Walt-O-Matic

Apparently because I like John Clease’s comedic timing, cursing magicians, and canceled 80s comedies, that means I like Carl Sagan’s COSMOS.

Here’s the sad part: they’re right.

Not very long ago, I noticed my iPhone was no longer pushing data up to MobileMe, and further investigation showed that my laptop was also having problems syncing.

The MobileMe icon had an exclamation mark in it, it told me there was 1 conflict, and if I tried to resolve it, nothing happened. If I attempted a sync, it’d attempt it, but I’d get a system log full of errors with no obvious signs of successful data synchronization.

I was seeing ominous system log failures in Console like this repeated many times:

Conflict Resolver[276] [110660] |ISyncRecordGraphNode|Warning| Warning: Failed to look up record with Id: 09000000-0000-0000-1234-430001005678
Conflict Resolver[276] [110660] |ISyncRecordGraphNode|Warning| Warning: Failed to get entityName for record with Id: 09000000-0000-0000-1234-430001005678 (record = (null))
Conflict Resolver[276] [110660] |Conflict Resolver|Error| failed to look up parent relationshipName for entityName: (null) (exception = *** -[NSCFArray initWithObjects:count:]: attempt to insert nil object at objects[0])

There were other strange messages like this:

Conflict Resolver[276] [110660] |UI Helper Proxy|Error| failed to look up UIHelper for attributeName: calendar on entityName: (null) (exception = *** -[NSCFArray initWithObjects:count:]: attempt to insert nil object at objects[0])

And this (my personal favorite as it has a sense of humor):

Conflict Resolver[276] [110660] |UI Helper Proxy|Warning| No data type returned for property “calendar” on entity “(null)”, displaying on blind faith…

And also this:

Conflict Resolver[276] [110660] |Conflict Resolver|Warning| Conflict Resolver: *** -[NSCFArray initWithObjects:count:]: attempt to insert nil object at objects[0]
Conflict Resolver[276] *** -[NSCFArray objectAtIndex:]: index (-1 (or possibly larger)) beyond bounds (2)

My guess is that MobileMe has some globally unique identifier that represents one of my sync-able objects, and for some reason it’s missing. That in turn throws off some collection count, and when things don’t balance out between what was expected and what was loaded, a software exception happens.

At that point, I was fairly sure I needed to converse with MobileMe support. Apple has free chat-based support services, but it’s buried. Really buried. Really, really, buried.

1. Go to http://www.apple.com/support/mobileme/
2. Expand Syncing With MobileMe
3. Click Troubleshooting MobileMe Sync issues
4. Click Chat Now at the bottom of the page.

It’s a good idea to have your machine and your MobileMe data backed up. TimeMachine, SuperDuper!, or Carbon Copy Cloner will help you do this.

Unregister and Re-Register the Machine

Bring up Apple / System Preferences… / MobileMe. Go to the Sync tab, and at the bottom click on Advanced…

Then select the computer name in the list, and a Stop Syncing Computer… button will appear. Press it. Confirm with Unregister.

Then press the Register Computer button. Press Done.

Check and set the synchronize with MobileMe to Manually. Leave the preferences panel up, you’ll be back in a second.

Blow Aware the Sync History

You’ll need to reset your sync history which requires the iSync tool and a Python script run at the command line:

1. In the Finder, choose Applications from the Go menu, then double-click iSync.
2. Choose Preferences from the iSync menu.
3. Click the Reset Sync History button.
4. When the window opens to confirm the reset, click Reset Sync History.
5. Close the Preferences window.
6. Quit iSync.
7. Open the Utilities folder (located inside the Applications folder) and double click the terminal.
8. Paste this command in the Terminal and press return: “/System/Library/Frameworks/SyncServices.framework/Versions/A/Resources/resetsync.pl full” – without the quote marks. There will be no output.
9. Quit terminal.

Re-sync

Select the items you want sync’d. Change the Synchronize with MobileMe back to Automatically.

You might get prompted with a request to sync immediately. If you are trying to push everything on your machine to MobileMe, stepping on what’s there, then press cancel and follow the next step. Otherwise confirm and skip the next step.

This next part is optional and you do at your own risk, assuming you have a backup and this is what you intended to do:
Press the Advanced… button. Click Reset Sync Data… and select the direction you want to sync, most likely computer to MobileMe.

At this point, the conflict disappears, but you’re no long able to sync either. Move on to resetting the preferences.

Resetting Preferences for MobileMe

In the MobileMe preferences pane, select the Account Tab, and click Sign Out… Confirm with Sign Out.

Now provide a bogus username and password like: blah@me.com / blahblahblah

You’ll get a name or password invalid message, but your log will show some interesting stuff.

/Applications/System Preferences.app/Contents/MacOS/System Preferences[352] Warning: Removed .Me password
[0x0-0x2c02c].com.apple.systempreferences[352] com.apple.CSConfigDotMacCert-blah@me.com-SharedServices: Already loaded
com.apple.launchd.peruser.501[206] (com.apple.CSConfigDotMacCert-blah@me.com-SharedServices) Throttling respawn: Will start in 6 seconds
com.apple.launchd.peruser.501[206] (com.apple.CSConfigDotMacCert-blah@me.com-SharedServices[470]) Exited with exit code: 1
/System/Library/CoreServices/FileSyncAgent.app/Contents/MacOS/FileSyncAgent[462] PIDFilePath() => ‘/Users/yourname/Library/FileSync/01254cc20d18/.pid’

Sign back in. Once again, go to the Sync tab and set Synchronize with MobileMe to Automatically.

You will get a message that says “A computer named [your machine name] is already registered with MobileMe synchronization server.” If so, press the “Use same name” button.

You may see this in the logs, not to worry:

System Preferences[352] First pass at computer registration failed with error: Error Domain=DotMacProxyErrorDomain Code=-100 UserInfo=0×200e73280 “A computer with this name is already registered with MobileMe.”
System Preferences[352] First pass at computer registration failed with error: Error Domain=DotMacProxyErrorDomain Code=-100 UserInfo=0×200e5fa20 “A computer with this name is already registered with MobileMe.”
System Preferences[352] LightweightMallornLoginSession is registered.

Now check off the items you want sync’d again and press the Sync Now button.

If you are trying to move all of your machine’s data to MobileMe, select the correct replace option when prompted.

And just to be sure that it didn’t do a partial sync, press Sync Now a second time, just incase the automatic setting jumped the gun before you finished selecting all the desired items.

Your syncing woes should be resolved.

But what about me.com?

At this point it’s a good idea to head over the your web account by going to me.com and logging in.

Check to make sure your data is there.

At this point in time there is known issue with me.com in which the calendar and the contact data does not appear. It is a known problem. Apple is aware of it. It is specific to your profile (other MobileMe accounts aren’t affected). And you need to contact support (see above) and open a trouble ticket. Apple only knows this as a “contacts and calendar loading issue” it has no formal title.

The error you see will be this message on a grey screen: MobileMe is unable to load your contacts. MobileMe could not load your information from the server. Try reloading the page. If this problem persists, contact MobileMe Support.

You can try and clear out your Safari cookies and cache, but realistically this won’t work as other browsers, like Firefox, show the same thing.

1. First click the log out button and close the MobileMe (me.com) window.
2. Click the safari title (next to the Apple logo) and select “empty cache.”
3. Next click the safari title and select preferences.
4. Click the security tab.
5. Click “show cookies” then hit “remove all”.
6. Now close the preferences.
7. After all that open a new browser window and log back into MobileMe.

Apple can fix the problem by escalating to the next level of support, and this is most likely more desirable than closing your MobileMe account and opening another, which will force your MobileMe account name to change.

As a glossed over quick introduction, MacHeist is a short-run sale of software packages for the Mac that has a twist. You pay $39 for a bundle of software, and some of that software is “locked.” A portion of your purchase price goes to charity, and the more money raised for charity, the more items in the bundle that get “unlocked.” Thus the more people buy, the more you continue to get. It’s a great scheme, only it isn’t working.

MacHeist, at the time of this writing, is conducting their third “heist” and after some amazing fluster of activity, new sales appear to have stagnated at an alarming rate.

Alarming to bundle purchasers, because if not enough sales happen, bundle purchasers won’t get all the amazing high-cost software at the extreme end of the bundle. What’s important about that statement is that it’s never happened before, and the problem isn’t the recession.

In informal polling, there appear to be two kinds of purchasers: early adopters and frugal purchasers.

The early adopter purchases the bundle early, knowing a good value when they see it, spurred on by the fact that there are additional incentives for doing so.

The frugal purchasers have their eye on either the final packages in the bundle, or are looking at the bundle as a whole. They don’t want to purchase the bundle until they know everything in it is unlocked.

And that’s the interesting part. If no one buys it, nothing gets unlocked. If everyone takes a risk, everyone gets handsomely rewarded, guaranteed. Thus each potential purchaser is waiting on the action of everyone else — it’s crowd mentality, only the driven behavior is idleness.

The secret ingredient is momentum. By carefully crafting a set of software incentives, under ideal circumstances the early adopter crowd overlaps with the late takers. This manifests itself as a steady stream of purchases.

It might be argued that The Directorate which runs MacHeist became victims of their own success and actually caused the problem by marketing the sale too well. Based on all the pre-sale puzzles, rumors, and incentives, there was a flurry of purchases in the early hours of the sale and projections seemed rather high.

However, one of the primary packages in the bundle required what looked like a high goal to unlock, the perception was that momentum was slowing. And perception drove reality. “Hmm, that doesn’t look like it’ll get unlocked, I think I’ll wait to see if it does before I buy,” is all it took to slow the influx of unlocking purchasers.

This was ill-timed, as it also happened to coincide with the reward for the first 25,000 buyers being removed from the table as the 25,000th bundle was sold. Days later, a only mere 5,000 more have sold and questions are being raised if the final packages will be unlocked.

The up-front fast burn created enough of a gap that people who were on the fence at different points became more segregated than usual. This didn’t happen in the last two sales.

So here’s my prediction: they have to fix this. Meaning, new incentives will re-emerge, the goals will have to be re-addressed, and it’s in the best interest of MacHeist to unlock the bundles anyhow at the end of it.

Turns out before I could finish this post, a new bonus was added, and that did stir a little traffic. But the real objective here is to convey there’s movement, specifically enough that the goal could be reached. That will inspire sales again, and in turn actually unlock the software. By re-calibrating the goal levels, this would solve the problem. In fact, the easy solution is to put all the last packages into one final, achievable goal.

The truth of the matter, however, is whatever happens will be remembered, if not chronicled in Wikipedia forever. If MacHeist goes down in flames for not unlocking all it’s bundled packages, people will be ever the more skeptical, and that means early adopters turning into late purchasers. That only exacerbates the problem, killing future sales opportunities.

By contrast, if the packages do get unlocked, whether by purchasers or by The Directorate making its own donation from the profits it receives, then MacHeist will be seen as more of a sure thing in the future, sliding more of the late comers and risk adverse customers into the early adopter side. This would actually increase future sales, because more gets unlocked sooner, enticing the skeptical buyers.

As such, “betting” on MacHeist with a purchase at this point still seems like a safe move. And, even if none of my predictions happen to come true, enough is unlocked already that the $39 price tag is still an awesome buy for the collection of software provided.

As a photographer that photographs models, there’s two primary goals that any website that tries to connect models and photographers should aim for:

All else is peripheral.

The idea is that if you’re a model looking for work, you post your portfolio online and photographers approach you with gigs. Conversely, if you’re a photographer, you post your portfolio and jobs start coming out of the wood work. The reality is that few sites can deliver on the promise adequately, not to a fault of the site’s objective, but due to design, business model complexities, or subtleties pertaining to the problem of brokering.

Naturally, for any such site to work you’re going to need a critical mass of both kinds of users just to have a wide enough selection to make this happen. As such, it’s important not to alienate users — something that is very easy to do with bad design or practice. It’s not enough that a site be free.

The closest site that I’ve come across that seems to have the right idea is www.ModelMayhem.com. It’s search capability is right on target. You tell it that you are looking for models in your local area that are some number of miles from your zipcode, that are between the ages of 18-24, female, 5′2″ – 5′7″, olive skin, with shoulder length black hair, green eyes, and poof — out pops a number of candidates.

This is the way it should work. You tell the criteria about what you need, and it finds people with those attributes.

The problem is the interface is klunky, the portfolio space limited, the navigation is horribly disorganized, and pretty much anything other than models is left wanting. Yet it’s still usable.

I wish it had a way to describe the kinds of services photographers offered and made them in a searchable fashion as well. Oh well, at least finding models isn’t problematic.

Such locator services are not a social network, nor are they a dating service. They’re supposed to be resources that connect professional with professional, with the added bonus of having a reasonable idea of what you’re getting. It frustrates me when a site is designed around chit-chat and messaging. Simply put let one professional find another, preview their work, and then get in contact with them by email; don’t obscure things. A site that works gets traffic, it doesn’t need fudging to get visitors.

Conversely, I just deleted my account over at www.aMuseBook.com, a web site that professes to do the same thing: connect models and photographers. I’d argue not only that it doesn’t, but that it physically can’t in my personal opinion — it’s a business model problem gone awry.

While better organized, and certainly much prettier, it’s search capabilities are downright awful. The best geographical resolution is state-level. So, if you’re looking for a model in Texas, that’s all of Texas. Additionally, providing search criteria for attributes just isn’t possible, which means locating a specific model by looks isn’t doable. And if you can’t find candidates, you aren’t going to be hiring.

Here’s another bad design choice that just seems obvious. If you want to find a model, you typically are looking for an age bracket, yes? Well, the site doesn’t let you search by age, instead you have to search by a specific birthday, which is stupid. Oh, and that’s a Day – Month – Year birthday at that. Even searching by year alone isn’t helpful, because simply year subtraction doesn’t give age.

Now while aMuseBook does give you more space to store your photos, it unlocks features using a point system. You gain points by commenting on people’s pictures and telling your friends to join. What becomes transparent very quickly is that the site is not structured to make contacts, but to get you to churn through pages so that Google Ads get thrown in your face generating an alternate revenue stream. I quickly got tired of being told in every email I have to “use” the site and it will ‘work’ for me.

Hogwash. If I can’t locate a model or post a comprehensive portfolio, then neither I, nor the models, are getting any serious value out of the site.

Adding insult to injury, the site gives you the ability to provide URLs to your own site; this sounds good at first, until you realize that many models and photographers keep their photos on Flickr. Why? Because Flickr is great for managing photos. But what does aMuseBook do for those sites? It blocks them out, showing up as www.*****.com, and when I questioned the site admin about it, I got back a response stating they didn’t want their site for depositing competitor URLs and not another portfolio site.

Wait a second. The service is there to help me find people by showing them my portfolio but they don’t want me to show them my portfolio if it’s elsewhere? Plus I can’t post my portfolio unless I leave comments that I wouldn’t have otherwise. That’s stacking the deck and gives unrealistic feedback. And when points are rewarded for clicking on ads, I’m pretty sure that’s against Google’s terms of service for AdSense.

If people are churning pages leaving “Nice smile” comments, how is one to know which comments are real (and therefore useful) versus people just trying to collect points? The information itself becomes devalued. Thus the business structure and the design alienates users in the short term, while the lack of utility alienates them in the long term. It can’t be viable.

And that’s why I deleted my account over there: It wasn’t usable or productive.

No wonder it’s so hard for models and photographers to get connected. I wish there were a simple directory that focused on doing one thing and one thing well, connecting professionals. It’s a hard problem, but the person that cracks that nut can steal a whole lot of business from all these other sites without trying too hard.

Fundamentally, the problem is that a brokering agent has to provide and organize information. Limiting it, not being able to search it, or failing to have a positive user experience drives away the very assets that are needed to make the site work. This appears to be a case where a well simple organized directory could be a winner-take-all.

---

UPDATE: I have found an awesome site for models, photographers, and makeup artists. It’s call Miss Online and it allows unlimited photo uploads with no point limitation schemes. It also includes discussions, groups, blogs, and email. The site is very active and quite attractive to use; advertising is at a minimum, and you aren’t coerced into clicking through tons of pages. Plus, and here’s the real proof: as a photographer I’ve had more exchanges with models with this one site than all the other sites combined. It does get you connected.

I used to be a pretty big fan of wimp.com, a site that collected links to all kinds of interesting videos. Not having much time on my hands, this was the perfect aggregation of interesting content.

Now, when I go visit a link, an ad usually pops up. I have to close it. Then the video starts. And, again, another ad slides up from the bottom, and I have to that add. Then all the while the video is playing, I have a little “AD” box overlaid hoping I’ll press it.

The new face of wimp.com — why I’m done.
While I don’t begrudge wimp.com, or other sites, from having advertisements, I really dislike intrusive ads in the video stream.

As such, wimp.com, you’ve just gone from being one of my favorite sites to one of my least visited sites (which means no ad clicks, no ad views, by the way).

The simple solution would have been placing an ad elsewhere on the page, even under the video.

Want to know where the ads should have gone? To the right of the directory listing of links. That’s the page I’m always coming back to.

Another great part of the internet just died for me.

UPDATE 12-Mar-2009: While showing this horrible predicament to a friend the site acted differently; it played the video with no ugly overlays or pre-ads, but displayed an advertisement when done. That is perfectly acceptable! Fantastic compromise guys!

UPDATE 18-Mar-2009: Ugh, now it’s worse — ads pop up while you’re watching, even after you’ve dismissed them.

In Bruce Schneier’s CRYPTO-GRAM, he includes a reprint of a fantastic article entitled The War on Photography.

Excerpt:

Since 9/11, there has been an increasing war on photography. Photographers have been harassed, questioned, detained, arrested or worse, and declared to be unwelcome. We’ve been repeatedly told to watch out for photographers, especially suspicious ones. Clearly any terrorist is going to first photograph his target, so vigilance is required.

Except that it’s nonsense. The 9/11 terrorists didn’t photograph anything. Nor did the London transport bombers, the Madrid subway bombers, or the liquid bombers arrested in 2006. Timothy McVeigh didn’t photograph the Oklahoma City Federal Building. The Unabomber didn’t photograph anything; neither did shoe-bomber…

As a photographer, I have been stopped by security guards, questioned why I was photographing a building, and probed who I was working for. Bruce explains while not only is this nonsense, but a waste of resources and money.

The article’s short. Take a moment to read it. It brings common sense back to the equation.

I’m a photographer, and if I take a picture of something, it’s because I like it and want to preserve it for others to enjoy too.

This morning I woke up to an email, it basically read this:

You received this email because you have created an account on Beyond.com. This is a one-time mailer. If you have any questions, please contact us.

I’m thinking to myself, “what?!?” Actually, I’m thinking something quite a bit more colorful.

Then there’s another message from Customer Service.

Then there’s another message with my username and password.

…right.

After deciding it isn’t some email spammer trying to get me to some foreign national site, I login. And what do I find? Someone had screen scraped an old copy of my resume and contact information and made an account for me.

At this point, I figure that anyone with any common sense should completely discount beyond.com’s credibility completely. Here’s why.

First, if any arbitrary user is able to make up accounts for someone else, then clearly the database of provided by beyond.com can’t be trusted. I know my information was wrong, so clearly any potential employer looking for candidates would actually be wasting their time — it isn’t an accurate representation out there. But more over, this represents bad business and security practice if someone other than the actual person can create an account.

Second, let’s assume that such a thing isn’t possible. The alternate conclusion is that beyond.com is scraping the web, making accounts, in an attempt to build a database to give the appearance they are more than they really are. Will some suckers sign on and “correct” the information? Perhaps. But I suspect many others will ignore it. Again, this is really not helpful for anyone trying to use beyond.com for candidates.

Bottom line, either side of the coin — something is wrong. Very wrong.

And, of course, removing that profile is painful and obscure. The help files toss around words like ‘deactivate’ rather than ‘delete’. Such things should make users of beyond.com question the marketing metrics of beyond.com as well.

To me, and in my personal opinion, beyond.com isn’t worth the pixels its printed on. In fact, it sucks.

REVIEW: Walt gives Beyond.com two thumbs down.

A while ago, I thought it might be fun to conduct an experiment and rank anything and everything that Amazon showed me. In fact, the rank wasn’t necessarily even important, I just wanted to see what would happen as recommendation after recommendation was ticked off. Would Amazon’s suggestions get better? Would it run out of suggestions? Would it result in an overflow message?

Well, I ranked over 10,000 items over the course of several months, ranging from computer books to perfume. What I found was that in the short term you could get Amazon to run out of things to recommend you. In the longer term, it got a little better recommending things, though the categories get broader, and if you stumble into a new kind of category, it leaps at the chance to have options again to show you. And, finally, nothing spectacular happened numerically when I crossed five digits.

That said, every so often, Amazon makes some amusing recommendations choices. However, this time it was the presentation that was amusing unto itself that I took a snapshot.

Know those split-books you had as a kid, where the page was divided? You’d get half an animal on top, and half an animal on the bottom. Allowing you to make a giraf-o-potamous, an elepha-gator, or a kanga-mander.

Amazon selected two products and presented them split-book fashion. Order, it turned out, was important:

It’s the top of a woman from 2002, and the bottom of another from 2007, put together it looks like one woman standing behind two cut outs on the product recommendation page. I couldn’t help but give each half five stars for creativity.

Just spent a few minutes circling trying to get Verizon news groups working. Here’s how I did it, you’ll need a news reader – OS X users, go try Unison.

First, I went to Verizon’s start page, http://start.verizon.net/, and discovered I could not login. Since I don’t use Verizon’s email or web services, I had no clue what my username and password was.

Calling 888-553-1555 and asking for Verizon FiOS Internet Technical Support (a 24/7 service) got me through to someone who was easily able to tell me my username and reset my password to an obscure string of characters. He then helped me log on to the web page and change my password, to which I did.

At that point, I made the mistake of hanging up and attempting to connect to the Usenet on my own. It didn’t work.

I called back, and again got someone who was quite helpful, and through some minor experimentation figured out what was going on.

• Some Verizon literature states the server is newsgroups.verizon.net, it is not. The correct news server is news.verizon.net, and you can ping it. The IP address I got was 199.45.49.11.
• Be sure that you don’t enter verizon.com – that will not work. You want dot net.
• Even though I had changed my password, it had not propagated – to access the Usenet, I needed my old password, the one the tech person read over the phone.
• Your username does not include @verison.net at the end of it, don’t type that. It was a bad habit my previous ISP started.
• The news server does not support SSL, stay with port 119.

The tech support person also mentioned there was a page for setting up newsgroups with Verizon FiOS online.

From what I can tell, Verizon has very good news group article retention.

Ever wonder how Google Images seems to zero in on images so well?

I just stumbled into the Google Image Labeler, and it’s addictive.

Google shows you a random image, and you enter in as many keywords as you can think of in real time. Meanwhile, a partner you’ve been paired up with does the same thing. When you match one of your terms, you progress to the next image.

You’re given a finite amount of time to do as may as you can, scoring points as you complete match after match. A score board is kept so you can see your ranking, as well as compete for the top titles.

So, while you’re playing this game with a mystery person on the net, you’re actually seeding Google with image tags, the ones where you both match are given validation that two independent people looking at the same image came up with the same tag.

Clever. And fun!