« More ways to ping | Main | Adventures in MS land »

July 11, 2003

De-publishing, stalking and/or fascinating technology

Have you ever created something, put it out in front of the world and then later revised it? I do it all the time. It helps me to get past the motivational barrier. If I spend too long perfecting, I'll probably spend forever perfecting. Better to get something out there and then refine it.

Usually this kind of "spiral development" is harmless at worst and a good thing at best. In the weblog world, some folks are frowning. The claim is that changing what you say is confusing and dishonest. You have to get it right the first time.

See Don Park's post about the Winer Watcher, and particularly the comments, for more.

Posted by Andrew Grumet at July 11, 2003 12:59 PM | TrackBack | Cosmos


When you write outrageous, abusive, or libelous things about someone, and then later delete those things and claim that they never happened, it is confusing and dishonest. Dave does this *all the time*. "Pulling a Dave" is an idiomatic expression in many circles; it means abusing someone in public and then deleting it and pretending it never happened.

It used to be that we had no way of calling Dave on this behavior. Now we do. Maybe that will keep him honest and force him to stand by the abuse he writes and publishes so recklessly.

Posted by: Mark at July 11, 2003 03:11 PM

I don't think anyone is saying "changing what you say is confusing and dishonest" or that you must "get it right the first time." I can assure that's not what I meant when I coined the term "de-publishing" ( http://www.tenreasonswhy.com/weblog/archives/2003/07/10/the_ethics_of_depublishing.html ).

I'm absolutely fine with people correcting spelling, grammar, and other non-substantive changes. I'm also fine with changing your mind, opinion, etc., as long as it doesn't involve eradicating evidence of your previously published opinions in the process.

My point -- and I think that other people have been making this same point -- is that your public words have impact and you must strive to be accountable for those public words.

This is particularly true when when posting inflammatory comments. Flaming people then hiding the words, refusing to acknowledge them, and trying to pretend they never existed is unethical. I'm not fine with is flaming someone publicly, then trying to cover it up by erasing the evidence. That's Nixonian!

Why delete? Why not just post a correction or an apology?

This isn't about perfection. It's about accountability. Expecting accountability is the opposite of expecting perfection. In fact, it's the expectation of imperfection! The critical question is "How does you address your imperfections as an author credibly and ethically?"

I don't believe the delete key is the answer to that question.


Posted by: Greg at July 11, 2003 03:15 PM

Man, I wish I could go back and change the grammatical errors in my comment above! :-)

"I'm not fine with is flaming someone publicly..." should be "I'm not fine with flaming someone publicly...."

"How does you address your imperfections..." should be "How do you address your imperfections...."


Posted by: Greg at July 11, 2003 03:28 PM

Frankly, I think its disingenuous to say that Dave's being hounded because he doesn't get his posts perfect the first time. I don't think anyone would suggest that he not fix obvious errors or omissions.

Instead, he should do what most people have said that they do: disclose his major changes and not worry about the truly minor ones. Made a typo or stated something poorly? Fix it up. Said something you shouldn't have? Replace it with a mea culpa and apology to anyone who might have read it OR tack on an amendment stating that you don't agree with what you originally posted and that you wish to modify it.

Deleting significant items undermines the credibility of what you say because the first time people spot it (and they will if they're regular readers) they'll brand you a sanitizer in their mind.

You're going to say embarassing things occasionally. It's one thing to own them and apologize, it's another thing entirely to pretend it never happened.

Posted by: Bill Brown at July 11, 2003 04:58 PM

Throughout all of the hubub, the only potentially substantive issue isn't being discussed.

Supposedly, DW edited one of his posts to remove an accusation that M.Pilgrim had surreptitiously changed the behavior of his RSS validator.

So, in the interests of "accountability", Mark : did you change the behavior of the validator? If so, how (behavior:before, behavior:after) and why?

Posted by: brian at July 11, 2003 08:12 PM

So I'm sitting at my computer one evening with a glass of wine at my elbow, and I post an uncharitable remark about how Bill Brown seems to follow Mark Pilgrim around the web and post flattering comments about Mark on the same blogs where Mark slanders, say, Dave Winer. And later I feel bad about suggesting that there's apparently something unwholesome about Messers Pilgrim and Brown's relationship, and I decide to excise it and write instead about the virtues of Cabernet Sauvignon. Apparently Brown and Pilgrim would prefer that that slander live forever on the Web, rather than being relegated to the bit bucket where as it deserves.

I'm not sure how this injects accountability into the blogosphere.

Posted by: Django at July 11, 2003 11:44 PM

Django, IMHO, a person exhibiting accountability would apologize publicly, either in a new post or (preferably) in the same post, because an author has to recognize that once words are made public, they may have already made impact on readers. Erasing the words doesn't erase the impact.

Also, I think in a situation you describe it would fine, even desirable, to modify or delete the offending text if (a) you made note that you were deleting it because it was objectionable and you had thought better of it, and (b) you wrote or linked to your apology in its place.

The readers who were impacted may never come back to that page to read your apology, but since you can't know who may have read it and direct the revision to them personally, the best a weblog author can do is make sure the post in question acknowledges the error of your ways.

Posted by: Greg at July 12, 2003 06:24 AM

Thanks Brian for asking the question I wanted to ask. Isn't it ironic that Mark goes on a jihad about me editing posts, when he apparently made politically heavy change to his RSS aggregator, without comment or note. It seems perhaps Mark is trying to cover his tracks with a shitstorm. Why not, it's worked in the past.

Fact is, Mark and Sam are much more in control of RSS than I. They don't want you to know that. They never told anyone that their policy was to deprecate core elements. Yet that's what their validator did. I took a source code snapshot before they switched policies just in case they did what they eventually did -- change the validator without telling anyone they did.

This is one of the reasons I'm going to do a new validator, and not only open source it, but I will document its policies. Any changes in policy will be made public immediately.

Posted by: Dave Winer at July 13, 2003 02:32 AM

Bill Brown appears to be a real person.


Posted by: Dave Winer at July 13, 2003 02:34 AM

"Bill Brown appears to be a real person."

Was there some doubt about it?

Posted by: Bill Brown at July 13, 2003 03:54 AM

Well, why not... I've defended this ridiculous charge in about 4 different forums so far this weekend, so may as well copy and paste it here as well.


The validator's behavior was not changed. Only the documentation was updated. As I explained over on Simon's site,


...the behavior was not "strange", it's just that (at the time) we were the only 2 people in the world who seemed to give a shit about this particular edge case so we made a judgement call.

Now there is a mailing list called SSF-DEV that is looking into these issues, and I thank them for that, and the validator will be the first beneficiary of their efforts.


To the ancillary argument that it somehow "hurts" interoperability to drop pubDate for dc:date, I say "bullshit". Aggregators that care about dates already support item-level dc:date, because that's the only way to do it in RSS 1.0. The entire point of adding namespaces in RSS 2.0 was to co-opt the work done in RSS 1.0 modules. If necessary, I will produce the email I sent to Dave laying all of this out for him. It will not cast him in a terribly favorable light, but it's the truth: RSS 2.0 was specifically designed to co-opt the majority of work done for RSS 1.0, to take the namespaces and leave the ugly RDF syntax, and thereby to suck the life out of RSS-DEV, which is probably the only thing in the world that Dave hates more than he hates me.

Anyway, the decision to recommend namespaced elements over core ones was a judgement call, deeply rooted in (a) our understanding of the point of RSS 2.0, (b) the climate last fall when we developed the validator, and (c) our own judgement. Honestly, we didn't spend a lot of time thinking about it; we spent more time thinking about *which* elements were equivalent (also a long-standing discussion, ever since RSS 1.0 came out). Nobody cared, or even noticed, until Dave starting all the "funky RSS" FUD-slinging a few weeks ago.

On the issue of "well everybody knows you shouldn't recommend namespaced elements over core elements", I say "bullshit". Item-level pubDate was *new* in RSS 2.0. Previous versions of the RSS 0.9x branch DID NOT HAVE IT. It was introduced at THE EXACT SAME TIME as namespaces (and therefore the ability to use Dublin Core's dc:date, or even dcterms:modified or :issued or whatever you want for fine-grained semantics), and pubDate was therefore COMPLETELY SUPERFLUOUS.

We tried to explain this to Dave at the time but he was completely ignorant of namespaces (except that people were asking for them -- he didn't even know how to create one! and his software certainly didn't handle them properly (a problem which Jake has since rectified, not Dave)) and he was completely ignorant of Dublin Core. If he had heard of it at all, it was in the context of RSS 1.0 and RDF, i.e. the boogeyman. He had absolutely no clue that adding namespaces meant that this edge case (of the same thing being expressed two different ways) could be a problem.

We asked him to deprecate existing core elements (not take them out entirely, just formally suggest that they not be used anymore), but he has this weird definition of evolvability and backward compatibility that precludes any hint of deprecation of any kind. (And then he complains when people complain that RSS is too hard to implement on the client side. Hello, these are related concepts.)

Now, you may not be aware that the entire "funky RSS" business is directly attributable to me. I made the initial RSS 2.0 template for Movable Type


Hell, I'm the reason it's called "RSS 2.0" at all (instead of RSS 0.94).


I mocked up an early prototype of what *I* thought RSS 2.0 should look like.


As you can see, all of my design decisions were PUBLICLY DOCUMENTED at the time. I didn't see anyone else bothering to make an MT template at the time. Dave even pointed to mine and praised it for (a) supporting RSS 2.0, and (b) supporting guids, which were his baby.

(A side note: the template in that entry actually now redirects to the copy on feeds.archive.org, because when we made the validator, we made posted sample templates there based on my previous work, and then I later accidentally deleted it from my site, and people noticed and complained, so I set up the redirect.)

Sam and I *did* plan to incorporate a "best practices" layer into the RSS validator. We naively thought that the discussion of what best practices were would be short and sweet, and we could include it in the first cut. Well, we coded what *we* thought best practices should be (based on our own experience and judgement and biases), but WE WERE NOT WILLING TO DEPLOY IT WITHOUT COMMUNITY CONSENSUS. Which, as you know, never came: we've been around and around on the issue multiple times, and the issue never got enough traction.

So there's code in the validator that is never used. You can see it if you look at the code. When it spots pubDate, it internally logs an informational message "UseDCDate" or something like that. These are the ghosts of our first-cut efforts at the optional "best practices" layer that never came to pass, because we could never get community consensus on what they should be and we weren't willing to simply deploy our own judgements in such a public manner without some backup. You will *never see* those informational messages if you use the RSS validator through the web interface.

However, there are 1 or 2 edge cases (like a feed containing *both* pubDate and dc:date) where our judgement "leaked through" and became visible. As I said before, *nobody cared about these issues at the time except us*, so *of course* it's our judgement that wins the day. It was our judgement, sitting around waiting for someone to tell us otherwise. As you can see from the SSF-DEV links above, we are FUCKING THRILLED that a group has FINALLY convened to clarify the ambiguities in the spec and discuss RSS best practices. We have been trying to do that FOR ALMOST A YEAR NOW, and ANY SUGGESTIONS THAT SSF-DEV REACHES CONSENSUS ON will be immediately deployed into the validator. The validator will be the first, and primary, beneficiary of their efforts.

Everything you believe about the validator being evil, and my MT templates being evil, and all of my intentions being evil: IT IS ALL A PRODUCT OF DAVE'S RECENT "FUNKY RSS" FUD-SLINGING. Absolutely nowhere in the spec does it say "you can use namespaces except where it conflicts with core elements". Some people (like Fredrik Lundh) would say "well it's obvious that you shouldn't do that". Well, fuck you and your sense of what's obvious, Fredrik; I publicly documented my decisions at the time, and Dave and a lot of other people read them and linked to them at the time, and I didn't hear any complaints until Dave suddenly decided -- all by himself, and only quite recently -- that all my work on RSS 2.0 was shit.

I am soooo done with RSS, and people have the gall to wonder why...

Posted by: Mark at July 13, 2003 10:41 AM

Mark, please post this on your weblog, clearly on the record, with a permalink.

Your initial assertion that no one would care is false, by the way. I would have cared. Based on the feedback I got on RSS 2.0 I am certain others cared too.

Mark, finally, about it all being my fault, that's nonsense. I've never told you what to do, everything you've done has been your choice. I wish you would just stop the flaming now. It's so damaging. We'll know you're really finished with RSS, when you stop flaming about it.

Posted by: Dave Winer at July 13, 2003 12:08 PM

Great. I know my arguments are sound when you have nothing left to say but "oh poor persecuted me, can't you see that they're all flaming me for no reason".

I don't tell you what to post on my site, and you don't tell me what to post on mine. This statement is public, and on the record. Just like all those posts of yours that you delete an hour later. Just like those posts from last fall where you praised our RSS validator to high heaven, and praised my RSS 2.0 MT templates. Just like those posts from a few weeks ago where you suddenly changed your mind and started spreading this ridiculous conspiracy theory about how we were "taking over RSS".

Why am I wasting my breath. You used to be at least vaguely interesting, Dave. Now you're just a garden-variety conspiracy theorist. And like every good conspiracy theorist, you are completely and utterly impervious to logic, fact, or evidence. The fact that I deny the vast global conspiracy to steal RSS from you (as if it were a prize worth fighting for) is just further evidence that you must be right. "He denied it, so it must be true," right?

Go play in your little RSS ghetto, Dave. You can have it; it's all yours. I'll be over in my neck of the woods, working on a truly open format, and a truly open API. You are the damage we are routing around, Dave. I refuse to be an RSS sharecropper anymore.

Posted by: Mark at July 13, 2003 01:31 PM

All right Mark, have it your way. I'm trying not to engage you on the personal stuff. I've heard from lots of people that they hate this flaming stuff. I won't participate with you. Even though you address me with such contempt, I will still address you with respect.

One last question. May I have permission to re-post your explanation somewhere where it can be pointed to with its own URL, independent of this thread, so people an hear your side of it.

Thanks for considering this request.

Posted by: Dave Winer at July 13, 2003 02:06 PM

Spare us the sanctimony, Dave. No one honestly believes you're the victim here.

And no, you do NOT have permission to republish my entire post verbatim.

Of course you are free to publish small excerpts of it, for the purpose of commenting on it, under the fair use doctrine. Just as I am free to publish small excerpts of your work from time to time, for the purpose of commenting on it. Copyright law applies equally to both of us. Please include a link to this original thread so people can read the full context and decide for themselves whether your excerpts are misleading or biased.

Posted by: Mark at July 13, 2003 02:24 PM

Hi guys,

As an alternative, how about if I re-publish Mark's explanation? That may seem redundant, but context is everything.

Posted by: Andrew Grumet at July 13, 2003 04:31 PM

What I mean is, this thread contains a lot of extraneous stuff that distracts from the explanation. And since the substantial stuff is copy/pasted, it should't hurt to copy/paste it yet again in a more prominent, less noisy place.

Mark, you would of course have final say over how I used your material. My plan would be to turn it into a regular blog post, with a link back to the original source (this thread).

Posted by: Andrew Grumet at July 13, 2003 04:42 PM

Andrew, that would be perfect.

Posted by: Dave Winer at July 13, 2003 05:11 PM

What's the need to take it out of context?

Posted by: pb at July 15, 2003 06:31 PM

Re: "Bill Brown appears to be a real person."

I asked my wife -- who has a Masters' Degree in social work -- about this. She said that "Bill Brown' is the most common pseudonym in use. In fact, a recent study PROVES that a majority of crack dealers, blog nazis, and scofflaws, upon apprehension, give "Bill Brown" as their name.

Wait a minute, that was "Bill Bryson."

Never mind.

Posted by: Django at August 5, 2003 10:54 PM

Bill Bryson is a helluva guy. I heard that he once got mugged, but ripped off the mugger's arm, beat him with it, and stole his wallet.

Posted by: Bill Brown at August 10, 2003 04:47 PM
Post a comment

Remember personal info?