David Duey

Auto-Homing Data

Last week I’d had a little bit of an epiphany while reading one of Brad Feld’s posts.  I made the following comment to his post:

When I read your paragraph about legacy data, I started thinking about your Glue theme and the bad Terminator (T-1000) in the movie Terminator 2. When the bad Terminator was torn apart, its various pieces would return to collective (for lack of a better description). Wouldn't it be cool if your data could automagically find its way home?

Since then, the possibility of auto-homing data has been rolling around in my head.  The basic idea is to have your data return home to a central repository regardless of the original entry point for the data. 

For example, I typed the comment above into Brad Feld’s blog, but I would like to have a copy of the comment along with the comment’s context (date, blog etc.) for my own personal data store.  Currently, Brad Feld has my data; I can copy it and paste it into a document or a database, but I have to do that myself, there’s nothing that will do it for me.

With auto-homing data, I want my data to automatically find my data store and incorporate itself into the collective.  This may seem a little strange at first, but in many ways it’s similar to email.  With email, content is created and sent to an address; the content did not have to be created at a specific place or within a specific application, it just has to be sent to the correct address.

I think we have all of the technologies we need to make auto-homing data happen.  Something as simple as an html tag or a microformat could be used to demarcate data for auto-homing.  We have the transport protocols, POP, XMPP and other standards to help move data around.  Almost all data storage systems (i.e. databases) can interact with XML and JSON formatted data.  And, there’s a ping service, Gnip Central, that could be used to notify a system that new and updated auto-homing data is available.

Of course, there are a number of obstacles to overcome such as walled gardens and data spam (yep!), but I think those obstacles could be overcome.  I think it’ll happen; I think I’ll see the day when my data will automagically appear in the collective and I want to be a part of making that happen.

Smells Like Glue

Trying to do a deep dive into Gnip and Mashery. 

Complementary Products

Free RSS Readers
Not long ago, NewsGator announced that all of its client RSS readers would be free.  Someone asked me why they’d do that.  My response was that NewsGator’s management probably decided that free RSS readers would help drive demand for their more profitable enterprise services.  But, Joel Spolsky, in his Strategy Letter V, explains the complementary product thing much better than I can:

Demand for a product increases when the prices of its complements decrease.

In fact, Greg Reinacker explicitly states NewsGator’s goal in one of his blog posts, ”What we’re working to do is to saturate the market with our clients.”

The Chicken or the Egg?
I think it’s kind of interesting that years ago a small software company actually used the proliferation of its complementary product as the vision for the company, “a computer on every desk and in every home.”  Regardless of what you think of Microsoft, in the early days of the company, I think they did a great job of communicating their “change the world” vision by emphasizing their complementary product.

I suspect that most new technologies face the same chicken and egg problem that micro software faced 30 years ago.  I’ve been working with semantic markup (microformats and RDFa) lately and I see the same sort of problem with the semantic web stuff.  In order for semantic markup to be useful, there has to be content made available as semantic markup.  In addition, users (or computers) need something that can “read” and possibly interpret semantic markup.  I’m seeing signs of life in both areas; Technorati and Yahoo seem to be leaders with semantic content and there are several Firefox add-ons for micoformats.  Also, according to rumors, IE8 will be able to recognize microformats.

Not Implicit Enough
Currently, most semantic markup is “read” by a browser plugin, and once the markup is identified, it’s up to the user to do something about it (i.e. add an hCard to your list of contacts).  But I don’t think content and reader are enough for the semantic stuff to work well.  And, I think that’s one of the problems with the current state of the semantic web, it’s not implicit enough; users have to take some action.  Semantic markup is certainly better than nothing, but there’s still a wide, gaping void between providing content that a computer can recognize and process (i.e. semantic markup) and improving the user experience by implicitly using and consuming semantic markup. 

I think the killer complementary product or service for semantic markup will be a tool (add-in, application, web service, whatever) that not only recognizes semantic markup, but also interprets and processes the information on the users’ behalf.  I don’t think that capability requires artificial intelligence; I think it can be something as simple as tracking attention and user preferences much like RSS readers do today.  Twine is probably a good start.

Business Idea: ClipShots

I think the little SnapShots widget is kind of cool.  When you hover over a link, it pops up a little Ajax window with a scaled-down image of the linked page.  It’s cool, but I don’t think it provides much value even though it has “related searches.”

Here’s my problem:  I often link to definitions of words or phrases on other sites like Wikipedia.  I would love to be able to select just the stuff on the linked page that I would like to be displayed to the user and have that content appear in an Ajax window on the current page.  And, I'd want it to be readable.  That way, users wouldn’t have to jump to the link and then return; users could just hover over the link and read the selected content right on top of the current page.  Of course, if desired, the user could jump directly to the linked page from the Ajax window.

There are a few technical difficulties to overcome, but I think it’d be a hit.

Update:  I discovered ClipMarks which has some of the features I'd like to have, but no popup window.

Ratings and Reviews

Heather and I have been working away on the alpha release of Georneys.  The sole focus of the alpha version is on the “lost child” problem; to provide a means to have lost children returned to their guardian(s) quickly and safely.

However, as part of the Georneys “product mix” (service mix?) we’d also like to help families answer the common weekend question, “Where should we go today?”  We’ve found that parents often have a difficult time finding new and fun things to do with their young children.  Families typically do the same things over-and-over again; go to the same park, mall, and restaurant.  I think a lot of the repetitive behavior is because familiar things are a known quantity, and with small children, a familiar routine is often welcome.  But, I think a lot of the “sticking with the familiar” behavior is because it’s often difficult to think of new places to go and there’s little time to research and think about it, so the familiar is the default. 

As I’ve given more thought to the “Where to go, what to do?” question, I've seen a number of opportunities to use the internet to help solve the problem.  It’s really an interesting problem and I think most of the current approaches are a bit weak.  What I see very often is that places and events are visually presented on a map (Google Maps, Microsoft Earth and the like).  I think that’s probably a mistake.  If I’m looking for fun places to go, I’m concerned about the distance of the place and the time it takes to get there, but I don’t really care about its location on a map – until I need directions to get there.  I think maps are very useful, but when my question is, “Where is a fun place to take the family today that within an hour and a half driving distance that’s will fit my budget?” I don’t really want a map.  What I want is really good, targeted suggestions of places and events that the whole family can enjoy.  Reviews are good too, but I don’t (and probably can’t) spend my day reading reviews; I want a good summary.

The other issue I have with current services is that user reviews may not always be the best indicator of “what’s good.”  Actual behavior is probably a better indicator; I think it’s somewhat analogous to the attention issue with web content.  Behavior is important, so I think feeding a “where to go/what to do” system with actual data would be very useful.  Fred Wilson talks about some of these issues in his post titled “Taste Neighbors.”  It’s a great post and points to a lot of good opportunities in the user generated ratings and reviews space.

Needless to say, I’m excited about working on some of these problems.

Business Idea: Catch SPIT

This isn't really an idea as much as it is a "heads up" about an opportunity to help solve a problem looming in our future.  David Cowan made the following prediction:

I made the prediction that one day email spam will pale in comparison to SPIT (SPam over Internet Telephony). With free VOIP calls, spammers can now use computers overseas to generate voice messages that they broadcast to every 10 digit telephone number in North America.  (Scroll down to the "SPIT and SPAM" post on David Cowan's blog; I couldn't find a permalink for the post.)

SPIT will provide an opportunity for entrepreneurs to help solve the problem.  It's going to be a difficult problem that will probably require both technology and law (unfortunately) to combat.

HCI on the Cheap: Another Weird Idea

Brad Feld has discussed HCI on his blog a number of times.  Yesterday, I listened to a significant (couldn’t resist) interview of Brad on w3w3.com.  Near the end of the interview he mentioned HCI.  (I think there’s going to be a part 2 of the interview next week where he’ll dive a little deeper into HCI.)

Last night, when I went to bed, I started thinking about HCI; I was wondering if there’s a way to improve human-computer interaction using existing hardware.

Specifically, I was wondering if it’d be possible to take an inexpensive USB video camera and aim the camera at a the computer’s monitor and use the video input to translate hand gestures in front of the monitor.  In other words, use software to interpret hand gestures in front of a standard monitor rather than using a touch screen. 

It may not be worth the effort, but I wonder if it’d work.  I’m certain that it wouldn’t have the fidelity of a touch screen (or multi-touch screen) or a mouse, but it would be an interesting experiment.

Stolen Ideas

I’m still reading “The Myths of Innovation” by Scott Berkun.  It’s a relatively short book, but I’ve had little time to sit and read lately.  The book is full of good stuff; lots of food for thought.

While reading the book I came across a good quote by Howard Aiken:

Don’t worry about people stealing an idea.  If it’s original, you will have to ram it down their throats.

Browser Evolution

It’s obvious that desktop applications are dying a slow death even though we’re seeing things like Acropolis.  It seems as if the masses are happier using browser plug-ins like Flash and Adobe Reader than they are downloading applications via ClickOnce.  And, Google Gears is adding fuel to that fire.  So, it looks as if the browser sandbox is the place to play.  But, I don't think the browser offers a terribly compelling user experience; the internet is compelling, but the browser sucks.

Ajax is great, but I think things like Adobe Apollo and Microsoft Silverlight (both plug-ins, so to speak) are really going to make a big difference in the way people interact with online applications.  Unfortunately, the browser will still be in the way. 

What I’d like to see is a simple browser toolbar that launches pages onto the desktop.  Tabs are fine, but they’re still confined to the browser window.  I want something that allows me to stash pages anywhere on the desktop, iconize them, maximize them and treat them as something separate from the browser.

I’m going to create a simple browser app to do that so I can see if I really do like it.  Of course, it may end up on my long list of bad ideas.

Business Idea: Googlodeon

My four year old son loves to play on the computer.  He plays online kiddie games on the Disney and Lego sites and he loves to browse the various toys sites.  He looks at pictures of toys and watches videos.  In fact, my wife setup a “favorites” folder for him called “wish list” where he can bookmark pages of stuff that he wants for birthday and Christmas.

There’s only one problem, he can’t read or spell.  He can read his name and a few small words, but if he wants to see something specific, such as the Batmobile, he has to say, “Daddy can you come over here and spell ‘Batmobile toys’ for me?”  What he means is “Daddy can you perform a search for Batmobile toys?”  I simply go to his computer and type “batmobile toys” into the Google toolbar and do a search, then select something from the search that looks promising.

So that’s the idea: vertical search for non-spellers and non-readers (thus Googlodeon, Google + Nickelodeon).  How would it work?  I don’t know.  Maybe it would allow users to select images that they like and then refine the search based on images that they’re viewing. 

Most of my son’s browsing is based on digging deeper by clicking on images, but he quickly losses his navigation path.  He uses the back button, but he generally will not go back through ten pages of history.  When he’s exhausted his options (i.e. can’t find anymore stuff that he wants) he calls daddy or mommy to do a new search.  There must be a good way to provide children (and others) a way to search without requiring them to be able to read and write.  It sounds silly, but I bet there’d be a market for it.