The BFC Computing Weblog

Twittervision is combines Tweets and geo coding to show a realtime display of what people are Twittering. It’s quite a beautiful thing to watch, especially the 3D view.

I guess I wasn’t aware that Twitter fed all updates to third parties, so that’s something important to be aware of - it’s not just your followers who are seeing your updates.

I’m not sure it’s actually useful, but it certainly is neat, so probably it is useful to somebody.

I wonder if Twitter will be coming up with a trademark licensing program to allow apps like this to live peacefully.

The geeks are taking over society, re-making it in their own image.

“How’s this then?” you may ask.

Consider that reality is what you perceive. What you perceive is based on what you know.

So then, what is it that we know? It’s either what we’ve derived ourselves or what we’ve been told or read. Most of us learn far more from others than we figure out on our own.

These days, if an average person wants to know something, where do they turn? Some people go to the library, but most go to Google, or someplace more specific, like Wikipedia.

Now, to add to Wikipedia, you need to learn MediaWiki markup. Most people don’t want to learn this. Geeks have no problem diving in, so they do it. They build an encyclopedia based on their perceptions and biases. Consumers of Wikipedia believe it to be true. Not that Wikipedia is usually incorrect, but perceptions are formed based on what is included or not included.

How about Google? Google tells you what’s out there, and it’s ranked primarily by how many links are pointing to a particular article. Who makes links? The geeks do. Google is a ranking of what geeks think is important, to a large degree.

And, again, users of Google generally accept its rankings to be ‘good enough’ for their needs. They don’t usually ask, “but what else is true that Google hasn’t told me?”

From the blogosphere to major media, to presidential campaigns, much of what “true” is based on what is found online. And what is found online is what the geeks feel like putting there.

If the industrialists shaped the last century, the geeks are going to shape this one. Sit back, enjoy, and go have a look at what’s popular on YouTube today.

There’s some grumbling about that Google is going to start targeting media ads to people based on their search preferences.

There’s only one thing to say about that: bring it on!

Having just sat through a Clairol hair coloring commercial to view a short news clip, I can’t think of any less efficient use of anybody’s resources. I don’t have enough hair to even think about coloring, CNN’s paying for the bandwidth, Clairol is paying per impression - everybody’s time and money just got wasted, and some would say the planet just got a bit warmer.

At least if it were an ad for Just for Men I’d at least be in the target market, but they’d do much better to try to sell me a Netflix Roku box, because I don’t think I can hold out much longer.

Ouch.

Every flash-enabled web browser without a Flash-blocking feature (ala NoScript) is vulnerable to remote compromise.

Having this much exposure completely controlled by one proprietary 3rd-party closed-source vendor is bad for the ecosystem. There’s a Free Flash clone underway, but it’s not good enough to replace Flash for many sites that require Flash, and many sites now require Flash.

Please, website designers: Stop hurting the web. Make sites that can be used without Flash, and add all the glam you want around it. Because Flash isn’t an open standard this problem will always exist. AJAX and SVG can accomplish all or most of what Flash can do, and any talented designer can figure these out.

Update: Adobe has updated their info, and it appears the very latest version (9,0,124,0) is not exploitable, thus this is not zero-day, and I didn’t need to publish this article. Title was: “0-Day Flash Vulnerability In The Wild”.

Sure enough, Microsoft has come back and offered to buy only part of Yahoo! this time (the part it cares about, obviously). This doesn’t prove my conjecture that it only cares about Zimbra, but it sure doesn’t rule it out!

Blaine Cook, formerly of Twitter, reminds folks that architectures scale, not languages.

Some folks have been complaining recently that RoR doesn’t scale, yet sites like Yellowpages.com know how to do architecture and do just fine with it. This isn’t to say that Ruby and Rails both couldn’t be faster and better optimized for scaling, but ‘going wide’ should be easy with a good architecture.

The claim of request uniqueness in Twitter’s case is an illustration of an architecture challenge. I’m always amazed how well Slashdot does with that same problem.

Jeff Dwyer has published a new book, Web 2.0 Application Development with GWT.

Google Web Toolkit is a development framework which lets you write web applications in Java and have it automagically converted to AJAX/JavaScript/HTML and deployed to all kinds of devices for you. Jeff did a demo of GWT at the September 2006 DLSLUG Meeting.

Jeff is a really smart dude, which he proved to me when he worked as an intern under me at PreviousJob™ where we did some work playing with Catmull-Rom splines for fitting endovascular grafts into abdominal aortic aneurysms. I was able to say, “hey, this looks like a good use for Catmull-Rom splines,” and Jeff could go off and learn all the math and implement it. Good intern! So, Jeff was an easy hire once he was graduated. Any book he’s written is bound to be fantastic.

He’s now the CEO of MyHippocampus.com, which I barely understand. :)

Congrats, Jeff!

One problem with LinkedIn is that every contact in your network is treated the same. This could be improved by adding a trust weighting to every contact.

For example, I have a best friend from high school, and some guy I just met at a conference yesterday. I want to give the old buddy a 10 and the new guy a 1, because that relates to how much I can trust them.

But, why?

LinkedIn has great potential for me to find people I need to find. Say I need a designer, an artist, a graphics pipeline optimizer. I can probably search the LinkedIn graph for these kinds of people now, but all I really know about them is how many hops away they are and maybe there are some recommendations.

But with trust weightings, I’d get better answers. I’d much rather hire somebody on a 4-hop 10/10/10/10 path than a 2-hop 1/1 path, for instance.

And, of course, the trusts would be kept secret by LinkedIn, but they’d use them to help you find people.

Any LinkedIn users out there - would you actually be willing to set trust levels on your contacts?

Spin off Zimbra again.

Microsoft’s talk about gettting all Webby is just the party line.

Zimbra is the most significant competition to Exchange and Microsoft’s stranglehold on ‘The Enterprise’ and the real reason that Microsoft wants to own Yahoo!.

Of course this won’t happen - I think the Yahoo! guys are smart and bought Zimbra so that Microsoft would finally acquire them. And they know what the Exchange monopoly is worth to them, so they can hold out until Microsoft gives.

This group is looking to block the deal with government intervention on these grounds.

Amazon has half a disruptive idea in its TextBuy service.

Here’s how it works: You register your phone # with your Amazon account and then if you find something you want to buy (however you do that) you SMS the UPC Code to Amazon. Assuming you have 1-Click setup, it places an order for you and sends you back a text indicating either success (it’s going to ship) or failure (out of stock, not carried, etc.). You can optionally set a preference to not ship the item right away if you want to be able to cancel the order later on their website. In this case, the order gets executed, say 12 hours after you place the order.

Cool, huh? Sorry, I lied to you. That’s not how it works, it’s how it should work. Here’s how it really works:

• You SMS the UPC code to Amazon
• If Amazon stocks the item, Amazon replies with the first two search results for that code.
• To purchase am item, users reply with 1 or 2
• They get an SMS back prompting them for their email address and zip code.
• The user types in their e-mail address and zip code on their crummy phone keypad using 3-tap input.
• The service then calls the user’s phone and completes the checkout process using an automated voice menu system.

I’d use the first system I described but not this one. Coming soon, perhaps?