Wikipedia:Reference desk/Archives/Computing/2008 June 12

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Computing desk
< June 11 << May | June | Jul >> June 13 >
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.


June 12[edit]

Lightweight desktop Linux[edit]

Can anyone recommend a Debian based (preferably) Linux distro for an old(ish) laptop (Dell Gateway 2000, Solo 9100). Its for a friend who isn't exactly a computer expert, so it should be easy to use and configure. He'll be mostly using it for homework and the like.

System requirements are pretty tight. It's from the Windows 98 days. 233mhz Pentium II and 64mb's of RAM, 4gb hard drive (in two 2gb partitions).

Has some maybe strange hardware, like a combination DVD-ROM/floppy drive, but I'm mostly sure It's just two ATA devices in one box. Floppy boot capable, but haven't tried. Supposedly CD boot capable, but not booting of my Ubuntu CD-RW (hardware not powerful enough for it anyway) though Win98 can read it.

I can provide any additional information. Thanks in advance -WikiY Talk 00:18, 12 June 2008 (UTC)

Have you tried Xubuntu? I've had luck using that on old laptops. -- kainaw 00:19, 12 June 2008 (UTC)
If you really want to dig into it, you could go Linux From Scratch and install the X-Window packages afterward, then the packages for any desktop you put on it. JeremyMcCracken (talk) (contribs) 06:49, 12 June 2008 (UTC)
I've recently used Puppy Linux on a 400MHz PII, with 128MB RAM and 8GB hard drive. I then upgraded it to 640MB RAM and 40GB hard drive and tried Vector Linux. Both Linuxes worked well. However, neither would work on my really old 200MHz Pentium MMX with 64MB RAM (I think the 1st gen Pentium architecture didn't suit them). Anyway, I'll probably give Damn Small Linux or Xubuntu a go next. Astronaut (talk) 10:23, 12 June 2008 (UTC)

DLLs[edit]

Someone told me that even after uninstalling programs, there are a lot of left over DLLs that may take up space that could otherwise go towards something else. Is this true, and if so, how much space would these files typically take up? Would it be a noticeable difference if I was able to get rid of them if there were enough of them? And is there a program that could accomplish this?-- 00:37, 12 June 2008 (UTC)

If these were shared DLL files, you would get into some inconveniences. Kushal (talk) 00:55, 12 June 2008 (UTC)
It depends on the program. Some uninstallers only pretend to uninstall a program. They just remove the entry for the program from the Add/Remove Programs list. Others will remove all traces of the program. As for space, the most a DLL will take up is a few megabytes, and they are usually smaller than that. The thing I'd worry about would be old DLLs loading themselves into memory, which would slow down your computer. I tried a program once called DLL Toys but it didn't catch most of them. I also hear that Registry Mechanic will delete useless DLLs. But I've gotten the best results by going into the C:\WINDOWS\system32 folder and switching to the Details view. Most DLLs are stored in that folder. Then, I sort the files by company. So if you uninstalled Zone Alarm, for example, you would look for any files made by Zone Labs. You can see what DLLs are being used at any time by opening up a command prompt and typing tasklist /m. That's only scratching the surface, though. If you want to do a more thorough job, you'd have to run a registry cleaner and look through your C:\WINDOWS\system32\drivers and C:\Documents and Settings\ and C:\Program Files\Common Files folders for other pieces of the program.--Hello. I'm new here, but I'm sure I can help out. (talk) 01:05, 12 June 2008 (UTC)
Thanks for the help. When I opened up the command prompt and typed in tasklist /m, it says that "'tasklist' is not recognized as an internal or external command, operable program or batch file"; what did I do wrong? Nevermind, I figured it out by downloading the required application.-- 02:45, 12 June 2008 (UTC)
Are you using Vista? It may be that tasklist isn't included in it. I'm using XP.--Hello. I'm new here, but I'm sure I can help out. (talk) 04:01, 12 June 2008 (UTC)
No, I have XP. Tasklist wasn't included for some reason, so I downloaded it, and it works fine. Thanks for the tips.-- 09:31, 12 June 2008 (UTC)
You're welcome. :)--Hello. I'm new here, but I'm sure I can help out. (talk) 09:51, 12 June 2008 (UTC) [banned user user:Primetime ]

GoLive Woes[edit]

Hi guys,

I was silly enough to lay out a page in GoLive. Unfortunately, that program used pixel co-ordinates to position text on the page. It looks OK in Internet Explorer, but Firefox and Opera mess up the formatting. I'm not sure how to move the "your computer" line away from the "I can help" one without doing a massive rewrite of the code. The bullets are in Wingdings, but I can't get those to display outside of IE, either:

Thanks,

Hello. I'm new here, but I'm sure I can help out. (talk) 01:13, 12 June 2008 (UTC)

Your link doesn't work. Preemptively, though, I can think of one thing to check. Look at the css pertaining to the DIV or SPAN tag corresponding to the misaligned text. If it's positioning says position: fixed, change it to position: absolute. Fixed doesn't work in every browser, and I've had that cause problems like what you're describing. JeremyMcCracken (talk) (contribs) 02:48, 12 June 2008 (UTC)
Sorry about that. The link is fixed. I tried messing with that, but it doesn't seem to work.--Hello. I'm new here, but I'm sure I can help out. (talk) 03:38, 12 June 2008 (UTC)
I'm looking at it, and I'll post if I find anything, but that's the most jumbled, mangled mess of HTML I've ever seen.
I notice the picture isn't working in Firefox, either. I think I know the problem there. The line that begins with DIV name="22F", at the end of the line is z-index:-1" there needs to be a semicolon between the one and the quotation mark.
Check the line starting with div name="155"; it's also missing a semicolon at the end. JeremyMcCracken (talk) (contribs) 06:27, 12 June 2008 (UTC)
Tell me about it. It was even worse before I tried to clean it up! I haven't been able to get the placement right, though, so I'm trying to code it in Notepad. Thanks for looking, Jeremy and Antilived. I never was a fan of CSS, anyway. I just use tables.--Hello. I'm new here, but I'm sure I can help out. (talk) 09:39, 12 June 2008 (UTC) [banned user Primetime]
I suggest you just start over and code it the hard and elegant way. It will take more effort to make it standard than to rewrite it. --antilivedT | C | G 07:25, 12 June 2008 (UTC)

WebSite with WWW or without?[edit]

I am considering making a website, but I don't know if I should choose www.mysite.com or just mysite.com, what is better? if I choose with WWW, will it redirect to the other one (www.mysite.com redirect to mysite.com)? and vice-versa? Is the WWW really necessary? should I go for it? thanks in advance. SF007 (talk) 03:18, 12 June 2008 (UTC)

The version without really is still www.mysite.com. It just has to be setup on the server to have an alias of some sort to make mysite.com go to www.mysite.com (which is why some sites don't work if you don't put the www). When you register a domain, you're usually registering a second-level domain, where the "www" part is part of that second-level domain's subdomain. --Wirbelwindヴィルヴェルヴィント (talk) 03:30, 12 June 2008 (UTC)
I'd add that the www doesn't serve any real purpose from the technical side of things, and most of the time I just make sure that www.mysite.com and mysite.com point to the same place, as Wirbelwind suggests. The main reason for keeping the www (other than the chance that some people will type it in anyway) is that it indicates to a reader that we're talking about a website - when you see www.mysite.com on a billboard, you know that it is a url, while mysite.com isn't as blatantly clear. :) - Bilby (talk) 11:48, 12 June 2008 (UTC)
I'd go as far as to say anything with a ".com" is ubiquitous of a website, so the www is not needed as long as the ".com" is there; or .edu/.gov/.org/.etc. :P-- 10:37, 13 June 2008 (UTC)
.. and the www version actually involves less typing as (in Internet Explorer and Firefox at least) you can type "mysite" in the address bar followed by Ctrl-Enter to get http://www.mysite.com. AndrewWTaylor (talk) 14:19, 12 June 2008 (UTC)
Or in Opera, where it automatically adds it. --Wirbelwindヴィルヴェルヴィント (talk) 16:27, 12 June 2008 (UTC)
yeah, with WWW people know immediately that is a website. Thanks for the answers. SF007 (talk) 02:49, 13 June 2008 (UTC)
The "www" prefix came about in the years before the Web was as ubiquitous as it is today. Back then, an organization might have their online presence spread around, say, www.example.com, ftp.example.com, gopher.example.com, telnet.example.com, irc.example.com, etc., and those might well be different hosts with different IP addresses. Meanwhile, the unprefixed domain name example.com might be only configured for mail delivery, with only a MX record and no A record at all. (In hindsight, it might've been more convenient to have separate DNS record types for all these protocols, not just for e-mail. Of course, this wouldn't have been very practical unless the DNS architecture was made much more flexible than it actually is, to the extent that people could effectively invent their own non-standard record names.)
In fact, you're still likely to find such setups around: for example, universities, free software vendors and others who still distribute large files over FTP are likely to have a dedicated FTP host which is not the same as their main webhost. On the other hand, it has gotten to the point where, these days, one can pretty much expect every second-level domain name to have an A record and respond to HTTP request, even if only to redirect the browser to the actual webhost. Meanwhile, even as older protocols like Gopher have been outcompeted by the Web, the naming convention continues to be used for newer protocols: for example, the Wikimedia Foundation, who run Wikipedia, have svn.wikimedia.org, which, though it does respond to plain HTTP requests, is really meant to be accessed via the SVN version control client. They also have irc.wikimedia.org, which, while it responds to HTTP requests too, only redirects the browser to a page on Metawiki.
Meanwhile, for "vanity domains" that only serve web pages and maybe receive e-mail, there's little point in using the "www" prefix. It's still worth supporting it, though, if only as an alias, since many people (and some software) will assume that the prefix should be there. For example, www.vyznev.net redirects you to the corresponding URL without the "www". Mind you, I had to set up the redirect feature myself using mod_rewrite in an .htaccess file; with the default configuration provided by my webhost, the two hostnames would've just served up identical copies of the site. —Ilmari Karonen (talk) 12:44, 14 June 2008 (UTC)

KDE 4 in Kubuntu 8.04[edit]

Hello I have install Kubuntu 8.04 but it is in KDE 3.5 and I need KDE 4 how can make kde4 default .thanks —Preceding unsigned comment added by 202.125.143.74 (talk) 04:15, 12 June 2008 (UTC)

Heading added — Matt Eason (TalkContribs) 09:15, 12 June 2008 (UTC)
KDE 4 is a work in progress, and doesn't yet have all the features the average user expects -- I'd recommend waiting until KDE 4.1 or 4.2 is available. --Carnildo (talk) 23:05, 12 June 2008 (UTC)
Just type "sudo apt-get install kubuntu-kde4-desktop" in the terminal. 89.164.204.160 (talk) 15:10, 17 June 2008 (UTC)

Help me debug this program please!
[edit]

I’ve written the following program in C++ which is supposed to convert a given string of ASCII values into coherent text. Whenever I run the program in Borland C++ environment, I get the message “Conversion may lose significant digits” and the code that is emboldened in the following program gets highlighted. Please help me debug it!P.S—The output is supposed to be James Bond if I give the input as 10665771011153266797868.

Thanks. 117.194.226.179 (talk) 05:13, 12 June 2008 (UTC)

I see it- the "else" above the bolded line should be braced out. As it is, it's executing both the upper and lower commands on the inputs. JeremyMcCracken (talk) (contribs) 06:44, 12 June 2008 (UTC)

I'm sorry to say that doesn't help much. I just ran the corrected coding in my system, and the same error message appeared on the screen. By the way, I thought that if there are no braces after an "if" or "else" statement, it only executes the immediate next line?? 117.194.225.130 (talk) 07:18, 12 June 2008 (UTC)

I think the if-else is okay as written, but the code is hard to read because of the lack of spaces. The message “Conversion may lose significant digits” sounds like a compile-time warning meaning that you've implicitly converted from a wider integral type (the int returned from toupper) to a narrower one (char). It's a spurious warning in this case; the code is fine (at least that line is). It shouldn't prevent the program from running unless you've configured the compiler to treat warnings as errors. -- BenRG (talk) 09:19, 12 June 2008 (UTC)

I took the liberty of spacing it out a bit to improve the readability. Yes, toupper (and tolower as well) will return int sized values. There is an implied cast to fit the result into a char and you are being warned about the potential to lose accuracy with the cast. Astronaut (talk) 11:08, 12 June 2008 (UTC)

Your program flow is rather tangled and unclear. You should have five separate steps:
  1. Read in input string
  2. Convert input string to array of single digit integers
  3. Convert array to list of valid 2 or 3 digit ASCII codes (simplify your logic here - for your purposes, a 3 digit ASCII code always starts with 1; a 2 digit code never starts with 1)
  4. Convert list of ASCII codes to character string
  5. Apply capitalisation rules to decide whether each character in final output string should be upper or lower case and convert as required
Prorgram each step separately and insert debug statements to print out the result of each step before it is passed to the next step. After step 3 your array of ASCII codes should hold (106,65,77,101,115,32,66,79,78,68) and after step 4 your string before you apply capitalisation rules should be "jAMes BOND'. And adding comments will help you to clarify your own thinking as well as being a big help to anyone else who is expecetd to read your code. Gandalf61 (talk) 11:42, 12 June 2008 (UTC)
And while on the subject of improving your program, you could use more descriptive names for the variables. Astronaut (talk) 13:05, 12 June 2008 (UTC)

Google Maps error[edit]

I'm using Firefox 3.0 on Ubuntu and when visiting Google Maps there are always specific tiles that it cannot display and instead says the "we don't have maps at this zoom level" error message, no matter what my zoom level is. The tiles changes from use but there seems to be a pattern for it, whether it be diagonal or skip one or things like that. What could be causing this? --antilivedT | C | G 08:44, 12 June 2008 (UTC)

Ad blocker? As far as I can tell, the "we don't have maps at this zoom level" can mean either "we don't have maps" or "there was an error downloading the image". --Carnildo (talk) 23:07, 12 June 2008 (UTC)
I get that sometimes (Firefox 2 on Windows XP); I just thought it was Google Maps screwing up. I'd noticed that G-maps and Wikimapia would lose map background at the same time (Wikimapia uses G-maps for its backgrounds). JeremyMcCracken (talk) (contribs) 01:05, 13 June 2008 (UTC)

mountmgr.sys[edit]

what dose it mean when your lap top says " the file mountmgr.sys is corrupted —Preceding unsigned comment added by 12.35.45.66 (talk) 14:26, 12 June 2008 (UTC)

mountmgr.sys has a link to [[1]] where one of the posters point to http://support.microsoft.com/?kbid=297185 (I get Network timeout on the microsoft page). AFAIR, I have never seen this error myself. file mountmgr.sys is corrupted may have more information as well. Please give us more information on your case. If you have any questions, please do not hesitate to ask. Thank you. Kushal (talk) 00:39, 13 June 2008 (UTC)

SATA HDD for this board?[edit]

Hi, Does this motherboard support SATA HDD? [2] —Preceding unsigned comment added by 59.91.254.43 (talk) 14:27, 12 June 2008 (UTC)

I think not. It would have appeared under "Internal I/O Connectors". -- Meni Rosenfeld (talk) 14:49, 12 June 2008 (UTC)
Looking at the image, I don't see any SATA connectors either. Useight (talk) 15:43, 12 June 2008 (UTC)
Ditto those above, I don't see any SATA connectors in the picture. JeremyMcCracken (talk) (contribs) 18:10, 12 June 2008 (UTC)

Quantum computer hardware for videogames?[edit]

Quantum computer technology can one day be used for videogames right? ScienceApe (talk) 16:36, 12 June 2008 (UTC)

Maybe. What makes you think they would have any advantages over traditional computers for this purpose? -- Meni Rosenfeld (talk) 16:40, 12 June 2008 (UTC)
I have no idea. Do they have advantages for videogames? ScienceApe (talk) 18:39, 12 June 2008 (UTC)
None that I can think of. -- Meni Rosenfeld (talk) 19:21, 12 June 2008 (UTC)
Graphics won't be processed faster? ScienceApe (talk) 20:51, 12 June 2008 (UTC)
I don't think there is any evidence that they will. Quantum computers don't really do anything "faster", they just have a wider range of basic operations available. They are famous for being able to factor integers effectively, but I see no application of this for gaming. -- Meni Rosenfeld (talk) 21:00, 12 June 2008 (UTC)
Maybe, Ray tracing (graphics) could benefit from the wider range of basic operations if games (or their underlying packages) were modified to take benefit of the options quantum computing gives. Kushal (talk) 22:04, 12 June 2008 (UTC)
Everything is possible. But last I've heard, the largest computation performed by a real quantum computer was factoring the number 15. This means that by the time these mature, classical computers will probably have developed to a degree that any further improvement will not be distinguishable by our feeble eyes. The bottom line: Don't hold your breath until quantum computers come along and turn your gaming world around, it's not going to happen for a long while, if at all. -- Meni Rosenfeld (talk) 22:20, 12 June 2008 (UTC)
A lot of newspaper and magazine articles have created the misimpression that quantum computers would run classical algorithms faster than classical computers do. As Meni Rosenfeld said, they don't; what they do is run a larger class of algorithms called quantum algorithms. For certain specific problems, the best known quantum algorithm is significantly more efficient than the best known classical algorithm. But for the vast majority of problems the best known quantum algorithm is the classical algorithm, and in those cases you're better off running it on a classical computer. I think there are fast quantum algorithms for some path-finding problems which might be relevant to computer game programming, but it's hard to believe that you'd get any practical increase in frame rate from offloading such stuff to a quantum coprocessor. I don't think any known quantum algorithms would help with 3D rendering. Even if we all had quantum coprocessors in our machines right now, they probably would see very little use. -- BenRG (talk) 23:40, 12 June 2008 (UTC)
It seems a fair bet that, if quantum computing ever works in practice at all, it will, somehow, be used for games. Or porn. —Ilmari Karonen (talk) 12:49, 14 June 2008 (UTC)

open-source free maintenance tools[edit]

I need two of them:

-first a program that monitors my connection to the internet and tells me what program is connecting to what port, what speed, etc.

-the second is to find duplicated files on my HD.

GoingOnTracks (talk) 16:39, 12 June 2008 (UTC)

If you are on Tiger or Leopard, you can use iStat for the first one. Kushal (talk) 00:30, 13 June 2008 (UTC)
No, I am on Windows :(. GoingOnTracks (talk) 01:11, 13 June 2008 (UTC)
I know there are core utils for Linux that can be used to test for duplicate files. I'd use something like:
find . -exec cksum {} \; 2>/dev/null | sort >~/fileCksums.txt
Then look for duplicate checksums on consecutive lines in ~/fileCksums.txt. Maybe try doing that with Cygwin or something. --Prestidigitator (talk) 07:14, 13 June 2008 (UTC)
I'd recommend md5sum (or sha1sum) over cksum, just to minimize the (admittedly small) chance of false positives. Also, xargs would probably be more efficient than -exec. Oh, and you don't want to checksum directories. And the non-duplicates can be filtered out with uniq. So:
find . -type f -print0 | xargs -0 md5sum -b | sort | uniq -D -w32 >~/dups.txt
Note that some of the options I've used are GNU-specific (but if you use Linux or Cygwin that's what you'll have). For sha1sum, replace -w32 with -w40. —Ilmari Karonen (talk) 13:07, 14 June 2008 (UTC)

Escaping stuff in bash[edit]

I have a path like:
$ MYPATH="/some/path with spaces in it/somefile.dat"

I'd like to create the directory in another spot and copy the file there (I'm aware there are easier ways to do this but this is for illustrative purposes only). How do I properly escape everything in the following:
$ mkdir -p /home/me/$(dirname $MYPATH)

As-is, dirname gives an error ("extra operand 'with'"). So I double-quote $MYPATH:
$ mkdir -p /home/me/$(dirname "$MYPATH")

Great, so now I have /home/me/some/path, ./in, ./it, ./spaces and ./with. If I play stupid and pretend like bash will know what I'm talking about (and stick double-quotes in, even though they then become nested):
$ mkdir -p "/home/me/$(dirname "$MYPATH")"

It actually works, but it doesn't seem like this is the correct way. I would have thought "/home/me/$(dirname \"$MYPATH\")" was the correct way, but the inner, escaped quotes end up becoming part of the directory name.

--Silvaran (talk) 17:46, 12 June 2008 (UTC)

Looks like bash is smart enough to use the parentheses that are part of the $(...) notation to nest the double quotes also. (Perhaps this is one of the reasons people like -- and invented in the first place -- this newfangled $(...) mechanism.)
Me, I've always used the universal (if perhaps a bit dated-looking) backtic mechanism for command expansion. I just tried this, and it worked:
mkdir -p "/home/me/`dirname \"$MYPATH\"`"
Here, I did have to escape the inner quotes, for exactly the reason you suggested. —Steve Summit (talk) 20:44, 12 June 2008 (UTC)
I've always wondered why good shells don't extend the nestable $() syntax to quoting as well. I agree that letting bash figure out what you meant with the double quotes is icky. In a script I'd do it in two steps:
mydirname() { dirname "$MYPATH"; }
mkdir -p "/home/me/$(mydirname)"
--Sean 00:20, 13 June 2008 (UTC)
There's nothing wrong with "/home/me/$(dirname "$MYPATH")" at all. It may be surprising that you can include a pair of quotes within a command substitution that's within a pair of quotes, but the recursive parsing necessary to correctly match them up is required by the unix standard[3] so it's a pleasant and portable surprise, not a bashism. --tcsetattr (talk / contribs) 00:32, 13 June 2008 (UTC)
Yes, there is nothing wrong with mkdir -p "/home/me/$(dirname "$MYPATH")". It is actually the preferred syntax. Although the bash documentation doesn't specifically deprecates the backtick syntax, many expert shell programmers discourage its use. --Juliano (T) 15:00, 13 June 2008 (UTC)

Search keys[edit]

In search engines, "" helps you to define your phrase and search for that exact one. I think ~ or maybe three of them or something has a similar function. I am wondering if there are any keys\symbols\etc that can be used to exclude things you are searching for to make the search more refined. Simply south (talk) 18:30, 12 June 2008 (UTC)

I don't think you can generalize search engine behavior. Google allows you to use to '-' character to exclude items. For example, I live in the state of Virginia, and when I search for state government information, I used to get a lot of results for West Virginia as well, so I use something like Virginia Lemon Law -"West Virginia", which will filter out the WV results. --LarryMac | Talk 18:57, 12 June 2008 (UTC)
Also, the ~ is used (by Google) to search for synonyms, this is described here. That Google Guide site is packed with great tips for getting the most out of Google searches, and I'll repeat, you can't necessarily apply those tips to other search engines. --LarryMac | Talk 19:14, 12 June 2008 (UTC)
Thank you very much for these. Simply south (talk) 19:42, 12 June 2008 (UTC)

You can preface a search term with a minus sign. Some engines, which includes google I think, use boolean words (AND, OR, and NOT), which you can also use. JeremyMcCracken (talk) (contribs) 21:25, 12 June 2008 (UTC)

There's also inurl:, intitle: and site:. There's a list of advanced operators here. — Matt Eason (TalkContribs) 12:46, 13 June 2008 (UTC)

USB-to-serial converters, and drivers[edit]

Does anyone know why, when you plug a USB-to-serial (RS232) converter into a Windows (or Mac) machine, it's useless until you obtain and install the proper driver, but when you plug the same converter into a Linux machine, it tends to work automatically? (Me, I'm zero-for-three with several different Mac and Windows machines and several different converters, and I think three-for-three with Linux.) —Steve Summit (talk) 20:33, 12 June 2008 (UTC)

My understanding has always been that since the Linux community can't trust hardware manufacturers to supply Linux drivers for their products, they write their own drivers for all common devices and include them with the kernel. -- Meni Rosenfeld (talk) 20:46, 12 June 2008 (UTC)
Windows enjoys a large enough market share to force hardware manufacturers to write their own drivers. I don't know whether Microsoft does or should include every single device driver written for its operating system with its default installation. Some say that the modern operating systems are already bloated as they are. Kushal (talk) 00:26, 13 June 2008 (UTC)
I suspect Linux is able to support these devices out-of-the-box not because it has 2,574 handcrafted individual drivers bloating the kernel, but because it has one generic driver which supports not only the 2,574 USB-to-serial converters already out there, but also all the ones that haven't even been built yet. See also Juliano's answer just below. —Steve Summit (talk) 15:56, 14 June 2008 (UTC)
If you haven't thrown out your old RS-232 devices by now, you must be a unix geek anyway. You refuse to get on the upgrade treadmill? You will be punished... --tcsetattr (talk / contribs) 00:59, 13 June 2008 (UTC)
That is a mystery. The USB standard defines a set of "device classes" that have well-defined behavior. One of them, class 02h, is USB communications device class. The specification clearly defines how a "serial port" is implemented. This is an open specification, see http://www.usb.org/developers/devclass_docs/usbcdc11.pdf, sections 3.3.1 and 3.3.2, and the rest of the document (a USB-to-serial converter is trivially implemented, it is just a very small subset of this specification).
USB intends to provide an extensive protocol that supports many kinds of devices without requiring specific vendor drivers. Check the first paragraph of Universal Serial Bus: "[...] and allowing many devices to be used without requiring manufacturer specific, individual device drivers to be installed."
This has nothing to do with the Linux community not trusting hardware manufactures to supply drivers (although this may be true for other situations, it is not the case for USB CDC). It is part of the USB standard, just like mass storage is (this allows you to plug a USB flash drive into your computer and it just works). Windows doesn't provide a USB CDC driver because whatever reason Microsoft doesn't want to. --Juliano (T) 16:21, 13 June 2008 (UTC)
I was pretty certain there was a generic (device- and vendor-independent) specification; thanks much for the confirmation and the references.
(It's interesting you mention mass storage, because every time you plug a different USB thumbdrive into a Windows machine for the first time, it tends to do its little "searching for new hardware" dance, and it may even warn you that it "couldn't locate an appropriate driver and that your device may not work properly", although in that case, at least, it almost always does.)
One more reason to loathe Microsoft, I guess. (But it's not like I needed any more!) And it's certainly not a merely theoretical concern: it cost me and four other people a very expensive wasted hour of time last week, which is what got me thinking about it. —Steve Summit (talk) 15:56, 14 June 2008 (UTC)

sound cards; bipolar?[edit]

i know i should just go try it, but does anyone know beforehand any reason why a sound card shouldn't be able to feed the output from the audio out into the input to record it? Sort of a D to A to D thing? Gzuckier (talk) 20:53, 12 June 2008 (UTC)

Make sure you check your recording levels- if you have the "stereo mix" option, mute it or things could get ugly. Other than that, I'd think it would be okay. JeremyMcCracken (talk) (contribs) 21:27, 12 June 2008 (UTC)
Depends on if your sound card supports full duplex operation or not. If it doesn't, it can't record and play at the same time. --Carnildo (talk) 23:11, 12 June 2008 (UTC)

Sharepoint 3.0, Internal and External Sites[edit]

We just set up WSS 3.0 on a system running MS Small Business Server 2003 in a small office.

We would like to use sharepoint to host both internal project-management type workspaces, and an external site for clients and/or other professionals.

We are concerned about security, and ensuring that hosting a sharepoint site from here doesn't open up our network in any way.

How do you recommend we go about doing this? Are there any tips that can be offered? NByz (talk) 22:55, 12 June 2008 (UTC)

Just a word or two of warning without getting into the sharepoint technicalities. If your users are used to using MS file systems and networking (such as workgroups) to share files, get used to a major paradigm shift when implementing sharepoint. Be prepared for a lot of resistance. We implemented sharepoint at the corporate where I work and I can tell you although there are disk saving and file sharing advantages, sharepoint is quirky at best. It requires some intense setup and at least one full time person to support MS SQL server that it uses and often loses connection to, etc. I would seriously reconsider before using it. Sandman30s (talk) 23:14, 13 June 2008 (UTC)