Slashdot Mirror

We tried to deploy it but it kept eating our RFC1149 packets.

--hickmott

and what ever happened to the ipv6 mac address privacy fiasco?
It was always (mostly) a non-issue. With IPv4 you could:

1) Use manually assigned addresses. Fiddly and error prone.

2) Use server assigned (DHCP). Flexible but can be overkill.

With IPv6 you have a third option: (roughly) just use your MAC address the host part of the address. Instant simple, unique address. This was always an additional option on what you had already.You could always set your address manually or by DHCP.

Decnet-V had exactly this scheme, and nobody said anything about it. And GUIDs were fairly uncontroversial. It's only recently folks have worried more about their online privacy.

Anyway, a lot of people made a fuss, and the IPv6 folks said 'Oh yeah we'll think about that'. And some wrote Internet drafts and others bitched on /.

Finally, the drafts became an RFC (3041) and some read it and said "verily this standard will safeguard our privacy, let us go forth and implement it". And others looked upon it saying "yeah, this is not news".

• 3com - no listings (no support???)
  
• Cisco Systems
  
• HP - no listings in network equipment
  
• Juniper Networks - OS support
  
• Linux IPv6 HOWTO
  
• Lucent - interesting
  
• NetBSD IPv6 docs
  
• Windows XP - Installing IPv6
  
• RFC 2492 - "IPv6 Over ATM Networks"
  

I've read C&B.

How is your firewall going to protect from a new Outlook virus, or the IE bug-of-the-month?

And if I'm running RFC3093 IP over HTTP, with SSL, your proxy is almost certain to be clueless. Even assuming I'm not using steganography... and SOAP might count. ;)

Sure, proxies are more secure; but they can't be perfectly secure... and now you need a new proxy for each new application, which is the mess I was complaining about.

You're not the only one that feels this way. Harrison Ford thought so too. As a long time reader and poster to alt.cult-movies (before the bladerunner group was incepted) Murray Chapman was *the* guru, and collected some great info in his FAQ. Google Groups scares me sometimes...here's the relevant info from my original post in Jan 1997:

Ridley Scott and Harrison Ford have *stated* that Deckard was meant to be a replicant. In _Details_ magazine Oct 92 Ford says:

"Blade Runner was not one of my favorite films. I tangled with Ridley. The biggest problem was that at the end, he wanted the audience to find out that Deckard was a replicant. I fought that because I felt the audience needed somebody to cheer for."

Modula3 is better. OK, why not write it in Modula3?
Time to bone up on Modula3

Cut and paste troll - as most of Metrollica's comments are. This one is from the remailer FAQ.

they can always offer a fixed number of copies to download, forcing you ro gegister.

• we have all kinds of quality problems. There's a shitload of web apps out there that break with transparent caches, one way or the other, and often in subtile ways. There's even an RFC about some of them.
• when metering traffic independently of the cache statistics we found that we actually did not save any bandwidth worth mentioning. The statistics for the caches of course say different, but interface counters don't lie :-)
• customer satisfaction goes down the drain. The reason is, even if there is no problem with the caches, people blame any problem with internet and web site availability on the caches - and thus on us.

Okay, I did.

RJE stands for Remote Job Entry and is described here.

With sufficient thrust, pigs fly just fine. However, this is not necessarily a good idea. It is hard to be sure where they are going to land, and it could be dangerous sitting under them as they fly overhead.

I've already posted a similar comment in this thread, but since I formulated it rather bad and not too many people seemed to notice I'll make another try. And this time I'll cut and paste from this site.

One of the problems that the Drake Equation produces is that if you take reasonable (some would say optimistic) numbers for everything up to the average duration of technological civilizations, then you are left with three possibilities:

1. If such civilizations last a long time, "They" should be _here_ (leading either the the Flying Saucer hypothesis---they are here and we are seeing them, or the Zoo Hypothesis---they are here and are hiding in obedience to the Prime Directive, which they observe with far greater fiqdelity than Captain Kirk could ever muster). -or-

2. If such civilizations last a long time, and "They" are not "here" then it becomes necessary to explain why each and every technological civilization has consistently chosen not to build starships. The first civilization to build starships would spread across the entire Galaxy on a timescale that is short relative to the age of the Galaxy. Perhaps they lose interest in space flight and building starships because they are spending all their time surfing the net. (Think about it---the whole point of space flight is the proposition that there are privileged spatial locations, and the whole point of the net is that physical location is more or less irrelevant.) -or-

3. Such civilizations do not last a long time, and blow themselves up or otherwise fall apart pretty quickly (... film at 11). Thus the Drake Equation produces what is called the Fermi Paradox (i.e., "Where are They?"), in that the implications of #3 and #2 are not terribly encouraging to some folks, but the two flavors of #1 are kinda hard to come to grips with.

An alternate version of 2 is that interstellar travel is far more difficult than we think it is. Right now, it doesn't seem much beyond the boundaries of current technology to launch "generation ships," which power systems. An
alternative is robot probes with artificial intelligence; these don't seem so difficult either. The Milky Way galaxy is well under 10^5 light years in diameter and over 10^9 years old, so even travel beginning fairly recently in Galactic history and proceeding well under the speed of light ought to have filled the Galaxy by now. (Travel very near the speed of light still seems very hard, but such high speed isn't necessary to fill the Galaxy with life.) The paradox, then, is that we don't observe evidence of anybody besides us.

Sex Trek III: The Search for Sperm

That was an actual porn flick!
The impression of Cpt. Kirk is so funny, I fastforwarded the sex scenes just to hear the dialogues!
Oh, and there's that thing where Sperm's ears become "engorged" whenever he is aroused.

For an extensive list of Star Trek inspired porn, check out Adult Comics

Thanks to MS, email has gone from a simple text message to a mixed bag of html, lines ending with an '=' and messages with unreadable attachments which repeat the body text (and are often larger than the message itself). Since asking politely hasn't done away with the crap, perhaps this will have to do.

Um, those would be examples of messages in MIME format. MIME is defined by RFC 1521. It predates Outlook, so son't blame MS if you aren't using a MIME-conformant mail reader.

It'd just make the "mainstream" IT community reject linux, and its users, even faster.

Which will not be as easy. My mail client (mutt) just displays everything with my settings and let me be the judge what to read and what to ignore.
Of course this assumes an RFC2822 enabled user...

PPPoE Info

If you look at the RFC for IP: http://www.faqs.org/rfcs/rfc791.html, and TCP: http://www.faqs.org/rfcs/rfc793.html, you'll notice that MAC address is not listed in either of these.

ARP (IP->MAC translation) occurs below IP and allows hosts on an Ethernet network to communicate. IP, TCP, UDP, etc, doesn't care about MAC addresses. Therefore NAT doesn't care about MAC addresses.

If you look at the RFC for IP: http://www.faqs.org/rfcs/rfc791.html, and TCP: http://www.faqs.org/rfcs/rfc793.html, you'll notice that MAC address is not listed in either of these.

ARP (IP->MAC translation) occurs below IP and allows hosts on an Ethernet network to communicate. IP, TCP, UDP, etc, doesn't care about MAC addresses. Therefore NAT doesn't care about MAC addresses.

An old PC with voice modem and vgetty would do the trick. vgetty can turn any PC with a voice modem into an answering machine. vgettty would record the message. Then some script could go through the call list of numbers you have, and play the message to them.

The more and more I search, the more it looks like at least in VA, this could be somewhat illegal.

The Virginia ID card and drivers license form state:

The information provided on this application is for DMV's record-keeping purposes and may be disseminated in accordance with 46.2-345.

46.2-345 states:

G. Any personal information, as identified in 2.2-3801, which is retained by the Department from an application for the issuance of a special identification card is confidential and shall not be divulged to any person, association, corporation, or organization, public or private, except to the legal guardian or the attorney of the applicant or to a person, association, corporation, or organization nominated in writing by the applicant, his legal guardian, or his attorney. This subsection shall not prevent the Department from furnishing the application or any information thereon to any law-enforcement agency.

The Department of Transportation is NOT considered a law enforcement agency, is it?. I'm sure this can seen differently by others.

If your VA license number is your SSN, it probably violates other information reporting laws also.

There is quite a bit of info online about the uses of SSN's that might clear up some misconceptions about the use of SSN's in drivers licenses numbers, but also raises questions about this "new" use that a drivers license number that happens to be your SSN. Is the new use of the your SSN been expanded and not IAW the 1994 Privacy Act?

Specifically:

The Privacy Act of 1974 (Pub. L. 93-579, in section 7), which is the primary law affecting the use of SSNs, requires that any federal, state, or local government agency that requests your Social Security Number has to tell you four things:

1: The authority (whether granted by statute, or by executive order of the President) which authorizes the solicitation of the information and whether disclosure of such information is mandatory or voluntary;

2: The principal purposes for which the information is intended to be used;

3: The routine uses which may be made of the information, as published annually in the Federal Register, and

4: The effects on you, if any, of not providing all or any part of the requested information.

The Act requires state and local agencies which request the SSN to inform the individual of only three things:

1: Whether the disclosure is mandatory or voluntary,
2: By what statutory or other authority the SSN is solicited, and
3: What uses will be made of the number.

If your truely concerned I suggest searching google for SSN (expanded) and pricay statement. There is a SSN faq here
I had a PA which did not use your SSN and now a VA which the SSN is optional.

I am with a admin group trying to integrate a couple hundred UNIX and Windows machines into a single login using an Active Directory server, which provides us with Kerberos authentication, and an LDAP directory. (This was mandated to us "from above") The kerberos authentication of course was easy, however there is hardly ANY information about actually using LDAP in a production environment.. we are trying to use the active directory LDAP server to provide the POSIX gecos and home directory information for the UNIX clients... however the default Active Directory schema does not include RFC2307

Probably the most frustrating part is if you go on google and look for help, you see people mentioning that this works, but never any specifics. I assume you are just using pam_ldap to grab a password crypt from an LDAP server (which is a secure as giving everyone read permissions on your shadow file).

I think the best solution is to use an LDAP server to host all the user information that is normally in /etc/passwd. This is possible in Linux and Solaris using the nss_ldap module which lets you add an "ldap" entry to your network switch file, and use ldap instead of /etc/passwd. It seems the best solution is Kerberos for authentication and LDAP for everything else, which Active Directory can provide, in a mixed-OS environment even.. but has anyone been able to successfully run nss_ldap against an AD LDAP server? (without using services for UNIX or other kludges) LDAP seems to be an integration nirvanna.. but without proper documentation I am afraid it will never see broader use..

100:1 ratio? On random data?
Considerations far more elementary than Shannon's limits rule out compression of statistically random data by even a single bit. Here's why:
There are 2^n bit strings of length n. Any compression method purporting to compress random strings (by even a single bit) must produce output of length at most n-1 for these 2^n inputs. But in that case the mapping is not unique, since there are only (2^n)-1 bit strings of length n-1 or less. (So decoding is not possible.)
Once every so often some "researchers" claim to have attained the holy grail of compression. Too bad we never hear of them again :(

From the comp.compression faq
this topic has generated and is still generating the greatest volume of news in the history of comp.compression
...
The advertized revolutionary methods have all in common their supposed ability to compress random or already compressed data. I will keep this item in the FAQ to encourage people to take such claims with great precautions

It's over here (Question 9, search for 'WEB, Gilbert').

Head over to comp.compression FAQ to read about compression of random data. Not exactly a new subject of controversy.

Their claims are 100% accurate (they can compress random data 100:1) only if (by their definition) random data comprises a very small percentage of all possible data sequences. The other 99.9999% of "non-random" sequences would need to expand. You can show this by a simple counting argument.

This is covered in great detail in the comp.compression FAQ. Take a look at the information on the WEB Technologies DataFiles/16 compressor (notice the similarity of claims!) if you're unconvinced. You can find it in Section 8 of Part 1 of the FAQ.

Their claims are 100% accurate (they can compress random data 100:1) only if (by their definition) random data comprises a very small percentage of all possible data sequences. The other 99.9999% of "non-random" sequences would need to expand. You can show this by a simple counting argument.

This is covered in great detail in the comp.compression FAQ. Take a look at the information on the WEB Technologies DataFiles/16 compressor (notice the similarity of claims!) if you're unconvinced. You can find it in Section 8 of Part 1 of the FAQ.

Yawn... see the comp.compression FAQ, compression of random data

Section 1.9 of the comp.compression FAQ is good background reading on this stuff. In particular, read the "WEB story".

Hey, if their algorithm works on random data, re-apply it to the output, and it will be compressed again. You can do this again and again, until only one bit is left!

Now, let's uncompress a 0 bit and a 1 bit. All software ever written and ever to be written in the future must come out, since there cannot be anything which compresses to anything else than a 0 or 1 bit, if compressed to a single bit.

Seriously though, the comp.compression FAQ is really worth a read, especially question #9.

Hey, if their algorithm works on random data, re-apply it to the output, and it will be compressed again. You can do this again and again, until only one bit is left!

Now, let's uncompress a 0 bit and a 1 bit. All software ever written and ever to be written in the future must come out, since there cannot be anything which compresses to anything else than a 0 or 1 bit, if compressed to a single bit.

Seriously though, the comp.compression FAQ is really worth a read, especially question #9.

Of course, you may use your own different definition of OOP. However, traditionally the inheritance and polymorphism are considered to be the essential parts of OOP.

What you're referring to is an "object-based" programming language, something like Modula-2. It supports object encapsulation, but it's not an OOP language yet.

Please read the comp.object.FAQ:

1.15) What Is The Difference Between Object-Based And Object-Oriented?

Object-Based Programming usually refers to objects without inheritance
[Cardelli 85] and hence without polymorphism, as in '83 Ada and Modula-2.
These languages support abstract data types (Adts) and not classes, which
provide inheritance and polymorphism. Ada95 and Modula-3; however, support
both inheritance and polymorphism and are object-oriented. [Cardelli 85, p481]
state "that a language is object-oriented if and only if it satisfies the
following requirements:

- It supports objects that are data abstractions with an interface of named
operations and a hidden local state.
- Objects have an associated type.
- Types may inherit attributes from supertypes.

Ethernet over cat...kinda like rfc 1149: A Standard for the Transmission of IP Datagrams on Avian Carriers

While this technology is cool, I think they should have put more money into R&D on RFC1149.

For those unfamiliar with RFC1149, here are the details: http://www.faqs.org/rfcs/rfc1149.html

Its probably a bit slower than barbed wire, but damn it... its more fun!

Time is an issue we rarely incorporate in our designs.

Excellent point. For exmaple, there's growing concern over data stored on "dead media"; enve stuff on readable media can be rendered useless by outdated proprietary file formats.

Some people joked about the Y10k problem a while back, but I think its quite possible that within the next century we'll be building systems with a mission lifetime of thousands of years.

Usenet: Sex, hamsters and duck tape

The collective wants you. Join Usenet, the largest information repository on the Internet.

Every day, thousands of articles are posted on Usenet, short for Users' Network. Since its inception in 1979-1980 at Duke University, an estimated terabyte (1012 bytes) or the equivalent of one billion pages of information and discussions have been generated in over 500 million posts by millions of people.

Usenet is essentially a global bulletin board. Like e-mail, participants read and post text messages, but it also allows anyone to contribute their experience and expertise by reading and replying to those messages and forming threads. For this reason, many people flock to Usenet solely to find instant-gratification answers to their questions.

Over 50,000 groups exist to cover topics including Rush Limbaugh and sex with hamsters using duct tape (no relation). Experts (and quacks) from around the world are more than willing to offer their advice for free, and there are plenty of people on Usenet who could use any knowledge you have.

To participate in Usenet, you will need a news reader program such as Forte Agent for Windows, Tin for AIX (available with your university account), or Yet Another Newswatcher for Mac. Netscape Messenger and Microsoft Outlook for Macintosh and/or Windows are competent news readers but lack some features. You will also need access to a news server, such as the one provided by the U of C or another Internet service provider.

Once you are online, you will need to choose a few news groups (such as alt.tv.star-trek). The easiest way to find a group is by browsing the Usenet newsgroup hierarchy which structures news groups by topic (for example, alt.tv.star-trek.voyager is for discussions regarding Star Trek: Voyager while alt.tv.star-trek is for Star Trek).

Most new users should start in a group under one of the Big Eight newsgroups which include "comp" for computer-related discussions (comp.os for example), "rec" for recreational topics (television, games, etc) and "sci" for academic discussions regarding science. The other large hierarchy, "alt," contains discussions on almost everything imaginable although "alt" can be somewhat chaotic since almost none

of the news-groups under it are moderated. In moderated groups such as those under "comp," article submissions are approved by a human being before they reach the actual newsgroup. This helps to reduce the volume of low-quality and redundant posts.

Users can find out about a particular group by lurking. Lurking, reading others' posts for a while to understand the general flow of ideas in the community, is a good way to acquaint yourself with the group. Another good idea is to read the FAQ (a list of frequently asked questions) from the newsgroup before posting since most of the common questions have already been asked and answered by others.

If you think you've come up with a question or a discussion topic not previously addressed, or would like to contribute to a discussion, be sure to read the FAQ for posting guidelines. For instance, discussion groups about current TV shows request spoiler space (a bunch of blank lines) before anything that may give away the plot.

There are a few other things to keep in mind when posting. The subject line should describe what your post or question is about and you should provide as many details as possible in the body of your message. "ALL YOUR BASE ARE BELONG TO US" as a subject line or contents of the body does not count. If you are requesting a specific file or piece of information, it is customary to put "REQ:" in the subject line of the article. Not all groups allow the posting of binaries, that is, any file that is not text. Those that do are often denoted by ".binaries." in their name and are nor particularly great for extended discussions since messages expire quickly, that is, they are deleted from the server because the attached files tend to be large.

If your news server suffers from a lack of interesting groups or fast expiry, there are alternatives. MailandNews.com, for example, offers free Usenet and mail service. For older postings, there are searchable archives of discussions such as DejaNews' Usenet archive.

Have fun but remember that other posters are mostly human. Don't say things that you wouldn't in real life. "Flaming" or being a "troll" (read: bastard) will get you plonked (permanently ignored) very quickly.

Hmm, hard to find a definitive source. But, some support for that assertion is here ("10Eh : 320x200 64k-colour (5:6:5)", "111h : 640x480 64k-colour (5:6:5)", ...) and here ("16 bit color depth is supported through several different bit arrangements, including 5-5-5 and 5-6-5.").

Some of us don't feel too kindly towards PayPal. And Amazon at least has a somewhat trusted name.

Scratch this. I just read the FAQ. They want $0.15 + 15% of the donation.

Sigh. I thought it was a good idea.

DISCLAIMER: I work for a corporation who is partnered with Amazon

John

"According to Mark Podlipec's XAnim site (May 10, 1999), he contacted Sorenson Vision to find out if he could license Sorenson Video for incorporation in the XAnim Unix X11 animation, audio, and video player. According to his Web site, Sorenson replied that Apple will not allow Sorenson to license Sorenson Video to others."

Furthermore, UDP for data is highly unreliable, and I wouldn't trust it across WAN's.

You're missing the point of UDP. UDP is just a *tiny* layer on top of IP, which adds a little extra information (basically the port number) so that the OS can deliver a packet to the right application. UDP can not be compared with TCP - it doesn't provide reliability and flow control, and it has absolutely no notion of a stream of data. If desired, these can be provided in the application layer (see UDP, TFTP, NFS, etc. etc.)

TCP is a reliable transport, but it's much, much more than that. First off, the fact that you're using TCP doesn't make the path between sender/receiver any more reliable. Your packets get dropped just the same as if they were UDP or any other protocol over IP. TCP provides reliability by retransmitting lost packets, but you knew that. It also provides flow control and congestion avoidance - this means detecting when the receiving end (and the router queues in between) are ready for more data, and throttling your transmission rate to match that capacity. It also means being "fair" to other streams sharing the same bottleneck(s). It does this by "backing off" the number of packets in flight, i.e. halving the congestion window, to reduce the data rate. These algorithms are a very active field of research - there's a *lot* more to TCP than meets the eye of a socket programmer.

When TCP loses a packet, that packet must be retransmitted. This is expensive because it means another RTT.

Anyhoo...

You can think of FEC for data transmission as being analogous to RAID 5 for storage. By adding one extra bit (the XOR of the rest of the word) you can lose any single bit and still know it's value. It's very simple. If the word is:

0 1 0 1

And I add an extra "parity" bit, where the parity bit is 1 is the number of ones in the rest of the word is odd, zero if it's even:

0 1 0 1 [0]

I can now lose any one bit (including of course the parity bit). Eg if I have only

0 1 X 1 [0]

Then I know the missing bit is a '0', because if it were '1' then the parity bit would be a zero.

Applying this to data transmission, you can see that by sending just one extra packet, you greatly reduce the chance of having to retransmit anything.

EG if I have to send you 10 packets over a link with 10% packet loss, there's a 65% chance that I'll have to retransmit one of those packets. (and a 10% chance that each retransmitted packet will have to be sent again, and so on).

However if I'm using FEC and I send one extra "parity" packet, then I only have to retransmit if TWO OR MORE packets are lost. The chances of losing TWO out of the eleven packets is only 30%, so you can see that for an overhead of 10%, I've reduced the number of retransmits by a factor of more than two! I hope those figures are right. I used this tool to calculate them. Of course there are a lot of knobs you can turn depending on how much overhead you can afford for the parity information, and what degree of packet loss you want to be able to tolerate.

Anyway, you can see that there are lots of possible improvements/alternatives to TCP - it's an old protocol and a lot of research has been done since RFC 793.

Compare your (ie America's) murder/rape stats to those in Europe and you may start to understand exactly why you don't want millions of people carrying guns.

Our problems with violence have more to do with economic disparity, lingering racism and segregation, and the war on (some) drugs, than with the legal status of firearms.

Within the US, there is a clear correlation between gun control and violence crime - states with strong gun control laws have more crime, states which respect the RKBA have less crime.

I really don't get this aspect of the US - even if you FEEL safer when carrying a gun, statistically you are in considerable danger.

Simply not correct. The statistics clearly show that those parts of the US that allow for the leagl concealed carrying of firearms have less violent crime, and that individual people with guns are best able to avoid being killed or injured by violent attackers.

Let me recommend the "Pro-gun FAQ", chock full of facts, numbers, and references.

Glad you mentioned this, because it reminded me how difficult it is to find good documentation about wiring Ethernet.

As a wise Slashdot reader once said, "You can't have too much overkill", so here is the wiring scheme shown 3 different ways. I hope it saves you the time of gathering it together yourself.

Slashdot doesn't allow the HTML PRE tag. Slashdot removes leading spaces, so I've used dots below. Another problem is that the lameness filter is lame. That lameness filter is definitely named correctly.

Use only Standard EIA/TIA T568B. This is also called the AT&T specification. T568A is NOT USED.

T568B:

When the hook of the RJ-45 Ethernet connector is underneath, pin 1 is on the left.

Pair 1 is pins 4 and 5, Blue and White/Blue.

Pair 2 is pins 1 and 2, White/Orange and Orange (Transmit Data + and -)

Pair 3 is pins 3 and 6, White/Green and Green (Receive Data + and -).

Pair 4 is pins 7 and 8, White/Brown and Brown.

/--T21 White/Orange
Pair2 \--R22 Orange
/--------- -T33 White/Green
/ /-R14 Blue
Pair3 \ Pair1 \-T15 White/Blue
\--------- -R36 Green
/--T47 White/Brown
Pair4 \--R48 Brown

Pin ColorPairName

1 wh/or 2 TxData +
2 or2 TxData -
3 wh/grn3 RecvData+
4 blu 1
5 wh/blu1
6 grn 3 RecvData-
7 wh/brn4
8 brn 4

I know this sounds crazy, but standard Ethernet uses ONLY pairs 2 and 3, for both half and full duplex. The other wires just sit there, unused. (It is possible to buy external adapters to use the other two pairs as a second 10- or 100 Megabit 100Base-T connection.)

The R1, T1, R2, T2 designations are for telephones. R1 is Ring 1 (the red wire at the telephone box). T1 is Tip 1 (the green wire at the telephone box). Ring and Tip are old names for the telephone wires, but if you talk to a telephone company installer, he or she will use those names.

Quite obviously, someone messed this up majorly, as in "How can we make this confusing?"

If you are new to wiring Ethernet start with the simple explanation at Johns Closet (as in wiring closet): Wiring: Color Codes, Terms, and Tools.

See the Leviton Do and Don't Guides

See the Wiring Guides at the Leviton Learning Center . See the Residential And Light Commercial Installation Practices (Tia-570 Compliance) [PDF file] guide.

Also see the wiring specs at FAQS.org: 9.0 Standard EIA/TIA 568 (Use ONLY T568B)

More information about wiring: Data Communications Cabling FAQ

Your local store will probably try to charge too much. Shop around for Ethernet cable and connectors. You need the real thing, cable marked "Category 5". Other cable won't work.

One last thought to those who are new to Ethernet networking. A Hub broadcasts all data to all computers. An Ethernet Switch sends the data only to the computer that where the data will be used. Therefore, switches are faster in cases where the network is sending data between more than one pair of computers at the same time.

--
Senator Biden (and Osama bin Laden) say that the Saudi government cannot continue without U.S. support: What should be the Response to Violence?

Glad you mentioned this, because it reminded me how difficult it is to find good documentation about wiring Ethernet.

As a wise Slashdot reader once said, "You can't have too much overkill", so here is the wiring scheme shown 3 different ways. I hope it saves you the time of gathering it together yourself.

Slashdot doesn't allow the HTML PRE tag. Slashdot removes leading spaces, so I've used dots below. Another problem is that the lameness filter is lame. That lameness filter is definitely named correctly.

Use only Standard EIA/TIA T568B. This is also called the AT&T specification. T568A is NOT USED.

T568B:

When the hook of the RJ-45 Ethernet connector is underneath, pin 1 is on the left.

Pair 1 is pins 4 and 5, Blue and White/Blue.

Pair 2 is pins 1 and 2, White/Orange and Orange (Transmit Data + and -)

Pair 3 is pins 3 and 6, White/Green and Green (Receive Data + and -).

Pair 4 is pins 7 and 8, White/Brown and Brown.

/--T21 White/Orange
Pair2 \--R22 Orange
/--------- -T33 White/Green
/ /-R14 Blue
Pair3 \ Pair1 \-T15 White/Blue
\--------- -R36 Green
/--T47 White/Brown
Pair4 \--R48 Brown

Pin ColorPairName

1 wh/or 2 TxData +
2 or2 TxData -
3 wh/grn3 RecvData+
4 blu 1
5 wh/blu1
6 grn 3 RecvData-
7 wh/brn4
8 brn 4

I know this sounds crazy, but standard Ethernet uses ONLY pairs 2 and 3, for both half and full duplex. The other wires just sit there, unused. (It is possible to buy external adapters to use the other two pairs as a second 10- or 100 Megabit 100Base-T connection.)

The R1, T1, R2, T2 designations are for telephones. R1 is Ring 1 (the red wire at the telephone box). T1 is Tip 1 (the green wire at the telephone box). Ring and Tip are old names for the telephone wires, but if you talk to a telephone company installer, he or she will use those names.

Quite obviously, someone messed this up majorly, as in "How can we make this confusing?"

If you are new to wiring Ethernet start with the simple explanation at Johns Closet (as in wiring closet): Wiring: Color Codes, Terms, and Tools.

See the Leviton Do and Don't Guides

See the Wiring Guides at the Leviton Learning Center . See the Residential And Light Commercial Installation Practices (Tia-570 Compliance) [PDF file] guide.

Also see the wiring specs at FAQS.org: 9.0 Standard EIA/TIA 568 (Use ONLY T568B)

More information about wiring: Data Communications Cabling FAQ

Your local store will probably try to charge too much. Shop around for Ethernet cable and connectors. You need the real thing, cable marked "Category 5". Other cable won't work.

One last thought to those who are new to Ethernet networking. A Hub broadcasts all data to all computers. An Ethernet Switch sends the data only to the computer that where the data will be used. Therefore, switches are faster in cases where the network is sending data between more than one pair of computers at the same time.

--
Senator Biden (and Osama bin Laden) say that the Saudi government cannot continue without U.S. support: What should be the Response to Violence?

I'd have to characterize that as a Nutscrape problem [..] Ditto for Eudora.

It's a Rogers problem: it breaks RFC 822.

(Gratuitous slams at Microsoft don't constitute an argument. Besides, ESR would be surprised to find that MS owns fetchmail. :-) )

I'd have to characterize that as a Nutscrape problem [..] Ditto for Eudora.

It's a Rogers problem: it breaks RFC 822.

(Gratuitous slams at Microsoft don't constitute an argument. Besides, ESR would be surprised to find that MS owns fetchmail. :-) )

This is less than impressive: most people know an email address is a "@" , and a local-part is a legal userid or alias on the underlying system. A mailbox containing an @ is prohibited... See http://www.faqs.org/rfcs/rfc822.html for details.

And the workaround for Rogers' goof provided their customers is to use the new account wizard to create a new acount/profile with a name that shows a @rogers.com externally, and as @rogers.com@rogers.com internally. This approach leaves the users with one account/profile containing all their old mail, preferences and bookmarks, and a new account with the mail missing.

I wonder how many support calls come in entitled "you stole my email"?

From information available PPPoE, even though wrapped, is still has a smaller header size than DOCSIS. PPPoE has a header of 24 bytes and a payload of between 0-1492 bytes, while DOCSIS has a header of 31 bytes at minimum ( with a max of 266 bytes ) and a packet size 0-1500 bytes. Which if taken to maximum with maximum packets sizes all the time is 98.4% for PPPoE and between 98.0% and 84.9% percent depending on the header size. (As packet sizes decrease the difference between the two systems INCREASE). This means even with PPPoE DSL lines you are using more bps for real data than on a DOCSIS cable modem system. Of course this analysis assumes that error rates are non-existant for both and that traffic comming from both nodes is at the same level. (The traffic levels is a safe assumption in this case, but error rates of DSL would probably be less than DOCSIS since it's collisions do no occur within the system, but at the head node. DOCSIS occurs at transmission because of contention on the line.) You webserver and mail will probably not perform as well if they were using the full line speed of the cable modem, but this is due to limitations on the upstream speed and not the frame size or frame overhead. If the server isn't being constantly slammed there will be little to no difference. But since you can never tell until you run the system you're millage may vary.

Info about: PPPoE DOCSIS

I'm investigating which precisely patents are these (some of them have patent numbers, some don't).

One of these is U.S. Patent 4,558,302 on LZW compression that the V.42bis standard uses, owned by Unisys Corporation. Unisys's policy since mid-1999 has been not to license the LZW patent to free software projects.

requiring Compression adds a substanial amount of work

I don't think we're likely to see V.42bis compression support in free software for several years, as the V.42bis standard requires the use of patented LZW technology, and Unisys refuses to license LZW for use in free software. Feel glad that patents last only 20 years after filing and not forever like copyrights.

Anonymous Cowards here aren't, because their IP addresses are still subject to subpoena, and there's a 2 week long window where Slashdot stores the IP address as an MD5 hash, which can be easily defeated.

How can MD5 be "easily defeated"? I was under the impression that MD5 is a one-way algorithm, so that the original IP address couldn't be recovered from the logs.