Slashdot Mirror

This book could be a good help. HTTP Developer's Handbook from SAMS gives you a great deal of information about the protocol

You can buy the book

OR

You can read the documentation of RFC 2616 - Hypertext Transfer Protocol -- HTTP/1.1 and save some money.

You can also read: HTTP/1.1 Specifications

Easy to understand and best of all FREE!!

I'm not an expert on the subject, but wouldn't that already be a problem with servers that do negative response caching? Granted, filling the negative cache with crap is less troublesome than the presumably more utilized positive cache, but I can't see how this would cause DNS servers to drop left and right.

You take the (oh... forgive me) Lotus Notes approach (I'm *not* a fan, but I understand this aspect of the software): it can be setup so the encrypt and decrypt happens transparently to the user between Lotus Notes servers. If you had something along this level between mail servers, then you might start getting into secure transmission of e-mail.

Yes, we call that STARTTLS in the SMTP world. However, that only provides hop-to-hop encryption, not end-to-end. There is no guarantee that each link in the delivery chain will employ encryption. While I'm not familiar with Lotus Notes, I daresay that once it forwards to SMTP it either stops encryption there, or uses STARTTLS for the first hop out with no control after that.

Please mod parent up, this is the most significant development in the fiaSCO so far.

For those who aren't familiar with the legend of Godwin's Law, cheack out How to post about Nazis and get away with it - the Godwin's Law FAQ. Although Godwin's Law is technically a USENET thing, it is frequently mentioned in regards to long /. threads, topics, and the like.

From RFC 2050:

2.2 Submission of Reassignment Information

It is imperative that reassignment information be submitted in a prompt and efficient manner to facilitate database maintenance and ensure database integrity. Therefore, assignment information must be submitted to the regional registry immediately upon making the assignment. The following reasons necessitate transmission of the reassignment information:

a) to provide operational staff with information on who is using the network number and to provide a contact in case of operational/security problems,

b) to ensure that a provider has exhausted a majority of its current CIDR allocation, thereby justifying an additional allocation,

c) to assist in IP allocation studies.

*Especially* for the less technically inclined it would be better to have as many different ways as possible to do something.

Will the end result of TMTOWTDI will be an obfuscated GUI contest?

SMTP based on RFC821 relies soley on the principle of:

1. User sends mail to target sender.
2. Mail goes to their SMTP server
3. Mail 'finally' arrives at the recievers SMTP server.

A much better solution would be based on user verification.

This in theory would work on the principle that the we are creatures of habit.

We all recieve legit Email from a small trusted group. Anything not based on the trusted group is potentially unwanted mail.

A verified Email transport would work like such:

1. Reciever builds a list of trusted Email senders.
2. The trusted list is uploaded to the recievers SMTP server.
3. A mail sender sends an email.
4. The senders SMTP server sends a message envelope to the recievers SMTP server. (contains just the senders smtp address).
5. The message is stored on the Senders SMTP server awaiting verification from the remote end.
6. The Recieving SMTP server checks the envelope against the user 'whitelist'.
7. If the sender is on the recievers whitelist - the RECIEVERS SMTP server confirms that this is legit.
8. Senders SMTP message delivers the message to the Remote SMTP server.
9. If the Email is NOT on the whitelist, the SMTP server sends a WAIT for further instructions message to the Sending SMTP server.
10. The user then can review the 'envelopes' and decide whether to recieve/remove the offending email.
11. If removed - the Recievers SMTP server sends a message back to the remote Senders SMTP server to say not to send.
12. If a response is never recieved by the Sending SMTP - the message is deleted after 30 days.

This has some added benefits:

• Legit mail recieves a higher priority. :)
• SPAM is not Blindly Sent but is only initiated at the 'Recievers request'.
• Network Traffic is cut considerably.
• The cost of storage of the SPAM is held at the remote end (SPAMMERS ISP).
• The Spammers ISP could legitimately then charge the SPAMMER for 'unsent' Mail storage. :)

This is only a thought -- and would need to round out the idea - however it seems feasable that this is possible.

Interested in others comments.

Most Spam filtering software already includes 'WhiteLists/BlackLists'.
Moving this into the SMTP transport at the server end seems the next logical and automated approach.

There are probably more MS-trained people out there looking for jobs than Java or Linux experts. A PFY with an MCSE who knows a bit of Visual Basic and .NET will probably work for less than a bearded *nix guru.

There is an easy solution to this. As we all know, port 80 will never be blocked, because otherwise how would we get our pr0n? The Web is the Net.

So, just tunnel everything through port 80. There is an existing protocol for this defined in RFC 3093 called FEP (Firewall enhancement protocol). Problem solved.

You Have Been Trolled. You Have Lost. Have A Nice Day

If a suitable subnetting scheme can be designed and is supported by the equipment concerned, it is advisable to use the 24-bit block (class A network) of private address space and make an addressing plan with a good growth path. If subnetting is a problem, the 16-bit block (class C networks), or the 20-bit block class B networks) of private address space can be used.

YHBT.

192.168.0.0/16 certainly does exist. The first three bits has not dictated the netmask for years. See RFC1817 for more information on this. Here's a relevant excerpt (emphasis added):

Classless Inter-Domain Routing (CIDR) ([RFC1518], [RFC1519]) is deployed in the Internet as the primary mechanism to improve scaling property of the Internet routing system. Essential to CIDR is the generalization of the concept of variable length subnet masks (VLSM) and the elimination of classes of network numbers (A, B, and C). The interior (intra-domain) routing protocols that support CIDR are OSPF, RIP II, Integrated IS-IS, and E-IGRP. The exterior (inter-domain) routing protocol that supports CIDR is BGP-4. Protocols like RIP, BGP-3, EGP, and IGRP do not support CIDR.

This solution will only work if it is exclusive of existing practice.

That was their first mistake.

Had they designed this as an SMTP Service Extension so that it could be integrated into existing mail servers, it would stand a chance of eventually being adopted. Sites could accept both, perhaps treating AMTP messages as SPAM-free for filtering purposes, until use was widespread enough to turn away messages that didn't have AMTP verification.

But to make an all-or-nothing stand will just doom the project. Sure, some rare people will want to run AMTP for cred and SMTP for the rest of their mail. Everyone else will wait for sendmail to create a service extension to do the same thing without having to rip out the plumbing.

They should adapt the routing algorithm of this protocol.

I actually believed you until you spoiled it with the laughable assertion that North Korea have access to the new protocol with QoS rather than the basic RFC1149 transport layer.

Nice to see that the 'Other Nations' are outside the US. And I'm glad its South (not North) Korea at No. 4, considering that Nuclear Engineering is at No. 7!

Those figures are a bit misleading. The North Koreans are taking the course through a FidoNET gateway in South Korea. The link across the border is by an RFC2549 connection.

When will the U.S. learn, and stop educating its enemies?

Nope. L is for Law. But here're the links.

Jargon File Listing for Godwin's Law, and How to post about Nazis and get away with it - the Godwin's Law FAQ (quoted below).

6. "Hitler!" Ha! The thread is over!

Nope, doesn't work that way. Not only is it wrong to say that a thread is over when Godwin's Law is invoked anyway (Usenet threads virtually always outlive their usefulness), but long ago a corollary to the Law was proposed and accepted by Taki "Quirk" Kogama (quirk@swcp.com):

Quirk's Exception: Intentional invocation of this so-called "Nazi Clause" is ineffectual.

Sorry, folks. Nice try, though.

Yeah, my legs used to get tired all the time. Then they upgraded to carrier pigeon and man was that fast. Where do you think "no carrier" came from?

i believe that's not microsofts invention, i think it's an RFC.

should get into moz sometime.
:D

Well, here is the obligatory Unix Shells FAQ in case you want to look it up. Then you can find the one that best suits your needs.

For those who might not have seen it:
[T]csh Programming Considered Harmful

I read the reasons not to use csh for scripting after I had spent about 10 years or so with csh from BSD days.

Even though I now run bash for an interactive shell (tcsh would probably be just as good with emacs style command line editing), I'm careful to limit myself to plain Bourne shell for scripts, since those are more portable than anything - I think that's why autoconf is successful.

The modern languages are good for serious scripting, as a glue language, for projects where you're willing to port the language and any needed libraries to the new platform.

But for a gopher build script, poking its head up in the dark of an unknown flavo[u]r of UNIX, sh is the True Path.

I know Apple is just a desktop OS, but remember that it's a powerful platform usable for literate users and not just for mom and grandpa -- myself including. I'm glad that they finally realized that. I have only one questions: was it only recently that they read Tom Christiansen's great article about csh-like shells? Seriously, it was published in 1996. Heaven't they read it before? Haven't they known about those issues which Tom points out before? I know Apple doesn't have a strong Unix background (hell, Apple doesn't even have a CLI background), still I find it somehow strange. Great news though.

According to RFC 821 (sections 4.3 and 4.2.2), the server can respond to new connections in with a 220 ("let's dance") or a 421 ("go away, I have a headache") response. Not a 554 ("you're lousy in bed") code.

You're citing an out-of-date RFC. 821 was superseded by RFC 2821, which makes it clear that 554 is a valid connection-opening response, to indicate that mail service is not available. (Indeed, 2821 spells out two codes for use at connection establishment -- 220 to accept, or 554 to reject access.) AOL is correctly using 554 to indicate that it will not provide mail service to your IP address.

A 4xx code would be improper in this case. 4xx codes indicate temporary failures. They mean that the client should queue its messages and retry them later, rather than returning a bounce message to the sender. That's not what is intended here -- the server doesn't want you to retry, it wants you to not try. A 5xx error code is correct.

I look forward to the day where this can happen you, too.

"As a Usenet discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches one."

RTFF (Read The ...er... Fine Faq)

An impropperly formatted response, like "2/30/2003", would probably get people's attention.

From RFC 958: NTP timestamps are represented as a 64-bit fixed-point number, in seconds relative to 0000 UT on 1 January 1900.

http://www.perthshirebigtreecountry.co.uk/Bigtree Fortingall.htm Or just use %20 instead of a space...

Thanks for the original link, and the instructions on how to get to the desination.

BTW, RFC2616 (HTTP/1.1) recomends the striping of white space from URLs. Not that it is often followed.

For those very few people left here who don't know what I'm talking about...

Godwin's Law is a natural law of Usenet named after Mike Godwin (godwin@eff.org) concerning Usenet "discussions". It reads, according to the Jargon File: As a Usenet discussion grows longer, the probability of a comparison involving Nazis or Hitler approaches one.

I did not get further than "Standard Brewing" before I thought of RFC 2324, namely the Hyper Text Coffee Pot Control Protocol.

No need to make a new standard. RFC 1149 already defines "A Standard for the Transmission of IP Datagrams on Avian Carriers."

I have yet to witness a compiler that did anything even remotely clever with register usage, much less correctly identify where the performance hot-spots were.

The fact that compilers don't do a great job of optimizing register usage doesn't mean the register keyword is useful. The (correct) statement was that most compilers have ignored it for some time. It doesn't matter if your recommendation to the compiler is a good one if the compiler ignores you on the grounds that it thinks it knows better.

On a lighter note, while googling to see how the compiler I use the most (gcc) handles the register keyword, I found this gem.

The resulting program will contain a request, on creation of the variable that it be placed in a cpu register, the operating system may ignore or honour this request.

Eh? I wasn't aware that the OS might decide to rewrite my program. If it can do that, can I get a Linux kernel patch that will automatically fix all the bugs in my software? It'd be really cool if it would also add all of the features I haven't gotten around to implementing yet...

The lines are ready and the technology works according to this RFC:
RFC 3251 (rfc3251) - Electricity over IP

13mm is tremendously below the minimum bend radius spec. As per the Data Communications Cabling FAQ:

"According to EIA SP-2840A (a draft version of EIA-568-x) the minimum bend radius for UTP is 4 x cable outside diameter, about one inch. For multipair cables the minimum bending radius is 10 x outside diameter.

SP-2840A gives minimum bend radii for Type 1A Shielded Twisted Pair (100 Mb/s STP) of 7.5 cm (3-in) for non-plenum cable, 15 cm (6-in) for the stiffer plenum-rated kind.

For fiber optic cables not in tension, the minimum bend radius is 10 x diameter; cables loaded in tension may not be bent at less than 20 x diameter. SP-2840A states that no f/o cable will be bent on a radius less than 3.0 cm (1.18-in).
...
Some manufacturers recommendations differ from the above, so it is worth checking the spec sheet for the cable you plan to use."

Ahh, the ever-popular patch pumpkin approach. Actually having the software physically stored on the token is a novel innovation, though. Have you thought of patenting it? (Sadly, you probably could...)

If I were in the original poster's shoes, I'd write a batch file that'd just make a lockfile on the same share where the program or data is located. Check for the lock; if it's not there lock the file and run the software. If it is there, pop up a warning. The lockfile should contain a user (or machine) name and a timestamp so you know who currently has the module locked. That way you can beat them up when they lock the file and go on vacation for a week. :-)

(Of course, the proper answer is to dump the buggy software. But I've worked in enough small offices to know that proper isn't always feasible...)

Regarding fonts not having copy rights, can you cite references for this?

Copyright FAQ, question 3.3

SMTP over TLS is mostly for encrypting mail so that eavesdroppers can't read the plain text.. It also allows servers to authenticate each other, but that doesn't help much... especially when the email has gone through multiple hops.

What you need to do is to (in some way or other) authenticate the sender. -- a completely different issue.

Working out the details of an appropriate certificate policy is not trivial, though.

The grotesque thing is that they probably go through a lot of hassle to prevent external access to these same IP numbers, thus defeating the whole purpose of having a public IP space in the first place.

Here's something kinda ironic. Presumably they don't move to network 10 because of the cost of reconfiguring every single system in their network. But when they go to IPv6 (it's gonna happen eventually), they'll have to do this anyway, and that would be the right time to go to a private network space. The same logic applies to all the companies that hang on to big chunks of the address space even though having it is a headache. So even as IPv6 eliminates the address space problem, it will free up big gobs of address space!

"Pacific Bell Internet Services jumped into the contentious music-downloading fray late Wednesday, filing a lawsuit against the recording industry and questioning the constitutionality of the industry's effort to track down online music sharers." (emphasis mine)

Joyce's Law: As a US lawsuit goes on longer, the probability of its constitutionality being challenged approaches one. :-)

The replacement for pkzip should be gzip. Not only is it specified in the open via rfc but it's implemented in internet explorer and friends.

Motif Faq

Subject: 2)* Is the Motif source code publically available?
[Last modified: Jan 02]

Answer: On May 15, 2000 the Open Group released the Motif source code for
Motif 2.1, using a public license, to the Open Source community. On January
29, 2002, Open Motif 2.2 was released.

okay, but i'm still left with questions. note these are not snide critisism cause I'm ignorant. I just want to understand.

why cant you just connect to the work mailhost or to comcasts mail host.

what has sending your e-mail out directly bought you?

And what does mobility have to do with anything here. that is, cant you see the comcast mailhost from anywhere and access it via an smtp connection that uses a password even if its not on their net?

if not then, mail programs (like mail.app) are happy to let you select the outgoing mail server so again its no big deal to switch from one to the other when at work/home. Am I missing something?

and finally my comapny recently started blocking port 110 connection FROM mailhosts outside the local network. thus you can only get mail sent to you through the companies mailhost. (they did this to force all e-mail to go through a virus sniffer on their host). Would this cause problems for sendmail?

yes I realize its outgoing but presumably it also gets info sent back regarding the success of the mail delivery.

Very good advice. I know from experience. I injured myself at the end of December running 10K in snow (very bad case of ITBS). I still had chronic pain in March. I've been physio ever since. I started running again in May with 3x 1km/wk, adding 500m per week since. I'm up 4x 2km a week, and it's slow progress. The physiotherapist has me doing all kinds of strengthening exercises to get my legs strong enough for real running (apparently my inner quad isn't strong enough to stabalise me knee properly), and lots of stretching too.

Going slowly when starting out running can't be emphasised enough. Keep to the rule of 10s: only increase one of speed, distance or frequency by no more than 10% per week. Start of slowly, even if you feel like you can push harder and that this approach will take months to get anywhere useful. It does take months. I hope to be hitting 10km again by late November, iff my injury has healed and doesn't recurr. I'm not old either - I'm 28. I spent my early 20s cycling everywhere, include 5 miles each way to work (I didn't own a car). Then I spent 3 years sitting on my arse working from home. It's depressing how quickly and easily the body gets weak.

Stretch *after* exercise every time. Get good shoes - NOT Nikes. Ask for advice on rec.running - lots of trolls, but lots of clued in people too. Check out Ozzie Gongang's rec.running FAQ.

Everything in society is not about terrorism, yet it is the overwhelming topic. I want to make and addendum to Godwin's Law:

Every slashdot discussion will eventually mention terrorism.

You're not thinking this through.

I run a gateway on it and advertise a server (B) behind this in the AEIOU space.

How? (to be more precise: Which DNS record do you have in mind?)

On the way back out, the gateway flips the stuff around, burying the internal address (B) in an IP option and sets the source to itself (A).

That's only one direction. The other direction is that your inner host may have to talk to another AEIOU host. The gateway can't know which address extension to add, so the client will have to do that. *Poof*, you need new client/server software. You have to be able to enter the full target address or the app has to be able to use extended DNS resolution. Coincidentally, these are also practically the only things which need to be changed for an application to support IPv6. The gateway requirements are practically the same, too.

The AEIOU addressing scheme suffers from the same problems which hold back IPv6. Do you really think it would be a big deal to route IPv6 if there were applications for it? If all else fails, just use automatic 6to4 tunnelling. Gives you exactly the same advantages and overhead as AEIOU, but when IPv6 becomes the norm, you can just leave out the tunnel and switch to native IPv6 routing without further modifications.

No, fast switching is alive and well:

http://www.cisco.com/en/US/products/sw/iosswrel/ps 1831/products_configuration_guide_chapter09186a008 00ca6c8.html">http://www.cisco.com/en/US/products/ sw/iosswrel/ps1831/products_configuration_guide_ch apter09186a00800ca6c8.html

http://www.networkcomputing.com/902/902sp2.html

http://www.cisco.com/univercd/cc/td/doc/product/so ftware/ios121/121cgcr/switch_c/xcprt1/xcdipsp.htm

http://www.faqs.org/faqs/cisco-networking-faq/sect ion-20.html

This FAQ explains how to access most of the internet using only a standard email client.

The above document explain how to access:

FTP
ARCHIE (deprecated)
FTPSEARCH (deprecated)
GOPHER (deprecated)
VERONICA (deprecated)
JUGHEAD (deprecated)
USENET
WWW
WWW SEARCH (using standard search engine like altavista, yahoo or google)
FINGER
WHOIS
[...]

All these protocols can be accessed via email, according to the FAQ. The FAQ has been around for a long time. This explains why many (most) involved protocols are now deprecated. I used this faq in the early '90 and I don't know how it works now. At the time, it was great. The last update is 2002/04/16.