Jump to content

Wikipedia:Reference desk/Archives/Computing/2013 February 14

From Wikipedia, the free encyclopedia
Computing desk
< February 13 << Jan | February | Mar >> February 15 >
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.


February 14

[edit]
[edit]

The logo is FAKE! It cannot be found on windows 1.x/2.x boxes, nor it is available inside the windows 1.x/2.x versions themselves. Moreover, even "Microsoft" word was written there in another way! Here is what the logo page at BrandsOfTheWorld says: Designer: unkown Contributor: unknown Updated on: Mon, 09/15/2008 - 08:32 Just wondering, how could this logo made by someone just for lulz about 4-5 years ago come now an official logo for 25-years-old product... — Preceding unsigned comment added by 212.34.248.162 (talk) 05:57, 14 February 2013 (UTC)[reply]

If you mean File:Windows logo and watermark - 1985.svg then yes it does look iffy. I'll put a tag on it requesting deletion if there isn't some citation showing it isn't wrong. Dmcq (talk) 12:57, 14 February 2013 (UTC)[reply]
I looked in to this before anyone commented and it seems to be a problematic case. Yes it wasn't on any boxes or in Windows itself that I can find. Nor on any advertising that I found (but to be fair, it was largely a couple of lame videos and some magazine adverts). And the actual Microsoft logo before 1987 seems to have been different (but Microsoft doesn't have always been to be consistent, e.g. the Microsoft in some of the accepted Windows logos is not written in the same way as the way Microsoft was written then). However there are an abundant number of sources that say it is the Windows 1.0 logo including some from Microsoft themselves and some quoting Microsoft themselves originating from the Windows 8 logo unveiling. It could easily be that most of these sources including Microsoft are wrong, relying on us and other sources quoting us. Unfortunately this self-reenforcing spread of ideas isn't unheard of and it's something very difficult to deal with given our normal reliance on reliable sources. Even more so in the case of images where we don't have any normal sourcing standards. The general preference is to find sources from before it appeared on wikipedia where possible but given our general reliance on reliable sources, you need to convince people there is a genuine problem first which can be difficult when there are plenty of reliable sources opposing your view and none supporting. Nil Einne (talk) 12:58, 15 February 2013 (UTC)[reply]
here are some images of the original boxart for Windows 1 and 2. And here is a clip that includes all the boot logo's. Edokter (talk) — 15:27, 15 February 2013 (UTC)[reply]
And commons:Category:Microsoft Windows 1.0 gives some magazine advertisments. A logo is something that a manufacturer shows not hides but nothing shows that logo. It is a fake and I have nothing showing that 1985 logo before 2012. How unlikely has something got to be? Dmcq (talk) 21:58, 15 February 2013 (UTC)[reply]
I've done some more careful searching and found it from September 7 2010 in Logopedia which seems to have been when it was created. Dmcq (talk) 08:38, 16 February 2013 (UTC)[reply]
Thanks for the research. I, too, was convinced that it was a later creation, either a retrospective one by Microsoft or a fake. Have you contacted THE GMoD aka Chris on Logopedia to ask if he created it? Dbfirs 12:30, 16 February 2013 (UTC)[reply]
I hadn't because their last contribution there was six months ago, but looking again I see they say they are on a number of other sites so I'll see if I can get a message through. Dmcq (talk) 17:54, 16 February 2013 (UTC)[reply]
Penyulap on commons has found an earlier one from Microsoft in November 2009. A person could misunderstand that as indicating it was the Windows 1.0 logo but it doesn't actually say that. Dmcq (talk) 01:02, 17 February 2013 (UTC)[reply]
But let's not forget someone in brands of the world was already saying it was the Windows 1.0 logo in 2008. And it was uploaded to wikipedia as the Windows 1.0 logo on 2010-02-19 from brands of the world. So it seems unlikely misunderstanding the 2009 Microsoft post is a significant factor in the later possible confusion. Nil Einne (talk) 04:23, 18 February 2013 (UTC)[reply]
I'm confused here. Why would you first say there was nothing from before 2012 when it was apparently appearing on brands of the world in 2008 and we know it was uploaded to wikipedia on 2010-02-19. And why would you later say it was created in September 7 2010 when as said, we already knew it was apparently appearing on brands of the world in 2008 and uploaded to wikipedia on 2010-02-19? The 2010 does require you to check out the image description page but the 2008 is right in this thread in the post that started it all. BTW, in case there is a suspicion the IP is lying I found an archive link demonstrating brandsoftheworld did say in their upload log it was uploaded in 2008 [1]. P.S. The magazine adverts look what I've already seen and as I mentioned, I've already seen Windows 1.0 screens. Nil Einne (talk) 04:19, 18 February 2013 (UTC)[reply]
Just stupidity and not reading properly on my part. Dmcq (talk) 11:42, 18 February 2013 (UTC)[reply]
As it turns out, I was right to be cautious about the limited number of results I came across. While researching the history of the files on wikipedia, I came across a big clue (check out the file discussion page for more). This lead me to [2] stuff for a developer conference in 1986 where the logo is used, apparently to identify Windows. From there, it's also apparent it wasn't just intended for developers but to identify applications designed for Windows to the public. While doing a quick search for such applications, I came across [3] a press release for Windows 1.0 from 1985 which shows the logo albeit without colour and with Microsoft Windows overlaying the logo rather then at the sidee. My guess is it was probably used in some other advertising material (but clearly not all). Funnily enough the clue was on wikipedia all along. For example, a look at the Windows 1.0 article history would have found it. (It was what lead me to here, albeit indirectly. I found the file log by a guess from the later file log which I did find via the article history. This was before I found it had been used in our article since 2005.) I guess I should have taken my own advice to a greater level earlier on. (As I mentioned elsewhere, in my earlier searches I did come across mention of the Windows SDK and had a quick search but didn't find anything useful. I was also thinking of press releases but for whatever reason never actually searched for one which was very easy as it turns out. If I had, I would have found all this before the mess blew up. I still haven't found any advertising brochures. And for that matter I think I've only seen two or so magazine adverts albeit multi page ones.) Nil Einne (talk) 06:17, 18 February 2013 (UTC)[reply]
Resolved: As Nil Einne has found out with a sterling effort - Not a fake. Dmcq (talk) 11:46, 18 February 2013 (UTC)[reply]

+- 3db floorstanders

[edit]

Just out of curiosity- are there at all any loudspeakers (high end-upper high fi- pro studio monitors ) that have an anechoic response 20hz-20khz in within plus- minus 3db tolerance? there seem to be quite a lot of them that go far beyond 20KHZ up to 30khz but not that many that go below 30Hz at +-3db or less.

I'm only interested in "standalone" ones ie - without any added separate subwoofers to extend LF. — Preceding unsigned comment added by 77.35.49.64 (talk) 07:21, 14 February 2013 (UTC)[reply]

SAS command for estimating the parameters of a censored distribution

[edit]

Suppose you hypothesize that your data on a single variable are drawn from, say, a lognormal distribution or a gamma distribution. Your data are right-censored. That is, for each observation either you have the exact value or you know that the value exceeds some particular value.

Is there a command in SAS that will estimate the parameters of the hypothesized distribution? Duoduoduo (talk) 14:30, 14 February 2013 (UTC)[reply]

Oh, for the love of Pete, how do I clear off the background hiss on my first podcast?

[edit]

This may sound pretty serious. Go here and listen. I want to start podcasting, and before I narrate / storytell my favorite articles I find online (and sometimes off), I need to know how to clean up the sounds I don't need.

I used Audacity, if that helps you any. I need to know what to do in Audacity in order to clean away the hiss so that my podcast recording sounds more flawless. So please, how do I clean off that hiss and retain my voice? (By the way, what other sounds / things did the hiss remind you of?)

Moreover, what else could I do to improve the podcast? This is a newbie-cast, so you know as well as I do that there's some help to be needed here. FYI: I spoke directly to my laptop; no fancy equipment besides the 'top itself. The microphone is embedded, about 1" - 1.5" to the left of my webcam.

And finally, please link a prime example of a professionally done podcast. I'd like to know what it takes (whilst on a limited budget) to match their sound quality. Thanks aplenty. --70.179.161.230 (talk) 15:14, 14 February 2013 (UTC)[reply]

You should use a standalone microphone — preferably a condensor mic. You should also be using a pre-amp between your PC and the mic (unless the mic has a built-in preamp). Without one, you have to turn up the volume and by doing so amplify whatever static you have. And to improve quality even more, you can use a digital audio interface (which is a standalone box) to handle the analog-to-digital conversion. Ideally, you want to keep as much of the processing away from your PC as possible since there is so much electromagnetic noise inside your computer. The pros also use XLR inputs wherever they can instead of the standard phone jacks.—Best Dog Ever (talk) 16:19, 14 February 2013 (UTC)[reply]

Postscript

[edit]

PS: How do I prevent the overlap of the word "at" over "ENG" where it says "Posted by ENG at 8:36:00 AM?" If anyone knows their way around a Blogger blog layout, a play-by-play to getting that fixed would be wonderful. Thanks again. --70.179.161.230 (talk) 15:32, 14 February 2013 (UTC)[reply]

When I pick that link, I don't get any sound. However, you really need a better microphone. The cheapest solution might be a headset with mic. There are far better microphones, but they are also more expensive. And, of course, make sure there is no background noise when you record. After the fact, you can maybe drop out the treble to reduce hiss, but that also impacts high frequencies in the voice, so it's not a perfect solution. As for avoiding overlapping words, leaving longer gaps when you record might help. StuRat (talk) 16:08, 14 February 2013 (UTC)[reply]
No, no, StuRat, when you say "As for avoiding overlapping words, leaving longer gaps when you record might help.", you're not gettin' it. The date-stamp is what's referred to: http://imgur.com/812pIAL <-- There. Do you spot it now? How would that get corrected on the blog layout? --129.130.18.100 (talk) 18:25, 14 February 2013 (UTC)[reply]

Follow-ups

[edit]

So everyone is asking to re-record? No one knows how to edit the existing file so that the noise gets removed and only the voice remains? --129.130.18.98 (talk) 19:51, 14 February 2013 (UTC)[reply]

Most fully-featured audio editing programs will allow you to apply various audio filters (such as low-pass filters, high-pass filters and band-pass filters) to your recordings. Often this takes the form of a virtual equalizer. If the noise is confined to a narrow frequency range, or a frequency range which is well separated from the frequency range of the parts you want to keep, these tools will allow you to drop out those frequencies which contain the noise, but not those which have the signal. - The results probably won't be as good as if you re-recorded it with equipment that avoided the hiss in the first place, though. -- 205.175.124.30 (talk) 20:24, 14 February 2013 (UTC)[reply]
P.S. I have no experience with Audacity, but Googling audacity noise removal tutorial provides a number of links that might get you started. -- 205.175.124.30 (talk) 20:26, 14 February 2013 (UTC)[reply]
There are some tools let you try to remove the noise by getting a profile of it from a period of silence. Again you should not expect perfection even if your noise is fairly constant and you have good periods of silence. Nil Einne (talk) 04:48, 15 February 2013 (UTC)[reply]

3½-inch floppy disk long term storage

[edit]

If stored in ideal conditions, is there a limit to how long a 3½-inch floppy disk will retain its data for? For example, flash drives apparently discharge over time so after ~10 years would likely have lost the data stored on them. 92.233.64.26 (talk) 16:09, 14 February 2013 (UTC)[reply]

Don't know, but you also should consider how long a floppy disk drive will be available to retrieve the data. That technology is already obsolete, and these days it's hard to find a new computer which has one. So, over time, those remaining floppy drives will break down and/or be thrown out. StuRat (talk) 16:13, 14 February 2013 (UTC)[reply]

I'm only interested in the disk itself from a technical standpoint, not the hypothetical availability of drives 92.233.64.26 (talk) 16:25, 14 February 2013 (UTC)[reply]

I have some 3½'s from 1997 which I can still read. I actually have all the data on them on the hard drive, but I've got the disks and a working drive and so I just gave it a go. I also have some 5¼'s from about 1990 and some 8" of unknown vintage (I guess from the mid 1980's when I worked on IBM System/34's) but I have no drives to read these or indeed any idea as to what is on them; I forgot to label them! I think with careful storage and infrequent or no readings (so that they don't physically wear), floppys would last almost forever. Of course, if you don't read them, then you have no idea as to whether or not they actually work. --TrogWoolley (talk) 16:59, 14 February 2013 (UTC)[reply]
So, as an optimist, that means that as far as we know, if the disk isn't ever put into a drive, there is no evidence that it ever wears out! Gzuckier (talk) 17:07, 14 February 2013 (UTC)[reply]
As an anecdotal piece of evidence, I recently read a floppy that I think was made/written in the early 1990s. Shadowjams (talk) 19:21, 14 February 2013 (UTC)[reply]
Of course you need to store them in ideal conditions. I've had floppies go mouldy within a couple of years or so when stored under normal room conditions. You may be able to clean these, but you may also damage them in the process. Nil Einne (talk) 04:46, 15 February 2013 (UTC)[reply]
Yes, storage conditions are important. I've had them go bad within hours by putting them in the pocket of my business shirt. HiLo48 (talk) 04:55, 15 February 2013 (UTC)[reply]
N.B. I should mention I'm primarily thinking of 5¼ disks. I'm reasonaly certain I've had 3½ go mouldy too and you can definitely easily find results discussing it but I think they may be slightly more resistant. Nil Einne (talk) 05:02, 15 February 2013 (UTC)[reply]
Do you mean real mould? I've seen it on VHS cassettes. I've been copying my old VHS tapes to DVD and I had one that I recorded in 1986, and there was some white gunk clearly visible on the side of the tape reel. My machine wouldn't track it, so I tried it on one at work. This tracked it fine, but the sound was out of sync on the DVD. Having been played the tape once through, my machine did play it fine to I was able to record it to DVD without problem. --TrogWoolley (talk) 13:02, 15 February 2013 (UTC)[reply]
Yes, I don't know what species of mould it is, but it grows on my old VHS tapes, camera tapes, 5¼ disks and even 3½ disks especially if the protective window cover is faulty and doesn't close. I blame the damp atmosphere here (more than a hundred inches of rain last year). When I tried recently to reformat some 3½ disks manufactured around 1990, I found that they were unreliable so decided to throw them out and buy some newer ones. Dbfirs 13:53, 15 February 2013 (UTC)[reply]
I should perhaps mention this experience was in Malaysia, a tropical country. (Although I wonder whether things would actually be better here in NZ.) I never actually identified it as mould but it sure looked like it and it seemed to be growing so it seems the most likely candidate. It definitely wasn't some random dirt. Storing the disks in sealed plastic bags helped of course. Speaking of VHS tapes, you could buy tape cleaners which lightly rubbed the tape with a cotton thing soaked in alcohol (well you didn't have to put alcohol) to clean off dust and mould like this [4] in Malaysia. I'm not sure but I don't think the price was quite so bad when VHS tapes were more common. My father made something similar for 5¼ disks by removing the cover and making it spin at a low speed when connected to an external power supply so you could carefully clean the surface with a cotton bud or similar although as it was manual you had to be careful not to push to hard and scratch the disk. Of course head cleaners for both floppy disk drives and VHS players were also common. BTW, even CDs can experience fungal growth. I think I mentioned before on the RD I experience this myself, I didn't realise it was a fungal growth until I came across info on someone from South America I think it was who identified it. Nil Einne (talk) 13:08, 16 February 2013 (UTC)[reply]
I've got a 5¼ disk from 1983 that's still readable, but all my oldest 3½ disks are in a format my PC won't recognise (MFS, probably). The oldest one I have that I can confirm as readable is from 1991. Wikipedia's article on digital permanence says the maximum lifespan for perfectly-stored magnetic disks is about 50 years. Horselover Frost (talk · edits) 02:40, 16 February 2013 (UTC)[reply]
I wonder whether there is a form of refreshing magnetic media. If you read USB memory every 5 years, would this upper limit of 50 years still apply? OsmanRF34 (talk) 13:53, 16 February 2013 (UTC)[reply]

DDR3 and paired modules

[edit]

I'm replacing some bad DDR3 ram in a laptop (DDR3-1333). My question is the consequence of installing mismatched pairs. So, if I put in a 2GB in one slot, and a 4 in the other, what kind of performance hit does that mean in practice? I've read that matched modules trigger dual channel, however I also read that if it's mismatched like this, 4GB of it would be dual channel, and the remaining 2 would be single channel ("flex mode"?). Can anyone with experience fill me in on this?

Two side questions. What's the practical performance difference between single and double channel, and what would happen if I were to put in only 1 dimm. Would that be worse than putting in mismatched ones (as far as memory speed goes)? Thanks for the help. Shadowjams (talk) 19:19, 14 February 2013 (UTC)[reply]

According to Multi-channel memory architecture#Performance there's not much difference, at least on the benchmarks they ran in 2007, and I suspect the benefit has decreased in the mean time since RAM is still getting faster while CPUs aren't. "Flex mode", aka "dual channel asymmetric" is supported by some motherboards but not others. -- BenRG (talk) 01:35, 15 February 2013 (UTC)[reply]
Thank you. That seems to be the conclusion I've come to as well. As it goes, turns out I have more serious issues with this machine, but this is useful advice for the future. Shadowjams (talk) 07:37, 15 February 2013 (UTC)[reply]
Actually with the modern SOC style CPUs where the GPU is integrated with the CPU ala Intel's Sandy and Ivy Bridge and AMD's Fusion, multi channel can make a reasonable difference for the GPU. See [5] [6] for Intel. For AMD's Trinity line, I haven't seen any dual vs single channel comparisons but it's well known that it likes memory bandwidth [7] [8]. AMD's memory controller isn't that great particularly compared to Intel's one but it's hard to imagine it's that bad that single channel vs dual channel doesn't make a significant difference but 1600 vs 1866 does. In terms of the CPU, it depends significantly on the app, in particularly now that CPUs are moving in the multicore direction (although there has been a reasonable IPC gain) if your app isn't sufficiently multithreaded you may not see a difference and similarly some apps are simply not demanding of the memory subsystem. But some apps may see a difference of 10% or more [9] particularly on a CPU with lots of threads (which is less likely on a laptop) comparing single vs dual channel. And I don't get why the motherboard matters so much, it could be a limiting factor, but remember with modern CPUs and integrated memory controllers, the CPU is far more likely to me a limiting factor. From what I can tell, most or all Intels IMC's support asynchronous dual channel (AKA Intel's Flex Memory Access [10]), it's possible it's limited intentionally as Intel likes to do, but probably not since even Pentium class ones seem to have it [11]. It originated back in the old days before they integrated their IMCs. AMD IMCs have never supported that AFAIK. BTW, there has actually been significant concern for a while that memory speeds will become a limiting factor as memory speeds while increasing aren't increasing as fast as CPU speeds. See e.g. [12] which is old but [13] is new. Architectural changes (e.g. better and smarter use of caches), the integration of memory controllers and other factors appears to have reduced the effect. It's probably less of a concern for home users since it seems system performance is becoming less important for many over the past few years as systems are simply good enough and combined with the move to laptops and now mobiles and tablets power usage and size have also become very important. But as the recent article shows, it's still a concern to some, particularly now with the move to GPU integration. Nil Einne (talk) 11:57, 15 February 2013 (UTC)[reply]
Resolved