]

It's not really meant to be pretty, but functional. Sometimes you have to sacrifice one to improve the other. I personally like it, but that's maybe just because I was already tired of the old Windows 7, and longed for something new. OsmanRF34 (talk) 17:46, 31 May 2013 (UTC)[reply]
I actually think it looks pretty good. And it does have a modern feel to it. I think there's two possibilities: either people generally aren't bothered by the design, or the other problems they have with Metro are so overwhelming that the looks are a mere afterthought! --SubSeven (talk) 18:15, 31 May 2013 (UTC)[reply]

There's some articles out there about Window's Metro (design language) that are negative but it does seem like it's been well received. E.g. (http://www.zdnet.com/blog/open-source/five-reasons-why-windows-8-will-be-dead-on-arrival/10275). There seems to be a current trend of backlash against so called Skeuomorphic design made popular by Apple (Think notepad app being exactly like a notable, calendar looking like a leather-bound desk calendar etc.). Personally I like that sort of design - it gives a clear hint at how it should work/be used, though I appreciate it can be jarring to some users and (theoretically) may hinder more efficient functionality...but form and function are always a trade-off :-) ny156uk (talk) 18:21, 31 May 2013 (UTC)[reply]

Here's a couple of instances where Metro was reviewed negatively: here and here. The second reviewer liked the Metro interface for mobile devices but felt it to be a bad match for a desktop setting. --Yellow1996 (talk) 01:14, 1 June 2013 (UTC)[reply]
It does look a bit like the original windows before application windows were allowed to overlap ;-) I'd have thought that nowadays the skeuomorph model would be Minecraft with 3d blocks instead of the 2d ones of Windows 8. That's where children nowadays learn about things like the tides or where meat comes from rather than the real world. Dmcq (talk) 10:05, 2 June 2013 (UTC)[reply]

Software calculating the astrological position of planets

[edit]

Way back in the good old times there was a software (ephem, if I remember correctly) to calculate the positions of the (astrological) planets within the signs of the zodiac. This computer is no longer working, and I can't find ephem anymore on the net. There is xephem instead, but it won't compile on my machine (looks like a problem with prehistoric motif libraries). I managed to compile xephemb, which seems to be some kind of batch-version of xephem. Now the output looks if it was designed rather for some hobby astronomers telling at which time they have to look in what direction to spot astronomical object. I don't mind writing a little code myself to convert the output to something that looks like an astrological chart. One output was marked "RA" and i figured out this would be Right Ascension. So I recomputed the 24-hours circle of the RA to the traditional 360 degrees and for those charts I can remember it looks not too far away. But, checking closely against some ephemeris on the internet, the program is off by a few degrees (I specially checked the slow moving planets to exclude differences from local time and the like.) Now I'm stuck. I remember "back in the good old times" I did check against the printed ephemeris and the difference was less than a minute (1/60hth of a degree).

So what could cause these differences? Is there anything wrong with simply rescaling the RA? Are there any (free) programs out there to do better? Any algorithms including data so I could write a complete program by myself (as catastrophic fallback)? 95.112.182.165 (talk) 21:53, 31 May 2013 (UTC)[reply]

Just make the positions up, like you would with the rest. :) 180.148.3.148 (talk) 23:56, 31 May 2013 (UTC)[reply]
Nah, that would be too obvious. The positions are the only part that can be checked easily. It's not like global climate statistics. 93.132.128.169 (talk) 10:21, 1 June 2013 (UTC)[reply]
XEphem uses VSOP87 data, and his data dump was last updated in 1997. (These data don't change much!) The code is documented with mathematical details about where the author truncated series. The errors should be insignificantly small (documented in the code as ~one part in ten million). Are you sure you're doing your conversion and post-processing correctly? Nimur (talk) 00:13, 1 June 2013 (UTC)[reply]
I'm calculating the positions twofold, once from HH:MM:SS to DEG:mm::ss using integers only, to avoid rounding errors, then second way (to check if I didn't mix minutes of hours and minutes of degrees or something like that) I calculate a float in 0..1 and calculate degrees from that, and those agree with each other. So I'm pretty sure it's not a plain calculation error on my part. I'm not quite sure about the RA. Equinox moves during epochs, and I'm not sure if and how different ephemeris take this into account.
I don't see any VSOP-data, but I have to pipe "Saturn,P" into the program to get the position for saturn 93.132.128.169 (talk) 10:21, 1 June 2013 (UTC)[reply]