The issue here is not how to define profanity or civility. The issue here is not whether you have the right to set rules for your blog. The issue here is not even whether Abramoff bought Democrats. The issue here is whether you're an honest man. The issue here is that your recent behavior appears to have patently political motivations.
I'd be very happy to see you prove otherwise, and in that event I'll promptly retract my accusation. The proof is in your hands: the 420 messages you're still withholding.
It's worth taking a close look at the facts, because it would be nice to know that WaPo's admirable steps into the world of blogging are being guided by integrity, and not by politics. This is an important matter that deserves careful examination.
1) A REVIEW OF KEY EVENTS
On Sunday, 1/15, Deborah Howell said "[Abramoff] made substantial campaign contributions to both major parties." That's false. It's also practically a verbatim recitation of an RNC talking point regarding a matter of extreme importance: the biggest DC corruption scandal in at least a generation.
That same day, WaPo readers started posting messages on a WaPo blog. (That blog has this incongruously pastoral name: "Maryland Moment." That blog was apparently picked for lack of a better place to go.) The first such message was posted at 5:26 pm. This message pointed out Howell's error.
One would hope that this message might have been sufficient to prompt Howell to check her facts and promptly post a correction to her fact-challenged article. Oddly enough, she didn't do that. In fact, neither she nor anyone at WaPo made any public response of any kind.
By midnight on 1/15, at least 376 additional messages had been posted, registering similar complaints. In the next 24 hours, at least 212 more messages were posted.
By Tuesday morning, 1/17, at 7 am, at least 667 messages had been posted, in all. At 7:12 am, WaPo editor Hal Straus posted an announcement suggesting that the large message volume was overwhelming WaPo's blog software.
He also pointed out that out of the total of "more than 700 [comments received] so far ... a few comments -- about a dozen -- [had] failed to make a substantive point and were simply personal attacks on Howell and others." That happens to be an incivility rate of 1.7%.
By Thursday, 1/19, at 11:30 am, a total of at least 875 messages had been posted, and almost 4 days had elapsed from the time the first message had been posted. During that 90 hour interval, this is, as far as I can tell, the number of times any WaPo employee made any public statement acknowledging the existence of these messages: 3. I'm referring to the aforementioned announcement by Straus, which was courteous and informative, but obviously made no attempt to address the substance of the complaint. I'm also referring to two blog messages posted by Derek Willis, a Post reporter.
On 1/19 at 11:30 am, Howell posted a statement. She said, tepidly, that there may have been "a better way to have said it." No correction was posted to the original article, and, in the opinion of many readers, she simply replaced her original false statement with another statement that was highly questionable, at best.
There was not a word of explanation for why it took 875 messages, over a period of 90 hours, before she could be roused to deliver a response, as tepid as it was. There was not a word of apology for making a false statement. There was no explanation for why she was failing to post a correction on the page containing the error. There was not a word of appreciation to readers who cared sufficiently to speak up. Instead, she simply characterized her readers as "angry."
I should note that when Howell joined WaPo 3 months ago, as ombudsman, this is the second sentence she wrote: "I have two goals in this job: to foster good journalism and to increase understanding between The Post and its readers."
I think the error Howell made on 1/15 is not "good journalism." I also think that ignoring 875 messages, over a period of 90 hours, and then dismissively describing them as "angry," is not a good way "to increase understanding between The Post and its readers."
She also said: "my mission ... is ... to deal with the needs of readers. I want to create a dialogue with readers ... Readers, you can help this dialogue by not only calling or e-mailing with complaints, but also by letting me know when ... there's something you don't understand about a story or a news judgment. The chasm that separates readers and journalists can often be a deep one, and I would like to help bridge that gap ... Mistakes should be acknowledged and quickly corrected ... Journalists should think of readers as their friends. Because if we're not in this business to inform and help readers, we're not doing our jobs."
In other words, when readers did what Howell had invited them to do, she first ignored them and then condescended to them.
Between 11:30 am and 3:55 pm, at least 202 more messages were posted, mostly attached to the new blog article ("Deborah Howell Responds").
At 4:22 pm, you (Brady) posted a statement indicating that you had "shut off comments." You said that "a significant number of folks" had engaged in "personal attacks, the use of profanity and hate speech." You said "there are things that we said we would not allow." You said "a significant number of folks" were refusing to "follow any of those relatively simple rules." A little later you said there was "a significant amount of people who refuse to abide by the rules we set out."
This might be a good time to mention that I cannot easily find a policy statement on your site. (I did eventually find something resembling a policy statement, but it was much harder to find than it should have been. I would link to it here, except that it's not only hard to find; it's hard to link to, which I think is related to the fact that it's embedded in a frame.) In other words, what you said was essentially this: "it's all your fault because you broke the rules, and when I get around to it, I'll let you know where I hid them."
This would also be a good time to point out the number of times you or Howell posted messages on your blog in an attempt to engage or guide your readers: zero. As far as I know, out of a total of 1,080 messages posted (that count of course excludes the hundreds of messages you deleted), there are exactly two messages posted by a WaPo employee (reporter Derek Willis). This is peculiar, especially since Howell invited readers to have a "dialogue" with her.
You also said "it's a disappointment to us that we have not been able to maintain a civil conversation," which is odd considering that you didn't lift a finger to be a part of the "conversation," or, while the "conversation" was underway, to deign to point out an example of something you thought wasn't "civil." Your phrase "maintain a civil conversation" conveys the impression that you made any attempt to be a part of the conversation. You didn't. This is one of many examples of why it's unwise to take your words at face value.
This would be a good time to point out that you said you "shut off comments." This tends to create the impression that you merely removed the ability to post new comments. But you went much further than that. You hid all the comments that had already been posted. This is one of many examples of why it is unwise to take your words at face value. By the way, on many other occasions you made the same sort of statement (e.g., "we decided to close commenting"), which tends to gloss over the important fact that you went much further than that.
On Friday, 1/20, at noon, you posted a transcript of an online chat. You repeated your charge of "personal attacks, profanity and hate speech."
This might be a good time to deal with those three charges. "Hate speech" has a specific meaning. There is no evidence anyone attempted to post hate speech to your blog. If you have proof of this, you should show it. If you don't, you should apologize to your readers and withdraw the libelous charge.
Your charge of "profanity" is hollow. There are still multiple easily-found instances of profanity among the 1,080 messages you reposted. You've had ample opportunity to find these instances and remove them. You haven't done so. Therefore a reasonable person is not inclined to believe that you're truly concerned about profanity.
As far as "personal attacks," that's a slippery term. Many messages used terms like "hack" and "liar." You're obviously free to decide that those terms represent an unacceptable personal attack. Trouble is, you reposted many messages that use those terms. You also deleted many messages which did nothing worse than using those terms. In other words, aside from applying rules that you keep well-hidden, it's also obvious that you're applying those rules in a highly inconsistent manner.
You and Howell both give frequent lip service to "transparency." To the extent that you use hidden rules and apply them inconsistently, you're about as transparent as the Kremlin.
It was apparently during this online chat (or perhaps shortly before the chat) that you were notified of the existence of one or more archives of your blog, in the hands of third-parties (a reader said "comments in question ... were saved by someone [outside WaPo] before they were removed [by WaPo]"). It was also during this chat that you made this promise: "over the next few days, we'll go back through them and restore the ones that did not violate our rules."
I think it's important to notice that you said nothing like this on Thursday, when you hid 100% of the messages that had already been accepted and posted. On Thursday at 4:22 pm, you simply said "we have shut off comments on this blog indefinitely." Of course you did more than simply turning off the ability to post new comments (which is what your words suggest): you hid 100% of what had already been posted. And, significantly, you didn't drop even a slight hint that you had any intention whatsover of ever restoring a single one of those comments.
A reasonable person paying attention to your actions (and not just your misleading statements which purport to describe those actions) reaches the following conclusion: you took down over 1,000 messages not because of profanity, but because of politics. And you intended to use false claims of offensiveness as an excuse to keep all those messages hidden. Then you were informed that the internet has a memory, and you wouldn't be able to get away with doing what you intended to do. It was only after you realized this that you stated your intention to "restore the ones that did not violate our rules."
Then instead of fullfilling your promise, and restoring all non-offensive messages, you proceeded to selectively restore messages based on politics. I think your reasoning went something like this: "it's a shame that I can't get away with keeping them all hidden, but I might as well take this opportunity to purge some that I find particularly inconvenient; no one will notice." But I did notice. And I know that you noticed that I noticed, because you said so ("I saw that 'analysis' "). And I think this is why you made no announcement when you restored the second, much larger batch, and I think this is why you took this step in the dead of night, on a weekend. But I noticed that, too.
But I digress. Back to Friday.
Later Friday, at 6:32 pm, editor Liz Kelly made this announcement: "Some previously posted comments have been returned to post.blog. Specifically, all comments that meet washingtonpost.com's standards for community interaction have been returned to the post 'Deborah Howell Responds.' "
Later Friday, at 11:10 pm, I posted an analysis reviewing the 42 messages that did not "meet washingtonpost.com's standards for community interaction." Number of instances of "hate speech" in these 42 messages: zero. Number of instances of profanity: 3 (and these are minor profanities identical to those that are easily found, at this moment, in messages you chose to repost). Personal attacks are another story. Is calling someone a hack and a liar a personal attack? Maybe so. But those words, in various forms, are used hundreds of times in messages you chose to restore. So you're being highly inconsistent, at best.
Anyway, here's an example of a message that failed to "meet washingtonpost.com's standards for community interaction:"
Ms. Howell, I look forward to the Washington Post printing a pie chart of this spreadsheet:
When one sees $172,933 contributed to Republicans and $0 contributed to Democrats, one can reasonably conclude "the remark in my column Sunday that lobbyist Jack Abramoff gave money to both parties" requires a full retraction.
As an ombudsman, does your job not entail representing the interests of the public by investigating and addressing complaints reported by individual citizens?
Please read the comments above and the comments to follow with the interests of the Washington Post's readship in mind.
Posted by: bartkid | Jan 19, 2006 1:42:10 PM | Permalink
You have failed to explain why this message, and many others like it, fail to "meet washingtonpost.com's standards for community interaction." And this would a good time to point out, again, that you keep your "standards for community interaction" well-hidden. You should explain why this message was deleted, or you should repost it, and apologize for deleting it. Likewise for many other messages.
This would be a good time to point out that you're failing to put your money where your mouth is. As I've mentioned, you admitted that you saw my analysis. You also said this: "As for the ones that you say were removed that did not violate any of our rules, I'd need to read them. If indeed there are some that fit that category we can put them back up."
You'd "need to read them?" You've had multiple opportunities to "read them." You certainly had an opportunity to "read them" when you read my analysis. So what's stopping you? You said "we can put them back up." We know you "can" put them back up. The question is, why haven't you?
You should explain why this message was deleted, or you should repost it, and apologize for deleting it. Likewise for many other messages. And, in the interest of "transparency," you should announce when you do so, so I don't have to spend my days burdening your server with refresh commands, in order to gather text which I then subject to various forms of computerized analysis.
You also said this: "to pull out a handful of examples and try and turn that into a huge conspiracy is a bit of a stretch."
Uh, you're the one who's pulling out "a handful of examples" and using it as the basis for sweeping accusations. Actually, you're not even bothering with examples. You're making accusations and not backing them up with as much as a single example. You've repeatedly charged your readers with "hate speech." Number of examples you've shown: zero. As far as profanity, many of the messages you deleted are 100% non-profane. As far as "personal attacks," I think you have to decide if words like "liar" and "hack" are off-limits. If they are, you need to delete several hundred more messages which at the moment are hosted on your site. At least be consistent, in other words.
On Saturday, 1/21, WaPo admitted that Howell's original statement was "incorrect."
On Sunday, 1/22, sometime early in the morning, WaPo restored about 948 posts on the Maryland Moment blog. Oddly, this was done in the dead of night, on a weekend, with no announcement. When a much smaller batch of messages was restored Friday afternoon, you saw fit to make an announcement. Why you made no announcement when restoring the much larger batch is a complete mystery. For this reason and others, I don't believe you when you claim you value "transaparency."
Also on Sunday, Howell published an article called "The Firestorm Over My Column." Presumably this would have been some kind of a response to the 1,500 blog posts that had been submitted. Oddly enough, she said nothing about blog posts. She began by saying this: "nothing in my 50-year career prepared me for the thousands of flaming e-mails I got last week over my last column, e-mails so abusive and many so obscene that part of The Post's Web site was shut down." (Emphasis added.)
Was "part of The Post's Web site ... shut down" because of "thousands of flaming e-mails?" I thought it was blog posts, not emails, that led to some kind of "shut down." Anyway, there were supposedly 1,500 blog posts, which is not exactly "thousands." Is it possible that Howell doesn't understand the difference between a blog post and an email? Is it possible that she's really that confused? Or is she deliberately trying to confuse the rest of us, by using the existence of perhaps genuinely offensive emails as an excuse to get rid of 1,500 mostly inoffensive blog posts? This sort of confused and/or confusing statement on Howell's part is another reason why I don't take her seriously when she claims to care about transparency.
Judging from these two recent columns by Howell, I'm inclined to conclude she's not capable of writing a column that is neither incorrect nor confusing.
On Monday, 1/23 I posted an analysis of the 948 posts you had restored early Sunday. Just as in my earlier analysis, I found that you had deleted messages based on politics, not profanity. Here's an example of a message you chose to delete:
well, its now three hours and counting since Willis claimed that "Abramoff did direct donations to Democratic candidates and committees. Our reporters have documents showing this to be the case" and also claimed that he was going to get those documents posted...
but instead of posting these "explosive" documents, the Post deletes Willis's claim....
Posted by: paul lukasiak | Jan 17, 2006 11:29:24 AM | Permalink
I've read that message over and over again, looking for signs of "personal attacks, profanity and hate speech," but I can find none. Maybe you can tell me where to look.
On Wednesday, 1/25, I noticed that you were deleting messages from Maryland Moment. Let's be clear: these were messages which you allowed to be posted originally, but which you then hid when you hid everything, and which you then reposted Sunday morning when you restored Maryland Moment, but which you then decided to delete again, selectively. Follow all that? In other words, you appoved of these messages before you disapproved of them before you re-approved them before you re-disapproved them. Follow all that?
When you mysteriously repost messages, and then just as mysteriously delete some of them, all in the absence of any kind of announcement, this doesn't enhance your reputation for "transparency."
And just as with earlier batches of messages you had deleted, these messages contain no "hate speech." They contain little or no profanity. Is "poophead" worse than anything you ever heard in a locker-room, in your years as a sportswriter? Is it worse than anything that Deborah "Adios Mother Fucker" Howell ever heard? I don't think so. I also notice that Howell didn't mind printing the word "whore" in her column, to make a point, which suggests that WaPo doesn't consider that word to be profane.
Also on Wednesday, you joined a panel discussing this matter. The panel included this very informative exchange: "Hamsher: Are you saying that these messages, which you are saying you pulled, never appeared on the blog? Jim Brady: Not at all. They were there and we removed them." (Emphasis added.)
Your answer gives us enough information to conclude that your explanations do not add up. In particular, your answer shows that you've been misleading us when you made claims such as this: "there's no way for you to see what we blocked." I will explain.
2) THERE ARE DELETIONS, AND THEN THERE ARE DELETIONS
Let's start by defining some simple terms. I use the following terms interchangeably: message, post, comment. These all mean the same thing: entries sent to your blog by readers. They are not to be confused with emails. An email is not a blog post. This is a very simply distinction that everyone understands. (Everyone, that is, with the possible exception of Deborah Howell, as I've pointed out.) Another important term is blog "article." This is something written by you, the host of the blog (or by one of your colleagues). Then readers post messages responding to the article you wrote.
A lot of unnecessary confusion can be swept away if we start by understanding that you deleted messages primarily under two very different scenarios.
A) You deleted messages in realtime, online, while the blog was open, live and active, and new messages were continuously being posted.
B) You also deleted messages during a period when comments for these blog articles were closed, dark, and inactive. In other words, there came a time when you shut down certain blog articles. This means that for those blog articles you hid all messages. Then you spent one or more days deleting various messages. Then you restored the messages that were left, meaning you made these messages visible again.
(We could also discuss a third scenario, C, which followed A and B. In C, the blog was open in the sense that existing messages were visible, but the blog was closed, in the sense that you had turned off the ability to post new messages. Scenario C doesn't warrant much detailed discussion here, because the implications of scenario C are easily understood, once we understand the implications of scenario A and B.)
It needs to be understood that A and B are very, very different. The mechanics of B are simpler and easier to understand. But there's a lot of confusion with regard to the mechanics of A (thanks, to a great extent, to confusing statements you've made). That confusion is unnecessary and it needs to be dealt with, so what I'll do next is explain A in detail.
A) You deleted messages in realtime, online, while the blog was open, live and active. This means a user wrote a message and pressed the submit button. The message traveled to your host computer. The message was then checked automatically by a very rudimentary software filter. I think this filter only rejected messages that contained certain words, such as "fuck" and "cunt." Those words are remarkably absent from the 100,000 words I closely analyzed. Certain other common swear words, such as "shit," "piss," and "cock," were, just as remarkably, not rejected by your software filter (or by your human moderators, for that matter), and are still visible on your blog. This is one of many examples of how WaPo, an institution that ostensibly is very sensitive to profanity, was woefully unprepared to deal with the simple reality that humans use profanity.
The software filter did its work in the blink of an eye. Then, assuming the filter approved the message, the message would then appear, instantly, on your blog. Once the message appeared on your blog, it was then visible not just to you and your WaPo colleagues, but it was also immediately visible to any visitor to your blog.
This is a very important point, so I'll restate it. Any message you saw, and read, and deleted, based on offensive content (or for any other reason), had become visible to the world at the exact same instant that it became visible to you, and remained visible to the world until you took the step of deleting it. (What was important about the recent panel is that thanks to Hamsher's questions, you were finally forced to make this clear. In other words, you filled in an important piece of the puzzle, despite yourself.)
Actually, it was visible to the world longer than that (i.e., visible to the world, under certain conditions, even after you deleted it): any user who saw such a message on your web page would be able to read it, stare at it, admire it, be offended by it, and cogitate over it for as long as they liked (minutes, hours or days), even if you deleted it from your blog seconds after it first appeared. That's because your web page, like most, does not have an automatic refresh feature (and there's no reason why it should). That means that if I navigate to your page, what will fill my window is the status of the page as of a certain instant. That exact information will remain on my screen as long as I like (minutes, hours, or days) until I close the window, or until I use a command that has the effect of refreshing or replacing the contents of the window.
In fact, the contents of the window can remain on my computer indefinitely, because it's a simple matter to save the contents of any web page. (It seems you either didn't know this fact, forgot this fact, or ignored this fact, when on 1/19 you hid 100% of the messages, and made no hint, at that time, that you ever intended to restore a single one.)
So let's be very clear about what's happened so far. In what was probably the vast majority of cases (i.e., cases where a reader managed to express how they feel about you et al without resorting to the words "fuck" or "cunt"), a software filter checked and approved the message, in the blink of an eye. This took place after the user pressed the "submit" button, and before the message appeared on your blog, visible to the world. For all practical purposes, this process was instantaneous. The typical user had the experience of pressing the submit button, and then seeing their message appear instantly. This is a big part of the appeal of posting to a blog: instant results.
I think that's all that needs to be said about the software filter. It didn't do much, for better or worse. It kept out messages containing certain words. That's it. It didn't slow things down. No one is complaining about the messages rejected by the software filter. I wouldn't refer to those messages as being "deleted." I would refer to them as being "blocked." (Part of how you have created confusion is by using the word "blocked" when it would have been clearer to use the word "deleted." And I think my analysis shows that you were being deliberately misleading. You said "there's no way for you to see what we blocked." For what was truly "blocked," by the software filter, that's true. But for what you deleted, it's not true, as I am in the process of explaining.)
The interesting part, where there are issues and complaints, is with regard to messages that were approved by the software filter. Any message approved by the software filter immediately became visible not just to you, but also to the world, and at the exact same instant that it became visible to you. In the moments after a new message was approved by the software filter, anyone loading the page (i.e., you or anyone in the world) would then see that message.
In some number of cases, you read the message and decided you found it unacceptable, and you then deleted the message. But even though you deleted the message, the message would still appear on the screen of any user who had loaded the page before you deleted the message (that is, a user who had loaded the page after the software filter approved the message, but before you deleted the message).
That interval is important (let's call it the "period of public display;" we can also refer to it as the public "lifespan" of an offensive message). Let's be extremely clear about how we define that interval. When does the interval begin? It begins as soon as the software filter approves the message. Remember that all the following things happen almost simultaneously: the user presses the submit button, the software filter approves the message, and the message becomes visible to anyone loading that page, which means you, or anyone in the world.
When does the interval end? The interval ends when you delete the message.
How long was this interval, typically? This is an interesting and important question. Let's figure out the answer.
3) LUCILLE BALL IN THE CANDY FACTORY (movie clip)
That 35 second clip (only 500kb) really does tell the story, so I highly recommend you take a look at it.
You've told us over and over again that highly offensive messages were arriving at such a rapid rate that you just coudn't keep up with the deluge. In other words, they were arriving faster than you could delete them. This gives us helpful information with regard to the important question I just raised: what was the average "period of public display," or lifespan, of a typical "offensive" message?
For context, first let's back up a step and consider what you've told us about the frequency of "offensive" messages, as a proportion of all messages. Currently posted are 1,080 messages (visible here and here). On 1/21, WaPo reported that "more than 1,500 postings" had been received. Therefore I assume 1,500 messages, in all, were submitted (and passed the software-filter hurdle), and 420 (28%) of these were deemed "offensive" and were deleted. I'm going to make an educated guess that you deleted about 120 under scenario B, and the rest (300, or 20%) under scenario A.
That 300 is probably conservative. You have said: "on Thursday, she put a clarification up, and we had already been getting hundreds and hundreds of comments about her column, and they were very, very nasty, using words that I didn't even know existed." This suggests that "by Thursday" you had already deleted "hundreds and hundreds" of comments.
Now let's consider one period of time, that was presumably fairly typical. Let's look at the thread "Deborah Howell Responds," and let's look at the 60-minute period that begins at 2:52:49. The first message posted after 2:52:49 was posted at 2:52:57. (This essentially means we're looking at the last hour of the thread, with the exception of the last two messages. I'm in a position to analyze this exact hour because an archive is available which was captured between 3:52:49 and 3:54:50. That former time is the timestamp on the last message included in the archive. That latter time is the timestamp on the next message, which can be found at post.com, but is not found in the archive. Therefore I'm in a position to know, within a range of about two minutes, exactly when the archive-snapshot was taken.)
For this period, you are currently displaying 38 messages. Given the ratios you've suggested, this tells us that during this hour, a total of about 48 messages were submitted, and about 10 of these 48 (i.e., 20%) you deemed offensive, and deleted, ostensibly soon after they were posted. This means that messages were being posted at a rate of one every 76 seconds. This also means that "offensive" messages were being posted at a rate of one every 6 minutes.
Let's assume an "offensive" message was posted at 3 pm, and became immediately visible to you and the world at that instant. How long did it take before you deleted it? In other words, how long was the lifespan of this message? Well, you've told us you couldn't keep up. You've told us they were coming in faster than you could delete them (like Lucy trying to keep up with the candy). Well, we know that they were coming in a rate of roughly one every 6 minutes (at least during a typical high-traffic period, such as the period we're currently analyzing). Since they were coming in faster than you could delete them, and since they were coming in at a rate of one every 6 minutes, this tells us what we want to know: how long it took you, typically, to notice an offensive message and delete it: about 6 minutes. In other words, we know that the period of public display (at least during a typical hour such as this one) was 6 minutes.
One more time, since this is very important. Offensive messages were arriving at a rate of one every 6 minutes. You said you couldn't keep up. This means that before you had a chance to delete the first offensive message, the next one would arrive, and so on.
If you were just barely keeping up, that means the first offensive message would arrive, you would spend about 6 minutes evaluating it (along with the other messages that were arriving during that time), and then finally at the end of the 6-minute period you would delete the "offensive" message. And this was exactly when the next offensive message would arrive. And so on.
If it was taking you less than 6 minutes to notice, evaluate and delete a particular "offensive" message, then there would no reason for you to claim you couldn't keep up. If the typical lifespan of an offensive message was less then 6 minutes, that means that you were keeping up just fine.
Let's consider what this means from the perspective of someone visiting your blog as a reader (e.g., someone like Malkin or Ververs, who both claimed they saw nothing terrible). A reader is likely to first focus on the most recent messages. What would they see if they focused on the 5 most recent messages? Roughly 100% of the time, they would see 4 benign messages, and one "offensive" message (since you claim the offensiveness rate was 20%). Yes, if they looked at older messages, they would presumably see no "offensive" messages. You had already deleted those. But looking at the newest messages, one "offensive" message would be visible, roughly 100% of the time (assuming that you were "just keeping up" with the flow of new "offensive" messages).
Now let's consider some slightly different assumptions. Maybe it's unfair for me to conclude that the typical lifespan of an offensive message was 6 minutes. After all, you needed to evaluate not just "offensive" messages, but all messages (however, I would argue that most of the messages were clearly benign and could be very quickly evaluated as benign; an "offensive" message might require longer consideration). Given that messages were arriving at a rate of about one per minute this means that if you were "just keeping up," you were devoting about one minute to evaluating each message. This would suggest that an offensive message would have a lifespan of one minute.
Let's consider other possibilities, for what could be the possible minimum lifespan of an "offensive" message. I do know that it's some interval greater than zero seconds. I know it has to be enough time for your human moderator to read it and evaluate it. A typical adult reading speed is 250-300 words per minute (proofreading is slower, and reading on a monitor is slower; I am going to use 200 wpm for my calculations; I think it's reasonable, and it leads to round numbers that are much easier to follow). A typical message, in this relevant universe of messages, is about 100 words. In other words, it would take your human moderator about 30 seconds to read a typical message. Therefore, I assume that the typical lifespan of an "offensive" message was at least 30 seconds.
Let's consider what that means. An offensive message appears. It has a lifespan ("period of public display") of 30 seconds (at least). 5 1/2 minutes after you delete "offensive" message number one, "offensive" message number two appears, and it, in turn, has a lifespan of 30 seconds. This means that a snapshot taken at any given moment, focusing attention on roughly the five most recent messages, has about an 8% chance of capturing an "offensive" message, during its 30-second lifespan (because 30 seconds is 8% of 6 minutes; in other words, even if you kill every offensive message when it reaches the age of 30 seconds, there will still be an offensive message visible about 8% of the time).
If a reader shows up about 10 or 20 times (in other words, if 10 or 20 snapshots are taken, as a window is reopened or refreshed; these are called "pageviews"), then there is a probability approaching 100% that this reader will have the experience of seeing at least one "offensive" message (this is making the generous assumption that you were able to delete an offensive message very quickly, i.e., 30 seconds after it was born; I think it's much more likely that "offensive" messages had a lifespan of 5 minutes or more, possibly much more). (Update: See excellent comment here, by Jerome a Paris, explaining that 20 visits would provide an 80% chance, not a 100% chance.)
This analysis focuses mostly on an example of a high-traffic period, but this is reasonable. That's exactly when many eyes were watching, including eyes that would be likely to notice, capture and remember any messages that were highly offensive, even if they had a short lifespan.
In other words, so much for "you couldn't see them." Wrong. We could see them, and there is an extremely high likelihood that many, many people would have seen them, if they existed. Did we see them? No, because, for the most part, they never existed, outside of your imagination.
Let's review the numerous indications that these ostensibly "offensive" messages never were visible to anyone, except to you, in your imagination. Archives here, here, here, here and here all capture snapshots of your blog at various moments. Now, you are absolutely correct in pointing out that such an archive should not be expected to show an offensiveness rate of 20% (even if it was true that 20% of the messages posted were "offensive"). Here's why. Assuming you were keeping up with the arrival of new "offensive" messages, then most of the "offensive" messages recently posted would already have been deleted, and therefore not visible in the snapshot. You're correct to point that out.
However, thanks to your recent admission that messages were deleted only after they appeared publicly, we know that 100% of these "hundreds and hundreds" of "offensive" messages were all displayed to the public, at least briefly. That is, we know that 100% of these "hundreds and hundreds" of "offensive" messages had a public lifespan that was greater than 0 seconds. Further, we're pretty damn sure that 100% of these "hundreds and hundreds" of "offensive" messages had a public lifespan of 30 seconds or more.
Therefore, just as any human observer, at any given moment, would have a very substantial likelihood of seeing at least one "offensive" message that had been posted very recently, the same is true for these archives. Therefore, when we evaluate these archives, for this particular purpose, we want to pay close attention to the newest messages. This is exactly what an ordinary observer would do if they were looking at the blog "live." And how many "offensive" messages do we see? Zero. Given that we have 5 observations, and given an assumed lifespan of at least 30 seconds, and given an overall offensiveness rate of 20%, it's quite significant that the newest messages in the captured archives contain no "offensive" messages.
5) SCENARIO A, IN A NUTSHELL
If there were indeed anywhere near as many "offensive" messages as you claim, a substantial number of them would have been publicly seen, noticed, remembered, captured, circulated and discussed. That didn't happen, because the messages didn't exist.
Note that while hundreds of people posted messages, it's almost certain that thousands of people were reading messages. And almost certainly at least a few of those readers were people (like Malkin) who were not sympathetic to the posters. Let's face it: there had to be more than a trivial number of observers who were gleefully watching for signs of "hate speech." You're claiming that one out of every five messages was highly offensive. Thanks to your candid recent admission, we now know that 100% of these "offensive" messages were visible to the entire world, for at least 30 seconds at a time (and probably for much longer than that). Let's face it: if such a message had appeared, even once, it would have been all over Free Republic and Little Green Footballs faster than you can say everyone-knows-that-libruls-are-a-grubby-mass-of-swamp-dwelling,-bottom-feeding-miscreants. That hasn't happened. Why not? Because these messages, for the most part, appear nowhere outside of your imagination.
6) OH YES, WHAT ABOUT THOSE OTHER KINDS OF DELETIONS?
Most of the above discussion is about deletions you performed while the blog was live, i.e., what I called scenario A. Let's also consider scenario B: deletions you performed after you hid everything.
You have tried to dismiss the importance of the various cached archives. I've explained why these archives give us important information with regard to scenario A (those archives embody five snapshots that would be expected to show at least one or two "offensive" messages among the most recent posts). It's even easier to understand why they give us important information with regard to scenario B.
In scenario B, you weren't in such a big hurry. You had pulled the plug on new entries, and you had hidden all previous entries. You had stopped the clock. You could take your time. So during this period when all messages were hidden, what did you do? You deleted a number of messages. Were you carefully identifying profanity? Absolutely not. I mentioned above that you ignored multiple instances of profanity, even though you had multiple opportunities to screen that profanity.
By the way, finding "shit" isn't rocket science. Your browser has a find command. It works. The fact that you have reposted multiple profanities, after multiple opportunities to screen profanities, tends to create the impression that your complaints about profanity are a crock of shit. My impression is that your concern is politics, not profanity.
So what do the various archives tell us about scenario B? They tell us that while the blog was dark, it appears that your primary focus was not expunging profanity. It appears that your primary focus was expunging certain political speech that you found particularly inconvenient. You knew you could no longer get away with achieving your original objective, which was to dispose of 100% of these 1,500 politically embarrassing posts. But you figured you might as well try to get rid of some of them, and you figured no one would notice, or be able to prove it. You figured wrong.
7) EITHER YOU WERE ABLE TO KEEP UP OR NOT. WHICH WAS IT?
There are many examples of your statements being inconsistent, on a variety of aspects, including this one.
When it's convenient for you to suggest that you were able to successfully keep up with the flow of newly-posted "offensive" messages, you do so. When it's convenient for you to suggest that you were unable to successfully keep up with the flow of newly-posted "offensive messages," you do so.
You took the former stance when asked why a saved archive, in third-party hands, revealed no offensive speech. You said this: "You were reading the ones that were posted live. There were a few hundred others that were removed the site altogether, and those would not be on the page you're looking at ... that screen shot is only what was live, not what we blocked. There's no way for you to see what we blocked." Similarly, here you said "The cached posts you see don't include any of the posts we removed. Simple as that. When we saw them, we took them down, which means they weren't live and thus not on that cached page."
In other words, you essentially said this: "we were consistently keeping up with the flow of new messages; we were thoroughly scanning them for offensiveness and quickly deleting the offenders, so therefore we can be confident that any saved archive will not contain offensive messages." If you had not always been keeping up with the flow of new "offensive" messages, then you would be in no position to confidently assert "those would not be on the page you're looking at."
(By the way, even then you would have been exaggerating. Since every "offensive" message was publicly displayed, at least briefly, you would never be in a position to claim that a saved archive was completely offensiveness-free. You would only be in a position to claim that a saved archive would contain, at most, one or two "offensive" messages.)
But you took the opposite position when you were asked to explain why instead of only just shutting off new comments, you also removed all existing comments. You said this: "shutting them all off together was just that it was the quickest way to remove the problematic ones that were starting to overwhelm our ability to get rid of them."
(The idea that you were overwhelmed also appears here: "The deluge ... overwhelmed the Web site's screening efforts.")
In other words, you're claiming you were concerned that older messages stored on your site included certain "offensive" ones, which you may have missed because you were "overwhelm[ed]." You contradicted yourself, in other words.
If you were concerned that older messages on your site might include offensive messages (which is your only excuse for hiding everything), then you had every reason to be concerned, for precisely identical reasons, that an offline archive would include exactly the same messages. Conversely, if you were completely unsurprised that an offline archive would be found free of offensiveness, then you had every reason to be confident that your in-house archives were also free of offensiveness. You can't have it both ways.
8) THIS IS WHAT EVERYONE KNOWS
Let's face it. You pulled the plug because you (et al) couldn't deal with the embarrassment, that so many readers would so thoroughly expose Howell's execrable hackery.
You just wanted to make the whole thing go away, which meant hiding everything and claiming an "offensiveness" alibi, and you thought you could easily get away with it. You didn't anticipate that a number of very helpful archives would appear. You also didn't anticipate that anyone would have the motivation and technical ability to use those archives to do a thorough analysis of your behavior.
By the way, if it was really just a question of not being able to keep up with the flow of new, "offensive" messages, that problem had a simple solution: disable the ability to post new messages. If the problem was as you describe (that you reached a point where you couldn't keep up with the flow of new messages), then that solution would have been perfectly adequate. You had no need to take down 100% of existing messages, that you had already screened and accepted, and which were roughly 99% non-profane. The truth is obvious: you took down those messages because they were a political embarrassment to you.
9) A FEW UNSURPRISING EVENTS, AND ONE VERY SURPRISING EVENT
Profanity: no big surprise
Anyone who has spent more than a nanosecond or two reading blogs is well-aware of the fact that humans swear. Surely an ex-sportswriter is well-aware of that fact. Surely Deborah "Adios Mother Fucker" Howell is well-aware of that fact. So surely no one is surprised to notice that certain quite ordinary swear words (e.g., "shit") were posted to post.com.
And the fact is that some of those swear words (e.g., "shit") are still present on post.com. This tends to create the impression that expunging profanity isn't really the issue.
Deleting a few messages: no big surprise
Anyone who has spent more than a nanosecond or two reading blogs is well-aware of the fact that a host might occasionally delete messages. Respected hosts do this rarely, and only for very good reason, and only after careful warnings. In your case, you did it frequently, not rarely, and you did it in the absence of good reasons, and you failed to issue any warnings whatsoever. This is all documented. Therefore, you should not expect to be seen as a respected host.
However, the fact of deleting selected messages is not unprecedented, and therefore not, in itself, a great surprise.
Turning off the new-message feature: no big surprise
Anyone who has spent more than a nanosecond or two reading blogs is well-aware of the fact that a host might occasionally turn off the new-message feature, even after allowing comments to appear on a particular blog article. This simply means preventing new messages from being posted. This is not a common occurrence on most blogs, but it does happen from time-to-time, usually for quite benign reasons. After a thread gets very long, a host might want to direct traffic to a new thread (there are still a few dial-up users, and they can experience a delay while loading a very long thread). Or a host might take this step after it's clear the conversation is simply going in circles.
In any case, such a step is not unheard of. In other words, turning off the new-message feature is not unprecedented, and therefore not, in itself, a great surprise.
Removing, en masse, over a thousand messages that had already been accepted and posted: VERY BIG SURPRISE
I've never seen this done. I've certainly never seen this done on any major site with a large audience. Blog readers/commenters are not likely to want to waste their time on a site where complete threads tend to dissapear at the drop of a hat. Part of what makes blogs popular and powerful is that they are persistent. Statements are made and do not go away, for better and for worse. Statements can be cited elsewhere, for better and for worse. You've just made a very loud announcement that you have no respect for this aspect of blog culture.
Your extreme, unprecedented act also prompts a very obvious question: why did you do it? You've been asked this question, and the answer you provided is completely absent of credibility.
10) CIVILITY BEGINS AT HOME
For someone who's done a lot of whining about how your readers are allegedly uncivil, you've been pretty fucking uncivil, yourself.
Inviting readers to post comments, and then ignoring them for way too long, is uncivil.
Blaming readers for not following rules that you've kept well-hidden, is also uncivil, especially since you never showed up to remind anyone about those secret rules.
Sitting on your hands while Hewitt describes your readers as residents of a fever swamp, is also uncivil.
For someone who is sensitive about name-calling, you've done more than your share. In other words, you don't mind making personal attacks against your critics, in the same breath that you decry personal attacks. I notice you compared Hamsher to "Columbo." The irony is rich. Columbo was famous for being highly motivated to uncover the truth. How odd that you, a journalist, would treat such a motivation as a reason for ridicule. And the pathetic truth is that if you were nearly as "transparent" as you claim, Jane's questions, and the kind of effort you see on this page, would never have been necessary.
11) IT'S FINALLY TIME TO CLEAR THE AIR, DON'T YOU THINK?
It certainly is, and you can do that, very easily.
WaPo is a terrific paper, or at least used to be. I used to trust what I read there. Recent messages posted to your blog indicate that I'm not alone. That you have taken certain steps to help WaPo become interactive is a terrific thing, and you deserve a lot of credit for that. Anyone who thinks that democracy thrives on healthy discourse feels the same way, and would like to see this matter resolved.
You can do that, very easily. You ostensibly have 420 messages that are substantially more offensive than the 1,080 that currently appear on post.com. It's time to find a way to make those 420 messages publicly available. No, I'm not suggesting they should be reposted in the original threads. I'm not even necessarily suggesting that they should be posted on your site. I am suggesting that you find a way to make them available to any interested observer.
One mechanism, for example, would be to create a file that is automatically sent as a reply to emails that are sent to a certain WaPo address. I'm sure clever minds can think of other mechanisms.
I'm sure that the information you're still hiding (those 420 messages) is not a matter of national security. I'm also sure that there is no need to consider copyright implications, since they are no different, in that regard, from the 1,080 messages you are currently hosting publicly. And I'm sure that there is no taint on the august reputation of your fine institution, simply by finding a way to let interested observers inspect those messages. I'm not asking that you print those messages. I'm not asking that those messages be hosted on post.com. I'm simply asking that you find a way to release them.
I should point out that it seems you've already done something like this. You've set a precedent, in other words. Vaughn Ververs of CBS News recently wrote "a Washington Post reporter provided Public Eye with screen shots of the comments that show that occurred in at least a dozen instances."
This language is a bit unclear, but it seems that Ververs was shown messages that you had deleted. I have a strong feeling that Ververs was shown exactly the messages that Straus mentioned, since the count (12) is strikingly identical.
In any case, if you can show those 12 to Ververs, you should be willing to show all 420 to any interested observer.
If these 420 deleted messages reveal that many critics of the Post are so potty-mouthed that they know words that even an ex-sportswriter has never heard before, then so be it. Let's find out. If these messages reveal that many critics of the Post engaged in "hate speech," then so be it. Let's find out.
But if these messages reveal that you're simply doing a sloppy job of covering up a sloppy attempt at erasing everything, then so be it. Let's find out.
PS: if you have trouble locating them, here's a tip about where to look: "We know where they are. They're in the area around Tikrit and Baghdad and east, west, south and north somewhat."