With BitTorrent Sync being available on Android, this means you can simply and securely make a backup of your Android device to your home computer using Titanium Backup and BitTorrent Sync from anywhere. The beauty of this method is that there's no middleman you need to trust with your data as it's encrypted from end-to-end.

To follow this guide, you'll need to install Titanium Backup and BitTorrent Sync on your Android device and BitTorrent Sync on your computer.

Photo Backup

BitTorrent Sync makes photo backup a trivial process. Simply launch the BitTorrent Sync app and select the BACKUP tab. Tap the add folder button and select the DCIM folder from your sdcard. Make note of the secret key that is generated.

Now, from the BitTorrent Sync client on your computer, add a new sync folder using the secret key generated on your phone and a location for the backup on your computer. Once complete, your devices should connect and being transferring data.

Note that, by default, BitTorrent Sync will not backup over cellular data, but that can be changed in the settings if you have the data available to do so.

App Backup with Titanium Backup

Note that Titanium Backup requires root to run.

1. Setup Scheduled Backups in Titanium Backup

   In order to keep an updated backup, you need to enable scheduled backups in Titanium Backup. There are two profiles that come pre-setup with Titanium; one that backups new apps, and one that checks for changed data in previously backed up apps. The defaults are reasonable, but I like to update them to run more frequently. I set the modified data check to run daily and the new app check twice a week (Wed and Sun).

   Check to enable both scheduled backups.

2. Setup BitTorrent Sync on Your Phone

   Just like with photos, you'll want to add a new folder to the BACKUP tab in BitTorrent Sync. The directory you want to add can be found in the Overview tab of Titanium Backup. The entry you're looking for is "Backup directory". It should be something like "/storage/emulated/0/TitaniumBackup". Make note of the secret key generated on the phone.

3. Setup BitTorrent Sync on Your Computer

   Again, from the BitTorrent Client on your PC, add a new folder using the secret key from your phone. The devices will connect and you're on your way to having a backup of all the apps on your Android phone.

Notes

• In order to have a continual backup, your computer will need to be on all the time. If that's a problem, you might benefit from a Raspberry Pi + BitTorrent Sync setup.
• By default BitTorrent Sync only backups over Wifi, this is probably what you want, but it can be changed if you have the data available on your cell plan.

The two tools that I'm now using are:

1. Git Annex + Git Annex Assistant
2. BitTorrent Sync

Both are really promising, but for different reasons.

Git Annex

Git Annex offers the ability to distribute files across many locations and uses Git to keep track of the location(s) any given file is available from. It also gives you the ability to recall a file from a remote location to your current device. This is really powerful for managing a large amount of data when you don't have room for it all on one machine.

The assistant functionality automates the syncing between repositories. It has a nice webapp which provides a GUI for managing remotes and sync strategies. These remotes can be additional git repositories, encrypted rsync destinations, local drives, EC2, and more. The assistant is still under active development, so it's not 100% stable or feature-complete, but it's definitely usable and the underlying functionality is pretty solid.

At the time of writing (July 20th) there's a crowd-funding campaign going on to support another year of full time development on git annex. If you want to support a good open source project, head ond over and contribute a few bucks (or bitcoins.)

BitTorrent Sync

BitTorrent Sync is the second tool I'm using for file syncing. It's much less complex then Git Annex. It's all about keeping folders in sync across multiple systems. The setup is pretty simple:

1. Add a folder to the client and generate a unique secret key.
2. From any other computer, choose a folder and give it the secret key.
3. Let the computers auto-discover each other, or specify an address.

BitTorrent Sync has clients for all the major desktop operating systems and an Android app.

What to use?

For the time being, I'm going to use both. I'm trying out Git Annex as a central store for anything I care about. Using it, I can setup a backup with the encrypted rsync remote as well as directly sync when I need to my different computers on an as-needed basis. This allows me to track all my old projects and miscellany in a single place.

BitTorrent Sync is going to handle more transient syncing needs. This includes things like my workspace and desktop wallpaper collection since they're ever changing and could be rebuilt from scratch if really necessary.

Web Access

In addition to the two syncing services listed above, I also have ajaxplorer setup on a remote VM and pointed at a BitTorrent Sync folder setup on the VM. This allows me to make files available when I'm on the road, at work or don't have access to a trusted machine with the client installed.

Further Reading

• BitTorrent Sync + Raspberry Pi
• Free/Open Source alternatives to popular syncing software

Periodically a story makes the top of Hacker News about someone who's just had their Google account disabled. Every time I read those stories, I think through what would happen if my account got disabled, and as you might imagine, it wouldn't go well for me, since I don't have any good friends who are Google employees who could bail me out.

I don't have any problem with what Google's doing, and I still intend to use some of their services, but it has become a massive single point of failure for me, and I don't have any recourse if something happens, since Google has no support.

I've decided to fix that by switching to separate providers for the services I depend on. Here's the breakdown of what I'm switching and what I'm keeping.

Switching

Email

This is the most important service I'm switching, since it's what I use and depend on the most. I've tried a number of alternatives, and I've decided to go with Fastmail, since they support custom domains, IMAP, SMTP, and have a nice web interface. It's also not expensive for 10GB of space at $40/year (or $32/year if you prepay 5 years). Fortunately, I've been using my own domain with Google Apps, so this switch will be completely seamless for everyone I communicate with.

Email Clients

As a side effect of dropping Gmail, I also needed to switch mail clients on my Android and iOS devices. The key feature I need in a mail client is identites, since I send mail from my custom domain, not the Fastmail account username. On iOS, this is simple, since the native mail client supports sending from a different email address.

Android however, does not support this in the native Mail application, so I had to find a good alternative. The one I ended up with is Kaiten Mail, which is excellent and well worth the $5. It has support for Identities, archiving mail and a plethora of customization options. My only gripe with it is the icon, which looks a lot like a trash can.

Contacts and Calendar

While Fastmail does provide contacts, they don't offer CardDAV syncing, which makes them unavailable on Android and iOS. My solution is to use a separate service, a self-hosted instance of Kolab for both Contacts and Calendar. This works with iOS and Android, since it has support for ActiveSync. It also has a very solid web interface for managing both.

I've also been experimenting with ownCloud's Calendar and Contacts, but I'm unimpressed with their web interface and I probably won't keep it. The main reason I'm trying it is the CardDAV and CalDAV support, since those protocols work well across all platforms, desktop and mobile. Kolab has announnced support for both in the next version, so I think it's still the better choice.

CalDAV/CardDAV Support (Mobile)

This is mostly for my own reference once Kolab has support for CardDAV and CalDAV - iOS has native support, Android does not, but there are good sync adapters available from Google Play for both - CardDAV-Sync ($2) and CalDAV-Sync ($2) which both support 2-way syncing.

Reader

I've been largely off of Reader for a while, since it's been obvious Google didn't care about it. The only place I still use it is on iOS, since it's the only service my favorite RSS reader (Byline) syncs with. Other than iOS, I've moved all my feeds over to both NewsBlur and Feedly, and have been alternating between the two. I'm waiting for the dust to settle after Reader shuts down to decide which I'll stick with.

Music

I'm setting up ownCloud to replace this as well. I'm using ownCloud's file sync to move Music up to a server and then for playback I'll use a separate tool (ownCloud has a built in music player, but it doesn't have many features for filtering albums or artists.) My intention is to setup, Ampache, Subsonic, Jinzora or something else on the server to use for playback and streaming.

This is an area where I might end up back with Google, since nothing I've tried has been very good and I also have everything in Amazon's CloudPlayer, so it's not really a single point of failure for me.

Keeping

Voice

I haven't found a good alternative to Google Voice; SendHub is close, but it's missing a number of key features like SMS to email and at $25/month for 500 minutes it's a little expensive on top of a cell phone bill. My current plan is to start experimenting with the Twilio API to see if I can duplicate the most important features of Google Voice myself.

Docs / Drive

I don't use Docs enough to merit finding an alternative and I already use Dropbox for syncing.

Search

Google still has the best search engine. It's also not strongly tied to my account, so I don't have to worry about losing data.

Summary

I have a few final thoughts on this process:

1. Google's great on the wallet

   Google's offerings are free, which is really hard to beat if you're trying to be economical. However, that comes at the cost of no support. Now that I'm out of college and have an income, but less time, I'm willing to pay for support and the assurance a service isn't going away.

2. Google's/Microsoft's/Yahoo's/etc... free offerings suffocate markets

   Looking around for good alternatives to products Google or Microsoft offer for free is really hard. There's no really good open solutions for online calendaring, or, until the recently announed death of Reader, RSS readers. Nobody wants to compete with Google.

3. Rent a dedicated server

   Nothing gives you more flexibility than having a big beefy dedicated server to run VMs on. When I want to test a new tool, all I have to do is clone a template VM and install it. If I like it, I keep the VM, if not, I delete or archive it and move on.

That's my experience switching from Google. If you have suggestions for good alternatives I'd love to hear about them. Feel free to post a comment or tweet @justjohn.

The solution to this problem when using VirtualBox is to setup two network adapters.

1. NAT Adapter

The NAT adapter allows the VM to share the host OS's network connection, including any VPN's it's connected to. The downside of NAT networking is that you can't directly access the VM from the host box using the assigned IP.

2. Host-Only Adapter

That previous problem can be solved by creating a second host-only adapter. This second adapter creates a separate network with the VM and host as members. This gives you an IP address for the VM you can access from the host for testing services on the VM or mounting a Samba/NFS/SFTP share to the host.

I never quite got around to posting here, but TransitWidget, an application started at a Google Technology User Group hackathon I participated in a while back is now available for free on Google Play. It supports all bus services that NextBus supports, including the MBTA, MTA and SF Muni. If you like the idea of having a widget on your home or lock screen that displays the time till the next bus, you might like it.

I've been experimenting with writing command line tools in node.js recently and it turns out to be a fantastic platform for doing so. It's really easy to get information from an API and format it for a console. As a result, I present console-weather, a simple tool for printing the current weather conditions based on a geolocation of your IP address.

You can install it via npm with

npm install -g console-weather

You'll need to get an API key from the Weather Underground API and store it in your WEATHER_API environment variable.

Usage: weather [OPTION]...
Print the current weather conditions and forecast.

Options:
    -c       Output temperature in celsius.
    -f       Print 4-day forecast.
    --color  Colorize the output.

I'll be posting more on the topic of command line node in the future; so if you're interested, stay tuned.

Github :: NPM

Changes: 

• Option to switch between 12 and 24hr modes.
• Options to enable or disable seconds
• flipclock.us now works better on mobile devices with smaller screens and in porttrait orientation.

 You can get the updated verions on the Chrome Web Store, or at http://flipclock.us/ 

MyGeode was a project I pitched last week at Startup Weekend Boston and developed largely over the weekend with the help of Amir Banihashem (design), Ryan Newton (website), Judith da Silva (business) and Mike Gioia (business). I've spent the last week finishing up the app and getting it ready for publishing.

I think the app has potential and I plan to continue developing it, it's based around the idea that tasks should have locations associated with them and using that location can help make you more productive. Imagine if your task list could alert you when you're near the grocery store, or when you get to work, it could know that, and remind you of your high priority tasks for the day. If that sounds attractive to you, take a look at the app and let me know what you think.

The version I released today is a solid todo list with the core location features: notifications when you near a task, a map view that dispalys all your tasks on a map and a task view that gives you all the information you need to know about a task. I'm excited about it and I've got a great list of features I plan on adding to the app, including routing capability to help you plan the most efficient route to complete your tasks.

I feel it's important to give a plug to Startup Weekend, since this is the second time I've gone and it's been an amazing experience both times. This time I have an app in the market a week after the event and I'm still actively working on the project from last year's event. If you're at all interested in startups or building a software product, you owe it to yourself to check out an event like Startup Weekend. It's 54 hours of insane fun and a great opportunity to connect with the local startup scene. If you do go, make sure to take advantage of the mentors available; it's very helpful to talk to people who've done it before and get their feedback during the early stages of developing your idea.

To conclude, you should buy MyGeode from the Android Market and let me know what you think!

res.header("Access-Control-Allow-Origin", "*");

I had a route that looked something like this:

app.get('/posts', function(req, res){
    res.header("Access-Control-Allow-Origin", "*");
    res.header("Access-Control-Allow-Headers", "X-Requested-With"); 
    res.send(
        { posts : ... }
    );
});

But it wasn't working in Chrome, I kept seeing the error "XMLHttpRequest cannot load http://... Origin http://fiddle.jshell.net is not allowed by Access-Control-Allow-Origin." even though I was setting the correct header.

Turns out that there was a OPTIONS request being sent before the GET request that didn't match my route and it was failing with the previous error. Problem is, Chrome didn't show the request in the Inspector's Network tab so it took me a long time to figure out what was happening.

The fix is a pretty simple one; you need to make sure your route in Express matches the OPTIONS request as well as the GET request. The easiest way to do that is to bind with app.all(...) to match on all requests. So the updated and working code is:

app.all('/posts', function(req, res){
    res.header("Access-Control-Allow-Origin", "*");
    res.header("Access-Control-Allow-Headers", "X-Requested-With");
    res.send(
        { posts : ... }
    );
});

It's also important to note that I needed to set the Access-Control-Allow-Headers header to allow "X-Requested-With". Without it, it generated a similar error: "XMLHttpRequest cannot load http://... Request header field X-Requested-With is not allowed by Access-Control-Allow-Headers."

$app->get('/user/{username}', function($username) {
    return "User profile for $username";
});

Combined with a powerful templating engine like twig that supports template inheritance, you can rapidly bind variations of a template to different url patterns. Using a base template with a standard page structure you can move all the common code for analytics, ads, etc... into it and you only have to write the page specific elements in each template. For storing data I ended up using Mongo for it's flexibility and because I could reuse components I've previously written for my foundry-core library.

For the front end I chose JQuery Mobile and was impressed by it's ease of use. It's pretty quick to get up and running with very little code required. They also have a decent amount of documentation on the basics which helped immensely when putting together the structure of the app but was lacking when it came to troubleshooting issues that popped up during development. My overall impression is definitely favorable, but I don't think it's quite good enough to replace native apps. There is a lot of loading time involved between pages that really slows down the experience. I think I could probably get around a lot of it by pre-loading templates and with some caching of content, but the effort required is probably better spent writing native apps. In it's current form it's a great tool for prototyping, but not yet a viable long term solution.

Unfortunately, we didn't win the contest on Sunday, that honor went to CaseReportal, a medical reporting/collaboration tool. But we do plan on pursuing our project since it has great applications and we were able to confirm the demand for such a product. If you want to learn more about it you can sign up for an account at http://pocketroster.me and try it out on your mobile device. If you think you'd like to use our product for your team or organization or if you have feedback for us you can contact us at feedback@pocketroster.me.

• Windows 7 Ultimate as the host OS

• VirtualBox 4 running the following 3 VMs (all Ubuntu 10.10)

• A file server running Greyhole and Samba for redundant storage accessible from everywhere

• A network core VM for locally hosting network services such as DNS and NIS. (oh, and a Minecraft server)

• A development VM with the appropriate toolset for PHP development (Netbeans 7 and Firefox).

The features that make this all work:

1. VirtualBox supports multiple monitors, so my development VM can span
   both my monitors, or only one if I need some windows-only tool on the other.

2. Windows is the host OS, so it has all the access to the hardware it needs
   for device syncing with my Android phone and my iPod Touch. It's also a
   plus when I want to do some gaming; I can close the development VM to
   free up resources and since VirtualBox saves state it takes mere moments
   to bring it back up when I need it (typically under 30 seconds.)

3. Ubuntu has excellent support for running in a VirtualBox container. I haven't
   had any issues with the VirtualBox guest additions and drivers.

In short, if you have a powerful enough machine, you can get the benefits of both
Windows and Linux simultaneously without having to give up anything due to
VirtualBox's great support within Ubuntu.

My project for the afternoon was a social wallboard. I had it in enough shape to do a short presentation by 5pm, but I didn't manage to get all the features that I wanted in (4hrs really isn't very long.) I have a good start however and intend to keep working on the project. Hopefully I'll have something usable in a few weeks after I've had a chance to solidify the OAuth authentication flow with the social services; I implemented the base code for Remember the Milk and Flickr during the Hackathon but I plan on adding a fair number of additional services over time (Twitter, Google Calendar, Delicious, etc...)

My primary takeaways from the day of hacking are that if you want to have a good demonstration after one afternoon you really have to cut back on the scope of your project (i.e. don't involve too many third party services; especially ones with OAuth) and it would have really helped to have a team of people to distribute the load across and brainstorm with. Overall though, it was totally worth it and I'm looking forward to the next chance I have to participate in a Hackathon.