(this is page 61 of 121)





Latest

Engadget's "International Launch Lineblog" reminds me of those hopelessly feel-good early 1970s/80s Coca-Cola commercials where the company would play that "I'd like to buy the world a Coke" song and show scenes of people from all over the world running in fields and smiling at babies. It's a pretty awesome cultural event when you think about it— around the world today, loads of people from all over the world are going to be getting in line to get their own piece of the Apple magic.

On the other hand, it kind of makes me wonder whether it is only through such a consumerist activity that we can have shared cultural events these days. I guess buying stuff has taken over the role of religion and even media as the great cultural shared substrate?

Hmm, well between that and the low battery life, I'm not sure the iPhone 3G is for me...

Comments Permalink

I've held off writing this post for a while, mostly because I was afraid it would seem like too much inside baseball for me to talk about how photo sites (which Tabblo attempted to be a superset of) basically still suck almost a decade after the first one launched. But the combination of recent usability testing I attended with my own frustrations using Tabblo and Flickr have overflowed the dam of self-restraint, so here goes.

Three Major ways in which photo sharing sites still suck


Most photo sites are still not geared towards pleasing the most important user, the person giving you their valuable time to see yet another album of your kids doing X, Y, or Z. Why is it that the predominant model for consuming photos online is a contact sheet interface (a grid of X by Y photos) along with a clickable "large" version that is often way too small? Sometimes you might get a half decent slideshow too, though these are very un-webby, and after you've seen 2 of any one theme, you've seen them all.

It is interesting that back in the day when Tabblo could freely call itself a photo site, this is the reaction I'd get from people when I'd talk about this consumption interface challenge: 20% of users would get it right away but 80% of people would give me this "huh?!?" look which was meant to make me feel loony. Here is the example I'd use to escape from Loontown: imagine that you bought a copy of National Geographic but instead of the beautifully art directed layouts of the articles, you'd have all the text of the piece followed by a number of contact sheets of all of the photography followed by a page per photo of all of the individual pictures. How many issues would you subject yourself to before canceling the subscription?

Most photo sites don't have fast enough interfaces for the author, especially around photo organization/selection: Try dealing with a collection of 25,000 images to get all of the pictures of a family member for a birthday. The best of the desktop applications can barely do this (and those are the GPU accelerated ones) to say nothing of how far behind the websites are. And manual tagging is only going to take us so far as it is in effect one of many kludges that enable batch operations (a poor substitute for snappy direct manipulation and search).

On this particular challenge, I would have thought that we would have made more progress by now, but as it turns out, I think most people see the assets stored/created on a photo site as a read-only projection of their photo library which tends to live on their home computer. This allows users to leverage the storage/CPU of the local machine for the heavy lifting, and use the web for sharing and output fulfillment. Apple's iPhoto is perhaps the best example of this: books, cards, and calendars can be assembled 100% on the client out of the entire photo library, and manufactured through a set of partner services. This makes activities like collaborative editing and composition more difficult (though not impossible), and sharing is at best very vanilla on dot mac these days, but it may just be something we have to live with, especially because...

Upload of photos is still horrendously broken, and getting more so as image resolution increases. This is what made me think of this topic in the first place again. Watch a novice user uploading photos and you will get all sorts of understandable errors at the boundary of the desktop and web metaphors. They will drag pictures into the browser window and have them open locally. The will go looking on the website for their C: drive. And perhaps most alarming, often times they will not know how to find their images on the local filesystem in the first place (which points to how expired the desktop metaphor is). Watch an advanced user and you will literally see the hair falling out of their head as they deal with the challenge of sucking a watermellon through a straw (the physical equivalent of uploading 12MP images via a 300kbps asymmetric cable connection). While there are solutions to this problem (a software agent running on your PC uploading everything, wi-fi camera cards, etc.), most of them fall short due to intermittent connections and the challenge of organizing/finding photos once they are living online (see above).

Until we as an industry (camera makers, browser vendors, OS vendors, output vendors, etc.) fix these problems, photo sites are going to remain relegated to the backwater of 4x6 production while the rich hybrid solutions (of which only iPhoto really works) continue to take output share while not really advancing the state of photo sharing. It's a shame too because of how inherently social story telling via photos could be.

Comments Permalink

Let's not mess up the web!

Good piece in a well-named InfoWorld blog, "Fatal Exception," on how all of this emphasis on richer UIs for web applications may be unwittingly forcing the web away from what made it great in the first place. Maybe it is the fact that I've been reading Zittrain's book this weekend about how our taste for glossy devices/experiences is causing us to unkonwingly eff up the generative (open) nature of the Internet (ironically, I am reading the book on my very closed Amazon Kindle), but with all of the Silverlight/Flex/GWT stuff landing inside our browsers these days, I tend to agree. Basta to the ever increasing richness of client applications inside the web browser.

Here is an example: there is a new web application called Flowgram that everyone keeps pointing to as the next twist on screencasting. Through a combination of Flash and URL refreshing, you can be taken on a tour of a set of "live" web pages that are clickable and whose embedded objects can be interacted with. Seems pretty cool right?

No, actually it is kind of annoying. Sometimes you just want to dive into the content at your own speed, in your own way and yet, from what I've seen of this seems all but impossible in Flowgram. Despite the fact that there is a timeline scubber thing you can use to jump around, the refreshes are far too slow— far slower than it would be to just load the underlying sites. For instance, I have been looking forward to seeing/watching/reading a presentation on Facebook's photo-related infrastructure, but because I've only found it on Flowgram, I've abandoned it twice, probably never to return again.

The web is great because it is so lightweight— let's try to make sure we don't throw that particular baby out with the bath water.

Tomorrow: why photo sharing sites still suck (including the one we built).

Comments Permalink

The ubiquitous computer

For the 4th of July I dedicated myself to Adam Greenfield's manifesto, Everyware, which argues that computing is about to disappear into every day objects that we will interact with without being conscious of the fact that they represent the post PC era of personal computing. The book is broken up into a set of "theses" that should really be called observations about how people behave in a world where every object is connected to the network and capable of serving as an input mechanism (sensor), a display mechanism (output), or both.

The best reference in the book is to a 1996 paper out of Xerox PARC on "calm computing" which basically argues that we will be able to deal with the information/data overload better in a world where we can interact with information/data via our regular contact with physical objects. Maybe, but only if these objects don't end up becoming the manifestations of Google AdWords 2.0, giving us an apocalyptic Minority Report-like future where every surface becomes a new opportunity for displaying an ad.

Everyware ends on a positive note by pointing out that we (as in regular every day hackers) are being given the tools to control our destinies through open source software, and more importantly open source hardware like the Arduino kits that are popping up all over the place. As folks like Jonathan Zittrain make the rounds warning us about losing control of our digital futures as we fall prey to the glossy experiences of the iPhone or the Xbox, it's good to remember that we can still assert our independence by voting with our tinkering.

Comments Permalink

When I bought my Macbook Air, I debated for about 20 seconds whether to spend the extra grand on the solid state drive. On the one hand it was the bleeding edge, and I've now come to realize that laptops last long enough and evolve fast enough to be worth spending the extra cash up front. What decided me against it was not saving $1,300 (though this helped), but what I call the VW Bug theory of technology purchasing.

When you are buying technology in an industry which is driven by mass market economics (which center around price/performance), it never pays to buy something off of a new branch on the evolutionary tree— at least not in the first generations. This was the genius of the original 1960s VW bug; because it used tried and true car technologies, its economics were riding the tail end of the experience curve. Ditto for spinning hard drives— because a few manufacturers have made billions of them, they've been able to learn how to optimize all of the critical metrics: density of storage, speed of access, and power.

I was worried that the flash speed of access improvement wasn't going to be significant enough, but as it turns out according to Tom's Hardware this week (one of my favorite hardware review sites), SSD drives suck a lot more power out of your battery than you might have thought. In fact, a lot more than those moving platters.

This is not to say that eventually all notebook hard drives will not turn to flash— I think they absolutely will. It's just that the first few generations of this technology will take a few black eyes. Maybe that's why Apple is now dropping the price of the Macbook Air.

Comments Permalink