Computing desk
< July 6	<< Jun \| July \| Aug >>	July 8 >

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.

July 7

1000 ft Ethernet Cable

I need to get internet from one building to another: let's say building 1 to building 2. The problem is that building 2 is approximately 1000 ft. away from the modem/switch in building 1. I've done some research and determined that it's not a good idea to just run a cat5e cable the whole length (due to loss), and it's not practical to boost the signal partway through (since the wire will be outside). I've thought about using baluns, but I'm not sure if they make them for that application, or which ones would work. Is there any other way to boost the signal so that one wire could be used? A different kind of wire perhaps?

Any suggestions would be appreciated. Thanks! -Pete5x5 (talk) 00:54, 7 July 2009 (UTC)[reply]

What's the budget? - Akamad (talk) 02:32, 7 July 2009 (UTC)[reply]

I'm sorry, but you're simply not going to be able to do it with Cat 5e because (as you've said) after ~100m/328ft you'll lose signal and there isn't a way round that. However, you're probably be better off looking at using a fibre connection instead and 1000Base-LX could easily handle the distance you're looking at and you could simply have a converter at each end to change it back to copper for your two networks (a lot of switches include the ability to plug in different gigabit modules of various types to do this automatically. Hope this helps! ZX81 ^talk 02:39, 7 July 2009 (UTC)[reply]

If you have a line-of-sight between the rooftops then a microwave link is a good way to do it - but it's gonna cost you $15k for the equipment. SteveBaker (talk) 03:00, 7 July 2009 (UTC)[reply]

Or, needless to say, you could spend $15 without the K and use a Pringles can, though the fact you are considering running a wire tells me that you want more speed and reliability than the Pringles company can provide you. Tempshill (talk) 03:25, 7 July 2009 (UTC)[reply]

I was going to suggest pigeons. Wireless! — Twas Now ( talk • contribs • e-mail ) 03:37, 7 July 2009 (UTC)[reply]

While some of the options such as fiber and wireless/microwave are certainly viable options, my immediate thought would be that the most cost effective solution would be to lay the problem in your ISPs lap. A second access point/modem/account. Once both buildings are "online", then the solutions to data sharing, communication, intranets, or VPN setups multiply. Just a thought. — Ched : ? 04:00, 7 July 2009 (UTC)[reply]

There might be 2 solutions using this ~330m ethernet cable: 1. at both ends use DSL modems (or something similar), since DSL has much more range than ethernet, cheap (reasonably) modems will be much slower than ethernet. Using modems made for this particular purpose, will be very expensive. 2. add 2 cheap ethernet switches in middle (and run an additional power cable along ethernet cable, to power them). -Yyy (talk) 04:06, 7 July 2009 (UTC)[reply]

Why not use several cables and connect each one into a hub or someshit to boost the signal half way through —Preceding unsigned comment added by 82.43.91.128 (talk) 07:10, 7 July 2009 (UTC)[reply]

Because the cables would be outside. These signal boosters would be outside in the weather, could get stolen, and would be difficult to power without a nearby outlet. Useight (talk) 14:58, 7 July 2009 (UTC)[reply]

You could stick them in weather proof casings, and run a power supply to them.. (a bit like telephone companies do)83.100.250.79 (talk) 15:29, 7 July 2009 (UTC)[reply]

Alternatively you could boost the signal using UHF amplifiers, though you'd need to know the attentuation to prevent the boosted signal being too high and frizting the electronics.83.100.250.79 (talk) 15:30, 7 July 2009 (UTC)[reply]

I can second the DSL-style approach. There are several companies that sell equipment for just this sort of purpose, with speeds ranging from a few megabit all the way to 100 mbit full duplex (as budget allows) and distances well in excess of 1000ft. Look for "ethernet extender" products on Google. --Jmeden2000 (talk) 18:54, 7 July 2009 (UTC)[reply]

Thanks for the suggestions. The budget is $4000 or less (preferably less, but I have a very nice budget for this project). I would prefer not to join multiple cables, but before this that's the only way I could think of. I will definitely look into the 'ethernet extenders' you speak of. Hopefully there's a way to do it with only one wire. I think the technology must exist for that amount of money.. -Pete5x5 (talk) 07:35, 8 July 2009 (UTC)[reply]

If you are going to be accessing file servers or suchlike from one building to the other I'd go for the fibre optics solution. Or even just to cater for the future. A building of any size can generate quite a bit of traffic. Dmcq (talk) 09:37, 8 July 2009 (UTC)[reply]

Any reason you cannot use WPA encrypted wireless? Like WiMax? --70.167.58.6 (talk) 02:32, 9 July 2009 (UTC)[reply]

pspwxp(wobzip)

is there any way i can install pspwxp using wobzip.heres the address http://wobzip.org if there is can sombody pleaz give me a link to download it and instructions on how to download it. thanks

Hard drive space

I only have about 6.5 GB left on my hard drive, so I deleted 3 GB worth of files and put them on discs, then I checked my memory, and it still said 6.5 How can I get my space back? ---Scarce |||| Talk -Contrib.--- 04:14, 7 July 2009 (UTC)[reply]

This might sound really obvious, but have you emptied the recycled bin? The space isn't actually reclaimed until you do that. ZX81 ^talk 04:41, 7 July 2009 (UTC)[reply]

My entire drive for the bin is erased. For future reference I'm talking about my main local drive ---Scarce |||| Talk -Contrib.--- 05:19, 7 July 2009 (UTC)[reply]

Try disk cleaning (Start Menu -> Accessories -> System tools -> disk cleaning) if you have Windows. There's probably too much junk in the temporary folder that can be cleaned. Although it's strange that there is no change in disk space at all. --98.154.26.247 (talk) 05:50, 7 July 2009 (UTC)[reply]

If you burned them to disc your CD/DVD writing software may have created a disc image on the hard drive first before burning the files. The disc image will take approx. the same space as the original files, hence you don't see a change. Try searching for ".ISO" files or check your preferences of your burning software to find its "temporary" or "working" folder and delete the disc images from there. Zunaid 10:59, 7 July 2009 (UTC)[reply]

I'll try it ---Scarce |||| Talk -Contrib.--- 21:19, 7 July 2009 (UTC)[reply]

I opened up the computer search program and searched .ISO, it couldn't find anything, I'll try the disk cleaning now ---Scarce |||| Talk -Contrib.--- 21:35, 7 July 2009 (UTC)[reply]

Try searching for files created in the last day or two and/or files greater than 100MB. You may find the offending files, or possibly some other files worth cleaning up. -- Tcncv (talk) 00:09, 8 July 2009 (UTC)[reply]

How? ---Scarce |||| Talk -Contrib.--- 02:09, 8 July 2009 (UTC)[reply]

If you are using XP, right click the C: drive and select search to open the search window. The search dialog has "when was it modified" and "what size is it" options that allow you to define your search. If you have Linux, a Mac, Vista or some other OS, let us know, so perhaps someone else can provide instructions. (I actually have Vista, but have yet to find equivalent built-in search functionality.) -- Tcncv (talk) 03:44, 8 July 2009 (UTC)[reply]

Chipset flashing?

Is there such a thing as chipset flashing? Is it different from BIOS flashing? Or do I just mean updating the chipset drivers? Thanks 78.149.207.75 (talk) 09:25, 7 July 2009 (UTC)[reply]

"flashing" generally means "changing software stored in flash memory". The BIOS is stored in flash memory, so it's possible to change that stored program by flashing that memory. The chipset isn't a program, it's a very complex application-specific integrated circuit. It's not stored in, or run from, any kind of memory; its operation is encoded into its physical structure. So it's not meaningful to "flash the chipset". I think you just mean the drivers. 87.113.26.43 (talk) 13:34, 7 July 2009 (UTC)[reply]

It seems possible that some chipsets could have an internal or external ROM. I don't know of any PC computer chipsets (north-bridge or south-bridge controllers, or audio or wireless peripheral controllers, etc) that make this option available to the user (though I have re-flashed my DVD-ROM drive controller). Generally, a controller chip might not have any ROM at all (a pure ASIC as mentioned above); or they may use a ROM intended to be programmed once at the factory. If there is an EEPROM, it can be flashed; but that does not mean that you can do it without special equipment. (The motherboard is usually designed to allow you to flash the BIOS without connecting any wires - this is not generally the case for all programmable devices). See programmer (hardware). Nimur (talk) 15:44, 7 July 2009 (UTC)[reply]

I actually wrote, then deleted, an addendum to my comment above, about just this subject (fearing we'd drift off from the questioner's presumed intention). Yes, there is a microcontroller inside most chipsets, which takes care of some very basic functions (it's the thing that handles wake-on-lan, and that beeps when the CPU isn't installed or the CPU fan isn't working). I'm aware of no public documentation for these (informally I understand them to be 8051 or H8 logic blocks on the same die as the rest of the southbridge logic), and there's no reason to, or advantage to be gained from, field-programming one. While being factory programmable would be nice (as it stops you making 100,000 bad units for want of a single line of asm being wrong) having erasable memory for it may be a problem. I'm not at all sure that the process by which the high-speed circuits on the southbridge die are deposited would allow for either flash or EEPROM to be built on the same die. If they can't be, having a rewritable program store for the microcontroller would require a second integrated circuit (built with whatever process is compatible with those memory technologies) inside the same package - a cost and complexity that chipset manufacturers would no doubt wish to eschew. Given the very basic program for this thing, it would be tempting to test it very thoroughly and put its code into Mask ROM on the southbridge die. 87.113.26.43 (talk) 16:21, 7 July 2009 (UTC)[reply]

InstantAtlas Error

When creating a map on InstantAtlas I get an error when uploading my .xml data file ( something about Attribute version having a fixed value of 1.2) so I create a demo data file instead. I can then replace the data.xml file with mine in the location I save it. On opening the map with my data file the data that should be displaying on the map doesn't, it just says no data in the comparison table, under the headings I used, and there is nothing in the ordinary table. The map highlights the Popp areas of my map in black when i hover over them. Any ideas? —Preceding unsigned comment added by 195.49.180.146 (talk) 09:50, 7 July 2009 (UTC)[reply]

malloc and structs

Ignoring the lack of error checking code, would this be the proper way to do this?

typedef struct
{
    int data;
} Mystruct;

Mystruct * lotsaStructs;
lotsaStructs = calloc(10, sizeof(Mystruct *));
lotsaStructs[0] = malloc(sizeof(Mystruct));
lotsaStructs[0].data = 5;

Thanks! Horselover Frost (talk) 13:53, 7 July 2009 (UTC)[reply]

No, that doesn't even compile. The following would be typical, though it's not clear whether you want lotsaStructs to point to a single array of Mystructs or to an array of pointers-to-Mystruct:

lotsaStructs = calloc(10, sizeof(Mystruct));
lotsaStructs[0].data = 5;
...

--Sean 14:26, 7 July 2009 (UTC)[reply]

I'm hoping to get an array of pointers-to-Mystruct. I guess the question wasn't to clear, sorry. Horselover Frost (talk) 14:40, 7 July 2009 (UTC)[reply]

TotoBaggins' solution (which is the sensible thing, in many cases) has all of the individual structs in a row (an array of structs). If you wanted to keep Mystruct being an array of pointers (which is what you'd do if you wanted to create the individual ones dynamically, or if they weren't all going to be the same size or the same kind of thing) then you'd do:

 Mystruct ** lotsaStructs;
 lotsaStructs = calloc(10, sizeof(Mystruct *));
 lotsaStructs[0] = malloc(sizeof(Mystruct));
 lotsaStructs[0]->data = 5;

your code was defective in two regards: firstly lotsaStructs is an array of Mystruct*, so you need to write Mystruct** Secondly when you're accessing data via lotsaStructs[0] that's an individual Mystruct*, so you need to dereference the pointer before retrieving the element: C uses the -> shorthand for that, although the line (*lotsaStructs[0]).data = 5 does the same thing (and makes the deref and the .data separate). Don't worry if this seems ridiculously complicated; once you understand this, there's really nothing left in C that's hard. 87.113.26.43 (talk) 14:50, 7 July 2009 (UTC)[reply]

Accessing Wikipedia through my iPhone

Can anyone help with this? Until a few days ago, I had no problem accessing Wikipedia through my iPhone. Now, I just can't. It starts opening en:, seemingly redirects to "en-gb [something - it flashes fast past this] and throws up a "Cannot Open Page" message. I have no problems accessing other websites. Cheers. --Dweller (talk) 16:02, 7 July 2009 (UTC)[reply]

OK: a totally random theory: have you set something on your iPhone's settings to have your language be British English? It might automatically change en to en-gb. Unlikely, but a possibility. Thanks, gENIUS 101 21:02, 7 July 2009 (UTC)[reply]

This might be related to changes being made to the MediaWiki:Common.js file – something to do with redirecting mobile devices to an alternate site. I don't know the details, but I'll ask the user to comment. -- Tcncv (talk) 23:54, 7 July 2009 (UTC)[reply]

Up until a few days/weeks ago, any iphone/ipod touch/etc requests to en.wikipedia.org/wiki/* had a banner at the top that offered "View this site on Wikipedia's Mobile Site". Upon clicking the link, you were taken to en.m.wikipedia.org/wiki/* (Main Page - Mobile. It was very succesfull, and much easier to browse. More recently, due to the ease of use, you are automatically redirected to the en.m. site. It works fine for me. I have my iPhone loading it right now :) The page might have just been down when you tried earlier. — Deon555^talk_{I'm BACK!} 11:11, 8 July 2009 (UTC)[reply]

As a side note, check out Wikipanion for iPhone (it's free). It's a great browsing app that caches articles for offline viewing. Much better than just viewing Wikipedia through mobile Safari. --70.167.58.6 (talk) 02:23, 9 July 2009 (UTC)[reply]

Still unable to "get in" - this has been going on for days now. The specific error message is "Cannot Open Page Safari cannot open the page because the server cannot be found". The url in the address bar when it finishes failing to load is http://en.wikipedia.org/wiki/ [whatever I've tried to open]. I can try downloading that app (does it support logged-in editing too?) but many others will just want to browse Wikipedia and not be able to access it. I've not changed any settings in my phone - it's just too much of a coincidence that it's happened simultaneously with the mobile site changes, so I suspect it's not my phone but the way the change has been implemented. --Dweller (talk) 11:26, 9 July 2009 (UTC)[reply]

~~Are there any other websites you are unable to access or is it just Wikipedia?~~ Durr, I can't read. Livewireo (talk) 16:18, 9 July 2009 (UTC)[reply]

Mouse totally f....d up

Mouse problems - just doesn't work right

Used to select with left button, hold left button and move to select a block.

Now mouse selects with right button (sometimes), left button causes block selection from last mouse position, scroll button doesn't work...

Same on another mouse.

No such control exists in windows to do this that I'm aware of?

I just tried open office - is this program known to fuck up windows??? Because I've never had anything like this before, maybe it's a virus, anyone know?

Appears to be the "stuck shift key problem" - what causes that - is it open source software? ? —Preceding unsigned comment added by 83.100.250.79 (talk) 21:48, 7 July 2009 (UTC) 83.100.250.79 (talk) 21:10, 7 July 2009 (UTC)[reply]

If this is the problem, it means your keyboard thinks the Shift key is being held down. Try tapping each of your Shift keys repeatedly until the problem goes away. If that doesn't work, unplug the keyboard for a bit and see if that fixes the selection problem. If you can't make it go away by tapping the Shift key and if unplugging it does cause the problem to disappear, it's time to buy a new keyboard. Tempshill (talk) 21:56, 7 July 2009 (UTC)[reply]

If the keyboard thinks shift is being held down - why no caps? 83.100.250.79 (talk) 22:07, 7 July 2009 (UTC)[reply]

I highly doubt it's an issue related to Open Office. Would I be correct in assuming that you've tried a restart of the machine? - Akamad (talk) 22:34, 7 July 2009 (UTC)[reply]

I hope you're right, I actually uninstalled open office, since the computer has been stable for many moons, and had just very recently installed open office. The problem didn't start immediately, but did seem to appear after I had been messing about with the image program (can't remember it's name).

Then the problem seemed to go away. I intend to wait a few days to see if the problem recurs, then reinstall open office, and see if the problem re-occurs.

It appeared to be a 'software' stuck shift key that only affected the mouse, not keyboard - does this sound familiar to anyone?83.100.250.79 (talk) 22:45, 7 July 2009 (UTC)[reply]

Have a look at the windows control panel applet for the mouse, and make sure some joker hasn't switched your mouse to be left handed. Also have a look at the mouse itself - if something got wedged under the buttons so your PC thinks one is being held down, you get all kinds of weird things happening. Astronaut (talk) 01:12, 9 July 2009 (UTC)[reply]

Choose a topic

The "choose a topic" at the top right of the page keeps dissapearing - sometimes it's there. sometimes it's not, sometimes it pops in, and pops out. I also had the problem above - has anyone else got this, or is my computer just screwed up (IE7)83.100.250.79 (talk) 21:31, 7 July 2009 (UTC)[reply]

IE 7 has done this to me before as well. To put it simply, IE is as buggy as hell. Try downloading IE8, or better yet, the new Firefox (both which are free)-- penubag (talk) 06:23, 8 July 2009 (UTC)[reply]

I've only ever had one problem on IE7 - the above, so does anyone know what it is about the ref desk header template that causes it to go wrong occasionally?83.100.250.79 (talk) 17:47, 8 July 2009 (UTC)[reply]

SImple MySql question

Is there another way to count the amount of rows in a table then using mysql_num_rows? When the rows get passed 5000+ and that query is ran it can cause slowness. —Preceding unsigned comment added by Ivtv (talk • contribs) 23:02, 7 July 2009 (UTC)[reply]

The SQL count() function is much more efficient, sometimes returning a result in constant time. From this discussion, the commands:

$row=mysql_fetch_array(mysql_query("select count(*) as howmany from mytable"));

$number_of_rows=$row["howmany"];

is much more efficient than

$table=mysql_query("select * from mytable");

$number_of_rows=mysql_num_rows($table);

-- Tcncv (talk) 23:44, 7 July 2009 (UTC)[reply]

Yeah. I keep forgetting about count. Thank you for your example. Only thing I would change would be the wildcard because its terrible on a query especially if you are just trying to get one specific piece of data. Should just query your key in this example. You know this though, and you were just giving an example so I thank you. Works perfect

Ivtv (talk) 02:34, 8 July 2009 (UTC)[reply]

My understanding and experience is that the assertion that "count(1) is faster than count(*)" is a myth that has propagated far and wide in various discussion groups. The count(*) function means "count the rows" and is typically optimized to use table statistics or indexes. Count(1) means count a collection non-null ones. Any DBMS worth its salt will produce the same execution plan for each. In neither case is it necessary for the DBMS to actually retrieve column data. MySQL COUNT function documentation specifically states, "COUNT(*) is optimized to return very quickly if the SELECT retrieves from one table, no other columns are retrieved, and there is no WHERE clause." -- Tcncv (talk) 06:48, 8 July 2009 (UTC)[reply]

Service Packs for Linux distros exist?

I was just wondering if there are any known linux distros that uses the concept of "Service Packs"?: something that can can be installed in a graphical and easy way, to update your system? (besides my own distro). Thanks __ Hacktolive (talk) 23:47, 7 July 2009 (UTC)[reply]

Isn't the term Service Pack a word more associated with Windows that Linux? With Linux being much more modular and everything being in packages I was under the understanding that all the recent distributions came with their own form of update system (like in Fedora you'd simply run yum update). ZX81 ^talk 00:29, 8 July 2009 (UTC)[reply]

Yes, and that is nice, but what about for users without internet? that update system (and also APT) just doesn't work properly and out-of-the-box. Hacktolive (talk) —Preceding undated comment added 00:39, 8 July 2009 (UTC).[reply]

For debian based distributions the install cd tends to be a big package repository. So for every new version you just point apt there and do a dist-upgrade. It doesn't need Internet but I don't know if it's "service pack like". --194.197.235.36 (talk) 01:15, 8 July 2009 (UTC)[reply]

SuSe Enterprise Linux has service packs. --Sean 16:03, 8 July 2009 (UTC)[reply]

Even in Windows, if you have absolutely no access to anything on the Internet in any way whatsoever, you cannot install a service pack. Someone, be it a friend or Microsoft, must download the service pack and burn it to a CD for you. With Linux, there is the same limitation. For example, if you use Fedora, the packages are in RPM format. You can use yum as a command-line updater or one of the graphical RPM management tools. With RPM, Yum, or one of the GUIs, you can update using packages burned to a CD. So, if you can get a friend or Redhat to give you a CD with the new packages on it, it is trivial to do an update.

Now, the concept of "service packs" is a bit different than just "updates". Microsoft gathers up a bunch of bug fixes and eventually puts them all out as a service pack. While updates come out every Tuesday (if I remember correctly), service packs come out every few months or even every few years. The same concept, applied to Linux, would be nothing more than making a CD of all the latest packages at some point in time and calling it a "service pack". When to release the service pack is arbitrary. -- kainaw ™ 16:10, 8 July 2009 (UTC)[reply]

At least for Windows XP Service Pack 2, you could order CDs with just the service pack (and I believe it's a general option available at times, but SP2 was a really big deal). I believe you could do that in an offline manner, as well (phone, mail). Also, Patch Tuesday

Wikipedia:Reference desk/Archives/Computing/2009 July 7

Contents