Hello!  I’m Dax and I do most of the programming here at 3 Dog Media.

After the Wordpress 2.8.1 upgrade, the built-in gallery feature stopped working.  Here’s how to recreate the error:

1. Go to Posts > Add New.
2. Click the Add media button and upload a file.

After uploading the file, a new tab should appear on the Add Media frame that says “Gallery(1)”.  That gallery will stay with your post after you save it.  After upgrading to 2.8, it does not and your gallery changes are lost until you save the post for the first time.  However, if you repeat the steps below with one small change, the gallery works as expected.

1. Go to Posts > Add New.
2. Click “Save Draft”.
3. Click the Add media button and upload a file.

The gallery will also work like normal when editing a post that has already been saved.  As you can guess, unless you know exactly how to recreate it, this could be a difficult bug to identify properly.  The gallery feature is used infrequently and when it is, it’s an afterthought or the culmination of a post that has been saved multiple times.

Here’s how to fix the bug.

For Wordpress single install:

$wpdb->query(“ALTER TABLE wp_posts CHANGE post_parent post_parent BIGINT”);

Stick that in your /index.php file after the line:

require(‘./wp-blog-header.php’);

Then log into your blog.  That should do it.  Delete that line from your index.php file afterwards.

For Wordpress MU:

Create a new file in your root directory “fix.php”, and put the following code in it.

require('./wp-blog-header.php');
$count = 1;
while($count < 2000){
 $wpdb->query("ALTER TABLE wp_{$count}_posts CHANGE post_parent post_parent BIGINT");
 $count++;
}

Then load yoursite.com/fix.php  Nothing will appear, but it should be fixed.  If you have more than 2000 blogs, use a different number.  You probably don’t.

There are cleaner ways to fix this error, but if you are capable of doing them, then you can figure them out from the code above.  I wanted to make sure even the most casual user could fix their blog.

WHAT WENT WRONG

In the WP system, posts, pages, and images are all stored in the same database.  To tie images to posts, thus enabling the gallery feature, and to assign child pages to their parent pages, WP uses a database entry called “post_parent”.  Think of it as a number that links a page or an image to the page or post it belongs to.  Think of it like that because that’s exactly what it is.  If you upload a bunch of images to a post and create a gallery with them, the database will have an entry for each of those image files with the post_parent of that entry set to the ID of said post.

Now why did this stop working for new, unsaved posts?

In previous versions of WP, the post_parent field was a BIGINT.  A BIGINT is a number that uses 64 bits of storage space.  That means the post_parent field had an upper limit of 2^63.  That’s the biggest number it could be.  If you had more than that many posts on your blog, the WP gallery and parent/child relationships would start breaking down because you would have run out of numbers.  That might sound bad, but even if you blogged every second of your life from the day you were born, you would need to live to 292,471,208,677 years old to break your BIGINT ceiling.

Now, you may be wondering why a 64 bit number has a limit of 2^63.  Why isn’t it 2^64?  When dealing with computer integers, the last bit is usually reserved to save the “sign” of the number — whether a number is supposed to be positive or negative.  That gives normal BIGINT’s a maximum of 2^63 and a minimum of -2^63.  That negative “sign bit” is important when doing calculations, but seems useless for keeping track of things like people and posts, right?  Not in this case.

In WP, any post or page that is not yet saved is assigned a temporary negative ID number until it is saved.  By using temporary negative values, the database isn’t filled with useless post ID’s every time you load the New Post page.  For some reason, in the latest WP upgrade, this BIGINT was changed from a signed BIGINT to an unsigned BIGINT.  The maximum number of posts increased to 2^64, but because of it’s lost “sign bit”, posts could no longer have a negative post_parent.  In changing it, WP broke the ability to add a gallery to a new, unsaved post.

It was probably changed to increase the theoretical maximum number of posts a system could support.  However, I think the WP core would break down because of the autoload options table way before then.  If you’re concerned about the number of posts in your system, a better way to cut down on them is to turn off the WP revisioning system by putting the following line in your wp-config.php

define(‘WP_POST_REVISIONS’, 0);

The BIGINT change probably broke a number of plugins as well.  I know we have several that use the negative temporary pre-save ID of a post.  Users should be able to treat new posts just as they do saved posts.

Hopefully this was helpful to anyone dealing with this bug, or at least was interesting to anyone who cares how their CMS works.  There is a lot of information in here that programmers already know, but since more and more casual users are programming with WP plugins and themes every day, I think everyone could benefit from some computer science fundamentals that explain the architecture of the system instead of just how to make things show up.

We are counting down the days now to attend our second annual SEO Oktoberfest in Munich. I have said this before and will say it again. Oktoberfest was the BEST party and BEST trip we have ever taken. Talk about getting the royal treatment? Marcus and Quadzilla don’t mess around when it comes to hosting an event.

This year it will be even better. More speakers, more attendees and more beer.

Wish you could come? How about bidding on a charity auction for the last ticket? Donate to a good cause and you could have the time of your life.

Click here for the charity auction. All proceeds will benefit AMAZONICA.

This is what you are bidding on:

You´re bidding on the last ticket for the sold out SEOktoberfest 2009 (22nd of September till the 24th of September)

The original ticket price is 5.000 euros, and the 15 tickets were sold out after only 20 minutes!

15 International Online-Marketing Experts, 15 Attendees, and 10 former Playmates – one conference – one big Party!

Here´s what´s happening at SEOktoberfest:

Tuesday, 22nd of September:
10 a.m – 4 p.m. – SEOktoberfest Conference
4 p.m. – 10 p.m. – Oktoberfest at famous Käfer Wiesn-Schnänke
open end – After Party at P1 Club

Wednesday, 23rd of September:
11 a.m. – 8 p.m. – We got a whole luxurious Spa completly rented out for everybody. 6 massage therapists etc. will get you fit again!
8 p.m. – open end – Friends & Family Dinner at Munich´s No.1 Wine-Restaurant Die blaue Donau

Thursday, 24th of September
10 a.m. – 4 p.m. – SEOktoberfest Conference
4. p.m. – 10 p.m. – Oktoberfest at Schützen-Festhalle
open end – After Party at a secret location

What´s included in the price?

Everything!

Every Attendee gets a whole Oktoberfest Outfit – Shoes, Lederhosen, Shirt etc., and everything you eat and / or drink, as well as all the transfers etc. – everything´s included!

Who are the Online Marketing Experts?

Todd Malicoat
Chris Winfield
Bob Rains
Johannes Beus
Brent Csutoras
Sickness
RSnake
Greg Boser
Barbara Boser
Todd Friesen
Marcus Tandler
Quadszilla
Frank Watson

Now that the heat has been turned up on Digg, scrutiny is also spilling over to all the “up-and-coming” services who have decided to jump on the “justifiable framing bandwagon.”

One of those services is BurnURL. The concept of BurnURL is pretty simple: You input a url and it spits out a shortened version that you can then use for distribution. But like Digg, they do not implement a simple 301. Instead, they wrap your page in a frame bar.

Also like Digg, they recently published a “don’t worry, we’re not evil” blog post assuring everyone that they have taken steps to prevent any negative impact on publishers.

We’ve always been a very publishers focused company (both with Readburner and BurnURL) and we intend to keep it that way.  We, from the very creation of BurnURL, wanted to make sure the focus was on publishers and their content versus on us and we’ve put quite a few steps in place to do that.

One of the changes we recently put live was to remove the Sharebar when we detected the user-agents of Google, Yahoo, MSN and Ask (these four account for the largest portions of search traffic).  These crawlers don’t need to see the Sharebar (as they’re not going to interact wtih it), so we don’t need to serve it to them.  Instead we feed them a 301 redirect.  This tells them the URL that was burned is the original content owner and it should be listed in the index on that topic.  Our shortened URL effectively gets ignored.

While I am certainly an SEO who appreciates “bot only” technical solutions for SEO related issues, I still find myself wondering how/why the operators of sites like BurnURL don’t understand the “big-picture” issues.

So let me try and explain them one more time.

1. The SEO Value Isn’t the 301 Itself – Bouncing Googlebot through a 301 served from a url Google has never indexed does not generate the type of “juice transfer” people think. The bump from 301 based adtracking/affiliate programs has been on the decline for quite awhile. I won’t go into the specifics in this post, but the short version is that Google is getting much better at figuring out when a 301 is being used as it was originally intended, (moving content that they most likely have already indexed to a new location) and when it’s being used as a bounce.

2. The SEO Value is in the Preservation of the Destination URL – The URLs generated by traditional 301 shorteners in and of themselves are not going to contribute much to improving your organic visibility. But they do play a significant secondary role by making sure that EVERYONE ends up on the URL that the PUBLISHER intended the content to be viewed from. That is extremely important, because it is from that point that the viral distribution process actually begins.

The Bottom Line:

If you are running any type of service that actively encourages the viral distribution of frame wrapped content you do not own, you are a steaming pile of shit. Period.  There simply is no way for you to spin it as some type of “we’re here to support publishers” bullshit. The only reason you do what you do is to gain exposure for your company off the backs of others hard work. So stop being a pile of shit and go figure out how to create something of actual value.

Over the last couple of months, I’ve been getting the urge to get back into blogging a bit. However, when it’s been awhile since you’ve attempted to construct a thought longer than a 140 character tweet, actually doing it is a bit harder than you regular bloggers might imagine. I keep finding myself sitting around waiting for a topic to come along that instantly compels me to start typing.

Well guess what? Today is the day. And the compelling topic turns out to be Digg’s new amazing DiggBar.  In cased you missed the announcement, here is the Digg’s explanation of what the DiggBar is:

“The DiggBar enables you to Digg, read comments, find related content, and share stuff from any page on the Web. And it’s presented in a short URL format, making it easy to share in emails, on Twitter, and via other services. In addition to finding it on all outbound links from Digg, you can generate the DiggBar using any of the following solutions.”

Here’s what it should say:

“The DiggBar is an incredibly clever framejacking tool disguised as a URL shortening service. The mass adoption of the DiggBar by the thousands of users who constantly distribute un-digg-worthy content through our most feared competitor, will allow us to generate millions of additional revenue dollars by injecting our ads in between our feared competitor and the destination url.”

Shortly after the release of the DiggBar, in an article about URL shortening services, Danny Sullivan wrote the following regarding the DiggBar:

“Like lin.cr, it does a 200 code. That means the page is actually on Digg itself – they’re making a page with the DiggBar and pulling in your content without permission into a frame. That’s not illegal, but it’s a tactic that died off years ago. It also means that if you use the Digg short URLs, none of the link credit passes to your page. It’s all kept with Digg.

There’s no need for you to give Digg all your link credit. If you want to shorten your URLs, use a service that does a 301 redirect.”

In response to Danny’s criticism, John Quinn Posted the following on Digg’s blog:

“Prior to launching the DiggBar, we reached out to Google and SEO experts to ensure we adhered to the leading best practices, as we framed and linked directly to source content via the DiggBar. This process involved gathering feedback from publishers to ensure the execution was as content-provider-friendly as possible. We took several steps to ensure that search engines continue to count the original source, versus registering the DiggBar as new content. We include only links to the source URLs on Digg pages to allow spiders to see the unmodified links to source sites. These links are overwritten to short URLs in JavaScript for users who have this preference.”

He then goes on to add:

“We launched a few additional updates early this week to address some lingering concerns in the SEO and publishing communities around the infamous (and sometimes mysterious) search engine ‘juice’. We always represent the source URL as the preferred version of the URL to search engines and use the meta noindex tag to keep DiggBar pages out of search indexes. For those of you interested in the technical details, we also include link rel=”canonical” information to indicate that the original URL is the real (canonical) version. Additional URL properties, like PageRank and related signals, are transferred as well. This is recommended by Google, Ask.com, Microsoft and Yahoo!.”

Sound’s great. (But not great enough for Digg to allow it on their site?)

But here’s the problem… Based on everything publicly published for us common folk, plus a ton of personal testing, I can tell you that the claims in Digg’s post are a flat out lie.

Lest’s start with the noindex part .  A page excluded from Google’s index either by robots.txt or via a noindex meta tag will develop juice, but it absolutely does not pass it. For that claim to be even remotely true, you would need to at least use “noindex, follow” (which Digg doesn’t) and from all my personal testing, that doesn’t work either.

Now for the canonical part. (aka RelCan)

From Google’s official blog post regarding the introduction of RelCan:

Can this link tag be used to suggest a canonical URL on a completely different domain?

“No. To migrate to a completely different domain, permanent (301) redirects are more appropriate. Google currently will take canonicalization suggestions into account across subdomains (or within a domain), but not across domains. So site owners can suggest www.example.com vs. example.com vs. help.example.com, but not example.com vs. example-widgets.com

Based on that, the big question is whether Digg is lying or a backroom secret handshake took place between Google and Digg which lead to Google giving Digg preferential treatment by honoring a cross-domain RelCan tag. I have no way of knowing because neither company is talking, but I did notice the the RelCan Digg uses contains a source tag at the end.

Could that be the secret code that tells Google to count it, even though they have been told to ignore the page? Maybe.

But lets explore the idea that Digg is just lying.

Think about it for a moment. You invest countless hours promoting your content. You get lucky enough to make the homepage of Digg, or you hit the Retweet motherload on Twitter. A certain percentage of all those people who see your content are going to copy & paste the link they land on into a blog post. (Thereby generating a link for your site).

Before the DiggBar, (and with legit shortening services) all those links would point to your url. Now, a large percentage of them are going to be links pointing to a page on Digg. Now if you are Yahoo, CNN, or the BBC, that isn’t really going to matter much. You don’t have to spend time thinking about building link equity, because you already have it. However, if you are a newer site struggling to build trusted link equity in the current black hole environment we live in, the mass adoption of the DiggBar is a serious issue.

I will be advising all clients to add some frame busting code to their sites so the DiggBar won’t work for the simple reason that regarless which scenario is accurate, they are both equally wrong.

Hopefully, others will do the same.

Update: 4-13-09

The DiggBar discussion on Twitter has been incredible. Here’s the most recent Tweets.

I am not sure when this feature was added, but I just discovered it thanks to a stumbler friend of mine giving me the tip. (Thanks Roby!)

Basically all you need to do is go to your ‘favorites’ then hover over the image and then a ’share’ button appears.

From there, you can choose what friends to share with and voila!  The days of sending to one friend at a time are over.

Thank you StumbleUpon. Nice!

We will be heading off to Pubcon next week and are really looking forward to getting away for a few days.  Things have been absolutely crazy and I am really looking forward to shutting down my computer.

On another note I wanted to send out a big bloggy thank you to Michael Dorausch (aka @chiropractic) for this very thoughtful gift I just received. 

In case you can’t read the note, this is our hangover kit we need to take to Vegas. Along with gum so we have minty fresh breath.

Very cool Mike and thank you. See you there.

Matt Cutts has gotten the ball rolling on this “don’t vote” blog meme that asks you to tag five friends to register to vote.  With election day being just about a month away let’s help keep this moving.  So join the meme and register to vote within the next month.  Not sure how? Don’t worry. Ask Google where you register to vote.

Here’s my tags:

Todd Friesen – Because…oh wait…haha nevermind, I forgot he can’t.

Guy Kawasaki – Because he’s way more viral than that other guy.
Brent Csutoras – Because I finally learned how to spell his name.
Chris Winfield – Because … well just because.
Katja Presnal – Because she is the queen of WOMM.
Neil Patel – Because I’m his biggest fan.

October is Breast Cancer Awareness Month

Today is the start of breast cancer awareness month. During the month of October you are likely to see more information about breast cancer than you may have seen all year. The difference is that you can help spread the word by focusing on the topic for at least one month. It’s easy. All you have to do is stumble, delicious, digg, reddit, kirtsy whatever article you see related to increasing awareness.

Help increase the flow of the topic on the web and help save the tatas.

Oh and one more thing you can do.  Turn Twiiter Pink.  Do you remember Green Friday, the ode to @gregboser?

Just imagine it looking like this…

So step up and go pink!