Dev::Blog

Russell notes:

The person who comes up with the app that compels a person to use their phone without considering the fact that it’s a phone is going to have a killer app on their hand. One could argue the opposite, that mobile phone apps *should* only be used in the mobile context, but I think that’s too narrow minded.

I had a half hour to kill last night with wireless access, but I didn’t have my Palm with me (I use Bluetooth/GPRS a lot.) I had a hard time trying to find things to do with just my v551, and it’s supposed to be “cool” and, um, stuff. Or so the sales guy told me. Since leaving the store (and convincing myself to spend more time thinking about mobile computing - it still seems like a limiting shift of mindset!), I’ve discovered that the mobile providers either a) don’t have a clue what to do, or b) don’t want you to spend any time on your phone doing anything that won’t make them money.

In related news, I found a bunch of information on J2ME programming on OS X yesterday. Sadly, my Fink installation seems to be messed up, possibly in connection with the XCode 1.5 install (yes, I ran the December patch, no luck), so I’m having trouble with the preverify command. Any pointers would be appreciated (I’m leaning towards “buy a new Mac because it’s easier than reinstalling everything on your current one,” but I’m also leaning towards “buy a new home,” and there’s a small conflict between the two goals.)

Update: a reboot and a “fink install libpool” did the trick, and I can now write apps for my phone. At least, that’s what the emulator’s telling me. Now what to write?

I’m still obsessing over the idea of alternate content delivery systems for mobile phones. How’s this:

Natural language IVR with txt input instead of voice recognition.

Voice recognition sucks. It’s amazing that it works at all, and the failure rate is just too high for anything other than an incredibly limited vocabulary. How many times have you seen some dude trying to look cool with his voice activated phone, only to have it backfire on him, repeating the same phrase again and again until he finally pulls out the phone and starts hitting buttons? In the meantime, I suspect more teenagers can touch type on a phone keypad than they can on a QWERTY keyboard. Imagine an interface where you can PAY BILLS (72999*224445555557777, or just 729*24557 if some kind of predictive system was put in). It’s a lot more clicks than press one for banking, press 1 for bills, but the whole idea behind natural language would be to skip having to go through the menu tree.

I don’t know if there’s a business model here, since most people who txt are probably more into visual displays and the web and all that, but something like this could work for something like the podcasts over voice idea I floated awhile ago.

From the “what do I know about anything” department: Peercasting Ready for Primetime? Lots of stuff about wise uses of client bandwidth that I suppose I’ll have to think about.

Earlier, I had some thoughts on podcasting and its potential impact on car culture. Of course, my brain won’t stop working, so here are some thoughts on podcasts and bandwidth from a mobile perspective.

The bandwidth issue is starting to raise its head. Even though bandwidth’s gotten cheaper since “Could Success Soon Kill Your Webcomic” was written in 2001, 27 megabyte podcasts are a bit bigger than 200K images, and it’s just a matter of time before people start feeling the burn from their ISPs - Dreamhost now gives me 192 gigabytes of bandwidth a month, which would let me send out, to make the math easier, about 20,000 10 meg MP3s a month without incurring extra charges, which would support 1000 listeners a day if it’s a Monday to Friday gig. Or something. This isn’t really key to my topic, I just want to get the rough idea across, and hey, here’s a possible tangent: is 1000 people enough to attract advertisers?

People are looking into using Torrents and such to deal with this, but then the bandwidth problem expands to hit consumers, which isn’t a big deal, unless the consumer’s on a mobile device using GPRS or whatever (non-WiFi) technology is available. While some parts of the world have unlimited data plans, Canada’s largest wireless data network provider really, really doesn’t (although their newly acquired subsidiary Fido still offers all you can eat GPRS for $50) In fact, it’s generally cheaper to listen to a podcast over the phone in Canada instead of stream the data to your mobile.

Now, I’m not the best user of cell phones (I’m all data, no voice), but think about it: podcasts over the phone network. Sure, it would tie up your phone line, but you’re not supposed talk on a cell phone while you’re driving anyway. You could call some number that would be a gateway to a web site’s (or multiple sites’) podcast library, and you’d just listen. This is actually one of the early business models for the home telephone, and maybe it’s worth a second look - on my wireless plan, it’s cheaper to talk for 20 minutes than it is to download a 10 meg file.

Or how about keeping the content PC-based, but using VOIP technology like Skype to deliver it? The hook here is that Skype already has a billing infrastructure in place for their calls to land lines, so this could enable the infamous micropayment structure that everyone’s waiting for, but that HTTP doesn’t seem to be handling very well (I figure, at least, or we’d see more of it, consumer preference be damned.)

Bandwidth is getting cheaper by the minute, and mobile carriers are responding, but at a slower rate. At the same time, newer and bigger content seems to be appearing faster than ever. 2005 will be an interesting year for delivery and subscription mechanisms, or at least, no less interesting than each of the past 10 years have been.

Apple’s upcoming iChat AV upgrade has me totally ready to start a plan to take over the world (or at least hold it a million billion dollar ransom). Kudos to Apple for going after new target markets!

I have an iSight, and I’ve only used it to establish 1 way video conferences with PC users so I can do voice chats (this was pre-Skype). Invariably, I’d have the camera pointed at the wall within minutes, if not seconds. We’ve had talk of video phones for, what, 50 years now, and it’s pretty clear that people aren’t interested. The only area where I’ve seen any kind of meaningful adoption is the remote business meeting, and the new iChat interface fits that well. Outside of the would-be world conquerer, the app could boost Apple’s rep in small tech companies, which are coincidentally the most likely to bring out the next killer apps.

I’m only now starting to get into the new feature list for Tiger, and while the user-centric stuff doesn’t excite me much (Safari RSS? Big whoop, although Spotlight has me all drippy), I’m super-excited to see what developers do with the back end technologies like Core Data.

Russell Beattie’s discovered PodCommuting, and he’s had an epiphany or two, and he’s onto something with his closer: “On-Demand Audio is the killer app for commuters,” but I suspect that’s just commuters with a car.

I’ve been listening to a few podcasts, but since I walk to work, anything that’s less than 15-20 minutes in length kind of gets lost on me the same way it does when I’m listening at home. I’ve yet to try bookmarking my iPod, but I suspect the problem is more with my attention span. A longer trip to the office would get me a lot further into the podcast swing of things, but it’s not just the trip time that’s a factor.

A lot of the spoken word I’ve downloaded is below radio-quality. That might change, but just as blogging doesn’t seem to have made a vast improvement in the world’s spelling and grammar, there will likely be podcasters broadcasting things I want to hear about at lousy audio levels for quite some time to come. When you’re walking through a busy intersection past a bus, the bus is pretty much all you hear. Ditto for some parts of mass transit. In a car, you (hopefully) don’t wear headphones, so you can adjust the audio within a wider range that’ll let you hear the message during the valleys without blasting your eardrums out on the peaks.

Podcasting is ultimately going to reinforce something Zed said awhile ago:

But your car’s an isolated bubble — wholly autonomous. It provides shelter, entertainment, a comfy chair. And you’re in control. There’s no one to tell you what you can play on the stereo, or how loud. No one to make you pick up your trash. You can smoke cigars, stay up late, sing out of key, pick your nose, or scratch your butt without anyone’s judgement or criticism. It’s an embassy of the self within the foreign land of society. Driving’s the only time a lot of people have to be alone, or to be in charge, or both.

Going further, the car is pretty much a private office for a lot of people, but there are some semi-obvious limitations to what kind of “office equipment” you can install within reach of the driver’s seat (does anyone understand how DVD players facing the driver are legal?) A mobile podcast downloader might fit in well here, possibly some kind of Bluetooth-equipped radio that talks to your cell phone and downloads your feeds on the fly, or a stand-alone solution that downloads throughout the day. We already have radios with Bluetooth that are designed to talk to your phone for handsfree operation, so this would largely be a software solution. Satellite radio is already taking off, but if the barrier to entry can be overcome through EDGE, we could see an interesting shakeup in the broadcasting world.

While in Montreal recently, I spotted a cool promotional idea in the washroom of the Eaton Centre (or maybe the mall next to the Eaton Centre; I can never tell where I am when buildings are connected): the soap dispensers were filled with product from a store in the mall. The store seems to be Fruits & Passion, and this was the cleverest thing I saw on my whole trip, but the sign next to the soap dispenser didn’t really give much help to tourists who wouldn’t be familiar with the brand. I mean “fruits and passion” could just as easily be the name of the soap instead of the store, and a map to the store would have been helpful. Still pretty cool though (and the soap smelled all pretty.)

I would have included a photo, but pictures with camera phones in washrooms are kind of discouraged, and it would have been doubly awkward what with not being too proficient in the local language and all…

I finally got around to installing Skype recently. Skype’s a voice over IP (VOIP) system for various computer systems, kind of like MSN Messenger on the PC, except they’re putting voice first and text second, and it’s positioned as a replacement for the plain old telephone service (POTS), complete with bridges out to allow calls to regular phones for about 2 cents a minute. This isn’t a new idea, and there are other services out there, but since computer to computer calls on Skype are free, it seemed like an easy way to check it out.

I’ve only tried one connection so far (Mac to Mac), and performance seemed on par with iChat (you can “Skype me” if you want, my id is ThrustLabs). Later this week I’ll buy some credits and try the land line options out.

Obviously, the key to getting this to work is to get as many people signed up as possible. Unlike VOIP solutions that cost $20/month, you don’t get a phone number with this system, so the only way that people can reach you is from a PC. With broadband adoption and PC proximity being what it is these days, it’s about the right time for something like this to take flight, but there might be a few more casualties before it catches on. Will Skype survive? Beats me. They seem to have some good ideas, and they’re making some effort to build a community and provide developer APIs, but it still has some ways to go.

What would a call centre with a Skype interface look like? Unlike traditional systems, you’d have a much higher technology base to work from (no more “please stay on the line if you have a rotary phone!”) I had a rant prepared about the death of the IVR (Interactive Voice Response, or the “press one for this” system that fronts most major businesses), since any computer that can handle Skype can surf the web (dedicated VOIP phones notwithstanding), but then I thought a few steps forward, and the same problems applied - no business wants their customers to call them. That’s why it’s so hard to find phone numbers, or even email addresses on many major web sites. Phone calls cost money, and they’re measured by the second.

What if a Skype-based call centre called you back? Rather than dealing with queues, if you called a business and everyone was busy, what if a special browser popped up, like an IVR for the computer, and you could hit a button to request an agent. You’d then keep looking through the interface, and if an agent was available during your session, you’d start to talk, but if you closed the browser (or your “please call me” button was disabled through javascript in response to some actions in the browser), the callback would be cancelled. It’s not perfect, and there’d be issues with how many rings to wait before giving up and so on, but it’s just one way that a next generation VOIP solution could do more than just replace the existing phone system.

I’ve been working with computer telephony and IVRs for more than 8 years, but VOIP is still new to me. I’m thinking it’s still new to the companies that are rolling it out, and it’s still new their potential consumers. Free/cheap long distance is a good hook, and it’ll be interesting to see what a company can do with a sufficiently large user base can do, or at least how long it’ll take for VOIP spam to kick in.