X-No-Archive considered pointless

Why X-No-Archive doesn't help as much as people might think

Because of who we are and what we do, a number of people who post to newsgroups like to hide themselves; use anonymous names, try to stop archives from retaining messages and so on. I, personally, am open with who I am and what I am but this doesn’t mean I decry the attempts of others to maintain anonymity.

One such attempt is the use of X-No-Archive: headers in newsgroup postings. This tells google to not archive the message. Other archivers may also use this convention.

Note an important thing about this though: the line must either be a header or the first line of the message.

This has a major implications on follow ups.

By default, a followup to a X-No-Archive message is not itself X-No-Archived.

The new message will not have the message header in place (if the original had it in the header) and any message body with the line in would be quoted (typically with a > symbol) and appear after attributions line (eg “Fred Bloggs wrote:“).

For convenience, I’ll simply refer to this header as XNA in the rest of this document.

I have been taken to task, before, for not XNAing my responses to a XNAd message. It was even, once, criticised as being deliberate. As can be seen from above, a followup to an XNA message is not, itself, XNAd without additional action on my behalf. And that action is actually quite an onus:

  1. I don’t automatically put XNA headers in my post because I don’t want my articles non-archived (tin allows per-newsgroup headers if needed)

  2. If a message has XNA header then I don’t even see them becuase I have my newsreader set to hide them. Most definitely X- headers (a “do what you like” header with no standing in RFC-1036 beyond that) don’t get copied into reply messages (otherwise X-Face: and other stuff would also get copied)

  3. If the XNA line is the first line of the message then when I quote the response then the quoted line (a) has a > in front and (b) isn’t the first line of text, and so is ignored by google.

This means that if people want me to honor the request to propogate I have to

  1. notice it
  2. go out of my way to ADD text to my message
  3. go against my desire to want to archive my text

Basically, XNA has no standing in Usenet; it’s a flag for google (and other archivers that may decide to follow the standard). It’s no better than X-No-Ahbou: or any other of the potential 1000 X-No- headers that people may decide to create off the top of their head (X-No-War-In-Iraq: perhaps?)

Now, it’s been argued that people can have good reasons why they don’t want their posts to be available forever and that by failing to honour the XNA request is in the same general category as outing a person (but not as bad); outing as in identifying them as a member of some minority that may cause them distress elsewhere (eg a closetted gay person, a BDSM practioner, whatever).

I disagree totally. The only place XNA has any meaning is in the google groups archive (and potentially other archives which also use this convention). I effectively have an archive; my own personal newsserver never expires messages and so has messages going back to Jun 6 2001. It most surely never follows any X- header request because it’s a Usenet server and X- headers are meaningless to Usenet. I can (and have) quoted messages from the past when they are relevant.

Contrarily, I consider if offensive that other people wish to prescribe and force behaviour on me that I do not wish, especially when their are no RFCs or BCPs that cover the situation, and (indeed) where the RFCs say that the header in question has no meaning.

RFC-822 (the one that defines headers and which RFC-1036 - usenet - refers to) says:

    extension-field =
                    <Any field which is defined in a document
                    published as a formal extension to this
                    specification; none will have names beginning
                    with the string "X-">

     user-defined-field =
                    <Any field which has not been defined
                    in this specification or published as an
                    extension to this specification; names for
                    such fields must be unique and may be
                    pre-empted by published extensions>

That means NO field defined by any RFC standard will begin with X- which implies that the XNA header is a user-defined-field.

Another argument could be that this is a fairly well defined convention and that while the standards don’t specify, there is no reason to not accept this. My response is simply that HTML format messages and MIME encoded binaries are also well defined conventions. If it becomes the onus on the person following up to check for and maintain a non-standard header because of the wishes of the sender, then why shouldn’t the sender also be allowed to send HTML messages, with the onus on the reader to get software that can handle it?

This is why we have standards documents.

What this boils down to is that the XNA concept is fatally flawed as a means of keeping responses to messages out of google. Responses (either by accident or by design or by simple not-caring) may easily quote a XNAd message and not have the header itself. In my case it’s simple “not caring”; if someone posts a message then they must expect it to be quoted and followed up, and along the way it may easily be archived.

Anyone who relies on XNA to be no more than a request for their own message is deluding themselves as to the state of Usenet.

As an alternative, people concerned about their anonymity could create a free webmail address (eg hotmail, excite, yahoo, whatever) and use that in their postings. If they XNA their own message, then the originating IP address (their ISP) won’t appear in the archives and any followup (be it XNAd or not) will merely quote their webmail address. It would take a deliberate act to copy any NNTP-Posting-Host headers into the follow up message (or an act of “pilot error” - ie a mistake). Their anonymity is maintained. If they don’t want people reading their words in 5 years time, then don’t post in the first place. Once a message has been thrown out to usenet it has an indefinite life time. XNA does not, and can not, change this.