Slashdot Mirror


Internet RFC Series Turn 50 (circleid.com)

An anonymous reader writes: This week marks the 50th anniversary for the Internet "Request for Comments" (RFC) series, which started in April 1969 with the publication of RFC1 titled "Host Software" authored by Stephen D. Crocker. The early RFCs were meant to be requests for comments on ideas and proposals, says Heather Flanagan, RFC Series Editor. Today over 8500 RFCs have been published, ranging from best practice information, experimental protocols, informational material, to Internet standards. An RFC has been published to mark the fiftieth anniversary to include retrospective material from individuals involved at key inflection points, as well as a review of the current state of affairs.

18 of 43 comments (clear)

  1. Even back then by fat+man's+underwear · · Score: 1

    People who should have known better capitalized WORDS needlessly. Why would a WORD like host need to be capitalized? It's just a WORD like any other, it's not an acronym.

  2. You know the infinite monkey protocol RFC is it by satsuke · · Score: 1

    How can we be talking about internet RFCs without mentioning the most important one?

    The Infinite Monkey Protocol Suite

    How else would we get the collective works of Shakespeare with an infinite supply of monkeys/typewriters/bananas?

    https://tools.ietf.org/html/rf...

  3. Here's a suggestion by Red_Forman · · Score: 1

    Here's my submission for their RFC: block advertisers so they can't make the internet suck so much. Dumbasses.

    1. Re:Here's a suggestion by WCMI92 · · Score: 1

      PCH.com has introduced unblockable video ads. I have seen Behati Princesloo so many times I want to murder her.

      --
      Corporatism != Free Market
  4. RFC-conforming implementation by sinij · · Score: 4, Interesting

    If you want to strike fear into any seasoned developer's heart, specify "RFC-conforming implementation" as a hard requirement.

    While RFCs offer much-needed push toward standardization, modern RFCs tend to be overly complex and often contradictory. Even standard implementations do not achieve 100% conformance. For example, OpenSSL and its widely-used TLS implementation does not 100% implement all SHALLs of RFC 5246 (TLS v1.2). Conformance to RFC 5280 (PKI) is especially abysmal, I know of no solution that comes even close to meeting all of it.

  5. In RFCs, all caps means a special, defined term. by raymorris · · Score: 4, Informative

    > It's just a WORD like any other, it's not an acronym.

    You almost figured out why. In an RFC, putting a word in all caps means it's NOT "just a word like any other", the dictionary definition does NOT apply. All caps means "this term is being used to mean something specific which is defined elsewhere in an RFC".

    For example, you said:

    Why would a WORD like host need to be capitalized?

    In an RFC, "a word" means what it means anywhere else. "Cow" and "print" are words.

    However, "a WORD" might mean a data item with the same number of bits as the machine's data bus. On a 32-bit machine, a WORD is 32 bits.

    In a RFC about a text-based protocol, a WORD might be defined as "a sequence of one or more printable non-whitespace UTF characters". In which case "printk" would be a WORD, as would "starttls".

    All caps means "we have a specific definition for this term, and we're using the term in that specific sense here".

    Perhaps the most frequently used all-caps terms in RFCs are SHOULD, MUST, and MAY. Specifically, MAY and may need to be disambiguated. "May have security vulnerabilities" means vulnerabilities might exist. "MAY have security vulnerabilities" means it's ALLOWED to be vulnerable - it's specifically okay to do anything marked MAY. (In this instance perhaps any security weaknesses in that part of the algorithm don't matter because it's taken care of when the chunk is encrypted at a higher level).

  6. Personal favorite never implemented. by UnknowingFool · · Score: 1

    I really wish they had implemented RFC3514. Thanks, Obama.

    --
    Well, there's spam egg sausage and spam, that's not got much spam in it.
  7. The most important RFC by Burdell · · Score: 1
  8. Re:In RFCs, all caps means a special, defined term by fat+man's+underwear · · Score: 1

    Strange that none of your examples are capitalized in the RFC then. How very odd. Just face it, it's wrong. No one is going to think the word "host" means the guy who started a party in a RFC.

    For example, the word "conversation" in the RFC doesn't refer to two people chatting about the weather, does it?

    Or how about "bit"? It wasn't a drill bit, or a bit of money, or a bit part in a movie, was it?

    So according to your rules, they should have been capitalized.

    Just face it, engineers and programmers are terrible spellers and worse at grammar. How many times (manytimes?) do you see "backup" when they mean "back up"? Hint: one's a noun, the other a phrasal verb.

    "All caps means "we have a specific definition for this term, and we're using the term in that specific sense here"."

    No, that's what the glossary is for. No one capitalizes the word over and over and over. That's just nonsense.

  9. Where do you stay those defined? by raymorris · · Score: 1

    > For example, the word "conversation" in the RFC doesn't refer to two people chatting about the weather, does it?
    > Or how about "bit"? It wasn't a drill bit, or a bit of money, or a bit part in a movie, was it?
    > So according to your rules, they should have been capitalized.

    Where, exactly, do you see "bit" and "conversation" defined in that RFC, or any contemporary RFC?

    I didn't say "all caps means a technical term".
    All caps means a term defined in:
    1. That RFC or
    2. An RFC which is referenced

    For example, most RFCs reference RFC 2119, which defines SHOULD, MUST, and MAY
    http://www6.ietf.org/rfc/rfc21...

    So yeah those should not be capitalized per best practice, because there isn't a specific definition included which is being referenced.

    Since that's the very first RFC, fifty years old, and usage RFCs like 2119 hadn't been written yet, you probably *can* find some instances where the very first RFC did not comport to modern guidelines fifty years later. Neither bit nor conversation are examples, though.

    Indeed, BIT probably *is* defined in a special way for some physical interface standards, such as 100base-tx, which defines a HIGH bit as being over a certain voltage, and a LOW being below another voltage, with an undefined error band in the middle. Capitalizing BIT would have indicated that one needed to refer to a given specialized definition.

    1. Re:Where do you stay those defined? by fat+man's+underwear · · Score: 1

      "Where, exactly, do you see "bit" and "conversation" defined in that RFC, or any contemporary RFC?"

      The same place I see "host" defined: ie, nowhere. The word "host" is dropped in the first few sentences:

      "The software for the ARPA Network exists partly in the IMPs and partly in the respective HOSTs"

      So according to you, again, this means that "IMP" is actually a word that is different from a real imp, and not an acronym?

      Where is "host" defined in RFC 1?

    2. Re:Where do you stay those defined? by UnknownSoldier · · Score: 1

      OT: Re: RFC 2116 You gotta admit that's a pretty cool / funny email address:

      sob@harvard.edu

      =P

  10. Re:In RFCs, all caps means a special, defined term by munch117 · · Score: 1

    No one is going to think the word "host" means the guy who started a party in a RFC.

    In 1969, no one was going to think the word "host" meant a network node, unless it was very carefully explained to them.

  11. To celebrate, post here below... by LordHighExecutioner · · Score: 1

    ...your preferred RFC. For me it is RFC 2324.

  12. Re:In RFCs, all caps means a special, defined term by fat+man's+underwear · · Score: 1

    Yes, I am sure in 1969 that RFC 1 was going to be in the hands of many non-technical people.

  13. I would have posted earlier... by Sparky66 · · Score: 1

    ...but I'm using RFC 1149.

    https://tools.ietf.org/html/rfc1149/

  14. x86 originally 16-bit by raymorris · · Score: 1

    x86 assembly was originally written for a 16-bit CPU.
    "16-bit CPU" means a CPU with 16-bit words.

    When Intel introduced "32-bit processors" (CPUs with 32-bit words), they found that some programmers had defined things to be "word" when what they actually needed was 16 bits. Updating the setting of word would make a lot of software faster (where the value should actually be hardware word), but would break software that used word values incorrectly (assuming it would always be two bytes). Intel chose backward compatibility, not breaking software that had incorrectly assumed a word would always be two bytes.

    It's like assuming that a size_t is the same size as an int. Sometimes it is, today. Sometimes it's not. Intel catered to those who made the error.

  15. Re:In RFCs, all caps means a special, defined term by Cederic · · Score: 1

    a WORD is two BYTEs, or 16 bits

    Ah, I see you've gone for the 8 bit byte option. Nonetheless a WORD does not have to be 16 bits. Your nitpicking is wrong.