Wikipedia:Reference desk/Archives/Computing/2006 August 13

Humanities Science Mathematics Computing/IT Language Miscellaneous 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 at one of the pages linked to above.

< August 12 Computing desk archive August 14 >

DHCP failure.

I'm using a Linksys BEFW11S4 wireless router, and DHCP is failing, despite the router's insistence that it's running a DHCP server. Here's what I get when I do a sudo dhclient eth2.

DHCPREQUEST on eth2 to 255.255.255.255 port 67
DHCPNAK from 192.168.1.1
DHCPDISCOVER on eth2 to 255.255.255.255 port 67 interval 8
DHCPOFFER from 192.168.1.1
DHCPREQUEST on eth2 to 255.255.255.255 port 67
DHCPNAK from 192.168.1.1
DHCPREQUEST on eth2 to 255.255.255.255 port 67
DHCPNAK from 192.168.1.1
DHCPDISCOVER on eth2 to 255.255.255.255 port 67 interval 7
DHCPOFFER from 192.168.1.1
DHCPREQUEST on eth2 to 255.255.255.255 port 67
DHCPNAK from 192.168.1.1
DHCPREQUEST on eth2 to 255.255.255.255 port 67
DHCPNAK from 192.168.1.1

And it repeats the pattern until I CTRL-C it.

Everything works fine if I assign myself an IP address, but this is, of course, not an optimal solution. Is there something obvious that I'm forgetting? I've tried to get an IP from several clients, but all of them exhibit the same issue of waiting for a DHCP lease and not getting one. (The Windows and Mac OSX hosts just kind of sit there, but I assume they're doing this. I can try to sniff packets with ethereal if it's really in doubt.) Reloading the router's default settings made no change; there's been no configuration change since the router was installed a year or more ago. grendel|khan 00:45, 13 August 2006 (UTC)

this is similar to what's happening to me on my fickle network (see above question). I have to assign static IP's or else I just get the shitty limited or no connectivity and a weird IP/subnet.

But, I assume your rotuer is 192.168.1.1?? So why is your router try to access what appears to be a subnet, 255.255.255.255 Wjlkgnsfb 06:27, 13 August 2006 (UTC)

Since a client has no IP address to speak of when it tries to get an IP address from a DHCP server, it has to broadcast over the network, that is, 255.255.255.255. DHCP communication is done this way until a client obtains an IP address. Splintercellguy 07:09, 13 August 2006 (UTC)

It's not the DHCP that's failing, it's the whole router. The BEFW11S4 series, regardless of stepping, has a fatal design flaw/bug, that kills them after a few hours of operation, by messing up the routing tables, and eventually disabling the router completely, hastened by lots of traffic. Linksys knows about the bug, but will not release a fix for it, even in their latest firmware. The worst is that this class is incompatible with most known third-party firmwares, so the only fix is throwing it away. ThunderBird (talk) 10:32, 6 April 2008 (UTC)

PCI 2.2/2.3

I've been looking at the Linksys WMP54G which I believe to be either PCI 2.2 or 2.3. Can this be used with a conventional 32 bit PCI slot as shown here? --Kiltman67 02:54, 13 August 2006 (UTC)

The datasheet says it's a 32-bit card. So, as long as both the slot and the card support the same voltage (either 3.3V or 5V), it'll work (IIRC having two notches means it's a "universal" card which works on both voltages, but I'm not sure). AFAIK, most PCs use 3.3V for the slots. And do not worry about the bus frequency (33MHz or 66MHz); it's autodetected, and the whole bus is slowed down to 33MHz if needed. --cesarb 03:59, 13 August 2006 (UTC)

I wasn't sure what to make of the two notches. The PCI slots in my PC only have one line for the notch to go over and I'm not sure if you need to have exactly the same number of notches and lines. --Kiltman67 04:37, 13 August 2006 (UTC)

You don't. It was designed so that, as long as it fits, it's supposed to work. If the card has two notches and the slot has only one, it'll fit. Which of the two notches your motherboard's slot has IIRC determines which voltage it uses. --cesarb 15:51, 13 August 2006 (UTC)

Prince symbol in Unicode?

Has there been any proposal to include in Unicode? NeonMerlin 05:20, 13 August 2006 (UTC)

Such a proposal would not be accepted. "Note, however, that the Unicode Standard does not encode idiosyncratic, personal, novel, or private-use characters, nor does it encode logos or graphics." [1] --cesarb 16:17, 13 August 2006 (UTC)

X on Java

Hello,

I'm looking for an X server that runs on Java, licensed under BSD-like terms (such as Apache License 2.0). I've already seen WeirdX and WiredX, but neither is licensed under BSD-like terms. Any help? Kudos to all.

Jdstroy 06:23, 13 August 2006 (UTC)

Mosaic browser

Does UIUC get money for each copy of internet explorer and netscape that is sold? (click on help-> about on IE, apparently spyglass is somehow related to UIUC) -Wjlkgnsfb 06:48, 13 August 2006 (UTC)

Considering that IE and Netscape are both completely free, no. --mboverload@ 11:03, 13 August 2006 (UTC)

??? It isn't free to make them...there's a certain cost Microsoft has to distribute each version of IE for free. Wjlkgnsfb 18:04, 13 August 2006 (UTC)

What cost? --mboverload@ 23:07, 13 August 2006 (UTC)

It's a tangled history, with the original NCSA browser called Mosaic, and Spyglass producing a commercial browser also called Mosaic, but our article indicates that "Spyglass licensed the technology and trademarks from NCSA for producing their own web browser but never used any of the NCSA Mosaic source code. Spyglass Mosaic was later licensed by Microsoft, and it was modified and renamed Internet Explorer. " Also, the original Netscape browser is said not to have shared any code with the NCSA program.

Your original question asked if UIUC received money for each copy of IE or Netscape that is sold. Since neither one is sold as a standalone product, then I'd say mboverload's answer must stand -- "no." --LarryMac 18:22, 13 August 2006 (UTC)

I believe there was a lawsuit between MSFT and Spyglass sometime in the mid to late '90's asking for an account of how cash MSFT had made off of IE. MSFT said there wasn't any money because it was free. This suit was later settled. As for Netscape, I seemed to remember I read a book called Overdrive by James Wallace that lays out the story about how Spyglass got upset that Netscape's (then Mosaic's) browser was so similar to theirs. They felt Mosaic, later Netscape was getting away making a browser based on Mosaic without having to pay Spyglass/NCSA/UIUC. There was a settlement in which Mosaic changed the company name to Netscape and that Netscape could use Mosaic's technology without paying Spyglass/NSCA/UIUC. The whole story is in p. 199-201 of Overdrive (ISBN 0471291064). So, in way, Spyglass and UIUC did get paid. - Thanks, Hoshie | 18:51, 16 August 2006 (UTC)

Actually, if none are sold, then the statement is vacuously true. It is not true that there is a copy sold for which UIUC did not get money, so (equivalently) they got money for every copy sold. --Tardis 20:32, 14 August 2006 (UTC)

y'all are being too logical about these questions (shoulda figured in a "computing" section) ;) ! You know what I was getting at...Thanks for the info, though. Wjlkgnsfb 22:21, 15 August 2006 (UTC)

Computer technological advance slowing down

Why is it that it seems that computers, especially microprocessors, are stuck in their speed race? Eight years ago it was so much faster... Have they reached some kind of technological limit or what? Thanks.

I think that there is a move, with microprocessors especially, from just faster processors, to more processors (dual-core). You have to take into account the physical energy required for faster processors. Some of this is lost as heat, and too much heat can damage the processor. Harryboyles 10:15, 13 August 2006 (UTC)

We have long known that new technology would be required to perpetuate Moore's Law. A number of physics limits are looming. For example, increases in switching speeds can mean increases in power consumption and in heat dissipation requirements. The physical size of atoms and molecules eventually intrudes. Components packed too closely are more vulnerable to various problems. Deposition masks using visible light cannot have details smaller than the wavelengths allow, so we must adopt alternatives. And when we can get no further doing "more of the same", we have to spend the time (and money) to research and develop and build facilities and assure quality and yields for the replacement technology.

It's interesting to compare recent progress in general-purpose CPUs with the GPUs used for graphics cards. Development of the latter has been considerably faster. --KSmrq^T 11:49, 13 August 2006 (UTC)

Heat is becoming a major factor. Also, CPUs are now becoming more efficient. It's not only speed that matters. You can't measure a processor's power on the Gigahertz rate only. You must also look at the level 2 caches and such. There are many other factors. Take a look at how much hard drive disk space has come down in price, and look at RAM. RAM has increased very little in the years. --Proficient 16:31, 13 August 2006 (UTC)

Broken digital camera.

I have a Pentax Optio S digital camera. After lending it to someone, I find that (a) the camera will not turn on; even with a fully-charged battery. When I attempt to turn it on, the power light lights up for 3-5 seconds and then turns off. (b) the lens is twisted slightly; the text on the zoom lens reading "Pentax Optio S" is now tilted to the left, as if the circular parts have been rotated. Does anyone have any idea what could be wrong?

If something's been twisted, it might be failing an internal diagnostic. Trying to make its moving parts move might cause further damage. I'm just guessing here, but I think I can be fairly certain that this can't be fixed by you, and that you'll need to take it to a camera shop and/or see about sending it to Pentax to get it repaired, if you think it's worth it. Whoever it is that you've lent it to should probably help you out, there. grendel|khan 13:40, 13 August 2006 (UTC)

Certainly the person who lent it to you will help you pay for damages arising. --Proficient 16:32, 13 August 2006 (UTC)

The person is my fourteen-year-old daughter, so this seems unlikely. Thanks all the same.

mIRC help.

I can connect to various networks and join channels just fine. But on efnet, trying to /join #help results in the error: #help Nick/channel is temporarily unavailable. I'd normally ask for help in #Help...

Maybe their #help channel is down? It seems pretty obvious to me. Does joining #help on other servers work? Wizrdwarts (T|C|E) 20:25, 13 August 2006 (UTC)

That's odd...

My current IP 71.247.125.144 (talk · contribs) (changes every few weeks), shows me making a change to wikipedia on 18 May 2006, I didn't even have Verizon in May, how is that possible?--71.247.125.144 15:07, 13 August 2006 (UTC)

More than likely the last person to have that IP (or person before them etc.) made a change. --Kiltman67 15:35, 13 August 2006 (UTC)

Indeed. Multiple people can have the same IP at times differing in history. --Proficient 16:33, 13 August 2006 (UTC)

That is why you want to create an account. It is easy and free. --Kainaw ^(talk) 18:57, 13 August 2006 (UTC)

Having an account doesn't help. You still get blocked, even though your account has never been used for vandalism, just because the I/P you currently have was once used by a vandal. This happens to me all the time. StuRat 01:32, 14 August 2006 (UTC)

That's if the IP address is blocked. However, now it is possible for admins to block only the IP address, not any registerd users using the IP address. It depends on the individual IP address and the blocking admin. I'd agree with Kainaw. Creating an account is easy and there are many benefits, especially if you are going to become a regular editor. There's a higher percentage chance of being able to edit with a user account, as other users have a better chance of solving the problem. Harryboyles 14:00, 14 August 2006 (UTC)

We have to convince the admins to stop using the old, stupid method. StuRat 19:56, 15 August 2006 (UTC)

Windows File Handles

Something I've never been able to find out about but has bothered me for quite some time is how to look up, given a file, what processes own mutex locks on that file. As near as I can make out, Windows (or at least Windows XP) calls mutex locks file handles, hence the name of this post.

The reason I care is that I extremely fed up with trying to delete or modify a file only to discover that I don't have write access to it, even when root (administrator). When I try to delete or modify, I simply get a message that access is denied; I'm not even told whether it's an ownership issue or a locking issue, but as I'm administrator, I assume it must be locking. I realise rebooting into safe mode would almost certainly let me get around the issue, but I'd rather find some sort of more elegant solution - at least one which doesn't require a reboot.

I figure if I can find out what process owns the extant handle, I can kill the process, which should dissolve the handle, letting me delete or modify to my heart's content.

Anyone have any thoughts?

-- Rick Weinberger

According to our File locking article, you can "force close" the file locks using Process Explorer. I can't say I've tried it myself. -- Finlay McWalter | Talk 16:07, 13 August 2006 (UTC)

I downloaded it, and it can indeed force close file locks, and you can look up locks by file name. This is exactly what I was hoping for, thank you. I'm sorry it didn't occur to me to check the file locking article myself; I only looked up windows file handles (and was unsuccessful). Thanks for the help. -- Rick Weinberger

Audio storage streaming thingamajig

Here's a rather un-emphazi'd question, but i'll try my best to explain. My uncle has a music device that stores music (I think about 250GBs worth) which you can plug into your home cinema system, stream to the web etc. Only problem is i've had very little luck finding one, probably because i have the name completely wrong. What's the name of one of these, and do you know any good ones? Thanks Benbread 16:16, 13 August 2006 (UTC)

    250 GB seems like a lot.  Are you sure about that number?

The Roku Soundbridge has some of those features, but it relies on your PC for actual storage. --LarryMac 19:06, 14 August 2006 (UTC)

And this review of the Sonos Digital Music System mentions, near the end, the idea of hooking it up to a Network Storage device such as the 250GB Buffalo Link Station. No streaming TO the web though. --LarryMac 19:31, 14 August 2006 (UTC)