Jump to content

Wikipedia:Reference desk/Archives/Computing/2008 February 25

From Wikipedia, the free encyclopedia
Computing desk
< February 24 << Jan | February | Mar >> February 26 >
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 25

[edit]

Overclocking a G4 Powermac

[edit]

I don't know the model or anything, but it is a 667mhz G4. It is running Ubuntu server. I'd like to overclock it, for the experience. I really don't care if the computer breaks, it's old and worthless now. Does anyone have a tutorial on how to overclock it? Software overclock preferably. —Preceding unsigned comment added by 75.31.170.32 (talk) 03:08, 25 February 2008 (UTC)[reply]

Good luck. Last time I looked into overclocking a Macintosh, it involved really old-school overclocking techniques: re-soldering resistors around the oscillator chip. --Carnildo (talk) 00:41, 26 February 2008 (UTC)[reply]

PNG/FLAC are lossy?

[edit]

I took an image and converted it into a 24-bit BMP (uncompressed), then I converted it into a PNG (with the same color depth - in this case 24-bit), then I converted the PNG to 24-bit BMP (uncompressed) using the same application which used to create to original BMP (MS Paint/FastStone Image Viewer/The GIMP/IrfanView) with the same parameters.

By comparing the md5sums of the BMP files I noticed that the BMP file was changed. Why?

I did the same thing with WAV (PCM, 2 channels, 44100Hz, 16 bit per sample) and FLAC. The WAV file was changed too. Why? —Preceding unsigned comment added by 89.0.155.165 (talk) 12:38, 25 February 2008 (UTC)[reply]

PNG isn't necessarily lossless, and anyway decompressing and lossless-recompressing doesn't guarantee you the same data- just the same image information. Also file headers are really crazy and probably change based on the date of last modification or revision :D\=< (talk) 14:20, 25 February 2008 (UTC)[reply]
Nope, PNG is lossless. You're as misleading as always. It's true headers can be different, but not for the reasons you mentioned. It's because there's more than one way to write a valid BMP/WAV file. Let's say you have one HTML file with "<p>here comes a paragraph</p>" and another file with "<P>here comes a paragraph</P>". Sure, the md5sums will be different, but the display will be the same. --Kjoonlee 14:32, 25 February 2008 (UTC)[reply]
Grrrr, you're wrong. Deflate is lossless but the file can be altered before deflate is run to maximize its efficiency. In fact, this is the default on many PNG implementations, including irfanview. I already made the second half of your point. :D\=< (talk) 15:58, 26 February 2008 (UTC)[reply]
You've not making much sense. --Kjoonlee 15:01, 27 February 2008 (UTC)[reply]
How does that not make sense? Just because the compression algorithm is lossless doesn't mean that you can't strip image information out before compressing to reduce the final file size further. :D\=< (talk) 18:43, 27 February 2008 (UTC)[reply]
You're contradicting yourself left and right. --Kjoonlee 23:11, 27 February 2008 (UTC)[reply]
Someone step in and vindicate me! Look, deflate uses lossless run-length encoding. So say you want to save a "lossy" png. A line in your image looks like Blue, Blue, Blue, Blue, Slightly-off-blue, Blue, Blue.. normally deflate would encode it as as 4 blue, 1 slightly-off-blue, 2 blue. But since you're saving "lossy" then your image program looks for optimizations and in this case changes the slightly-off-blue to normal Blue. so now deflate compresses Blue, Blue, Blue, Blue, Blue, Blue, Blue, or 7 blue, or one third the size of the non-optimized version. It has nothing to do with PNG itself so I wouldn't mention it except that so many image editors actually do this by default. :D\=< (talk) 04:54, 28 February 2008 (UTC)[reply]
So an image editor can corrupt the image before encoding it as PNG. That's not evidence that PNGs can be lossy. It's just evidence that the image editor sucks. --tcsetattr (talk / contribs) 05:12, 28 February 2008 (UTC)[reply]
Actually, there's no evidence of that either. --Kjoonlee 21:10, 28 February 2008 (UTC)[reply]
No matter how annoyingly self-assured and spunky you try to sound, you're still wrong. Second image from the top, I guess the direct link is blocked by his webserver (so 90s). What does "compression level" mean then if not lossiness? :D\=< (talk) 05:05, 29 February 2008 (UTC)[reply]
It means compression level. ZIP is lossless and it's got compression level as well. Please stop being wrong (or pretending to be wrong.) --Kjoonlee 06:18, 29 February 2008 (UTC)[reply]
As was said, after compression and decompression you end up with the same data but not necessarily the same headers. To check that this is indeed the case you can extract the raw data (i.e. without headers) and checksum this before and after. Morana (talk) 14:38, 25 February 2008 (UTC)[reply]

I used the same application to create both of the BMP files. The headers should be the same unless they stores the date up to the seconds or something like that. How can I verify the image store the same image? —Preceding unsigned comment added by 89.0.165.111 (talk) 15:02, 25 February 2008 (UTC)[reply]

By converting them to a raw format and comparing them. If I were in a pinch I'd try to load both versions in GIMP, do a colour inversion on one version, then paste over and adjust layer settings until I get a white or black screen. Then I'd merge the layers and check if all pixels are either pure black or pure white.
With FLAC files, get the original .wav file, and compress it using /usr/bin/flac -V, where -V is the verify option. Then you run /usr/bin/metaflac --list on the file and note the embedded md5sum. This should be the same for all FLAC files with the same audio data. --Kjoonlee 17:00, 25 February 2008 (UTC)[reply]
To detect almost-identical files, I sometimes use xdelta. If the delta from one file to the other is much smaller than the size of the files, they must have a lot in common. If the files are the same size and I suspect that only the headers are different, cmp -l can confirm it. --tcsetattr (talk / contribs) 21:55, 25 February 2008 (UTC)[reply]
Windows comes with a command-line utility called fc for comparing files. It's not very good, but it might be good enough for this depending on what's going on. For a binary compare you do fc /b file1 file2. I agree that this is mysterious. Neither BMP nor WAV has a time stamp, or any other field I'd expect to change in different saves by the same application. -- BenRG (talk) 22:19, 25 February 2008 (UTC)[reply]

'programmable' stack

[edit]

a. I'm wondering if such a thing already exists - let me describe it..

I can push a series of values onto a stack (memory), (these will be n-bit descriptors - such as an item list number) If I push the same value twice onto the stack hardware detects the match and doesn't write it a second time.. (I believe L2 cache does something like this - but can it be used in this example) Obviously I need to be able to pop values back off the list.

b. Also wondering about hardware supported z-sorts - as above but this time I push two values 'z-depth' and 'descriptor' - the hardware automatically sorts every new value down the stack - so that when I pop the values back off they are arranged in z order) - Is such a thing common - is there a device that has this functionality.. 83.100.158.211 (talk) 14:51, 25 February 2008 (UTC)[reply]

A stack is not a physical thing. It's just memory in the computer like any other. It's the software the treats it like a stack. Ariel. (talk) 10:08, 26 February 2008 (UTC)[reply]
I'm asking about a hardware stack. ie if a architecture (that exists) has ever supported such a thing as described above. Should have said...87.102.42.162 (talk) 14:16, 26 February 2008 (UTC)[reply]
Like this: Stack machine? Ariel. (talk) 01:06, 27 February 2008 (UTC)[reply]
Erm.. No. (thanks anyway) (or even yes..sort of) - I was thinking of a more covnentional processor (eg 1,2 or 3 operand instruction set in the language of that article) with a stack as an extension to the architecture.. The stack would work like a conventional software stack, but it would have additional co-processor functions that operated automatically on the stack eg a z-sort.87.102.93.245 (talk) 11:24, 27 February 2008 (UTC) (I assuming that the z values in my example wouldn't count as a 'machine program code' though that's probably irrelevent to my original question)[reply]

Juice sync with Sansa c250 on Tiger

[edit]

Hi, I am on a Macbook with Tiger 10.4.11. I recently installed Juice (software). Can I automatically sync my Sandisk Sansa c250 (which the friends at #rockbox recently helped me "unbrick") with the podcasts? Please suggest me how to do it if it is possible. Kushal 16:30, 25 February 2008 (UTC)[reply]

No replies? Kushal 18:00, 27 February 2008 (UTC)[reply]

Seemingly Bricked Thinkpad

[edit]

I've had a Thinkpad T40 for the past 2 years, and it has never given me any problems until 3 days ago when it inexplicibly restarted while I was using it. I had no need to use it for a few days, so I just ignored it. Yesterday I turned it on, only for it to freeze on the Windows XP welcome screen, so I restarted it. Thereafter it restarted at random points while using Windows (but never at the XP loading screen). I did a System Restore to a week ago and it worked fine and stayed on for about 2 hours before it started restarting at random again. Now it has degraded to the point where it restarts seconds after I turn it on (at the IBM logo). During all of these restarts it was receiving steady AC power from different outlets each time. Can anyone think of anything I can do to fix this, or do I have to get it repaired? 206.252.74.48 (talk) 18:24, 25 February 2008 (UTC)[reply]

I had my desktop computer do this for the last 6 years. I changed everything: motherboard, graphics card, memory, processor… Your situation degrades much faster so it could be a memory problem, definitely not the problem I had. Admiral Norton (talk) 18:38, 25 February 2008 (UTC)[reply]

Seems to be a hardware related problem. Just to make sure that it is not Windows misbehaving, could you try to run (NOT install) a LiveCD version of Ubuntu (Linux distribution)? Kushal 18:44, 25 February 2008 (UTC)[reply]

(I am the OP) I've tried a Linux boot CD, and here is my experience: I push the power button and nothing happens, so I push it harder - it turns on. It gets to my BIOS password entry and the keyboard doesn't work. I restart it myself. This time the keyboard works and I get the boot screen of the LiveCD and press enter to get it going. It starts to load Linux and the screen turns black (well, LCD gray, actually). It was then that I turned it off. Definately a hardware problem - but of what? The motherboard? How much would it cost to fix this thing, I've already spent too much this week. Chris16447 (talk) 23:51, 25 February 2008 (UTC)[reply]
Depends what the problem is. If it's a RAM problem then it's easy to fix (just swap out the bad RAM with good RAM); try running memtest86 if you can and see what it tells you. If it is a major problem with the motherboard itself, then the cost becomes quite expensive (hundreds of dollars just for the part, no to mention the labor, which will be a lot as well) to the point of being prohibitive. --98.217.18.109 (talk) 00:40, 26 February 2008 (UTC)[reply]

If it was the motherboard on a laptop, I would not replace it. It is just not worth the headache and sleepless nights of figuring the problem with the old one, finding a new one, buying it, waiting for shipping, installing it, making sure it works, and making sure it won't fry up the instant I press the power button, and so on.

If it is the motherboard, and if I had the money, I would just go out and buy a new computer. Kushal 01:26, 26 February 2008 (UTC)[reply]

I bought the laptop at half-price for $500. Thusly, I am willing to pay up to $500 to what the full price would have been. Will repairing a busted motherboard cost any more than 500 USD? 206.252.74.48 (talk) 18:16, 26 February 2008 (UTC)[reply]

If the motherboard was broke, and if you could fix it, it probably would not cost very much. I believe if you just busted your motherboard, you should be able to find one for less than USD 150, (that was a completely wild guess based on desktop boards, your mileage may vary). Kushal 18:00, 27 February 2008 (UTC)[reply]

SRS reporting

[edit]

Need information or tutorial on how to build SRS reports —Preceding unsigned comment added by LeLeRags (talkcontribs) 18:57, 25 February 2008 (UTC)[reply]

SRS could mean a ton of things. Could you be more specific, please? Kushal 02:35, 26 February 2008 (UTC)[reply]

I assume this means SQL Server Reporting Services. Google "SSRS tutorial" and try something like this link. 207.148.157.228 (talk) 14:58, 26 February 2008 (UTC)[reply]

where i could find the the different interesting programm codes that could run on borland version 3.0 of C++

[edit]

116.71.177.95 (talk) 19:07, 25 February 2008 (UTC)[reply]

I'm puzzled why you would want to compile programs using Borland C++ version 3.0. Check out the article I linked to - it's ancient, and speaks an obsolete C++ dialect. Is it for learning purposes? If so, there are better and free alternatives (such as version 5.5. of Borlands compiler). Is it because you want something to work on an old DOS machine? If you could explain a bit more, it would be easier for us to give a better answer. What kinds of programs interest you? --NorwegianBlue talk 20:55, 25 February 2008 (UTC)[reply]