Count It Off is my new web and iPhone app for simple calorie counting and weight tracking.

I’ve been trying to not be as large as I am for a few years now. I’ve tried a handful of ways to diet and have recently concluded that calorie counting is the best solution for me (everyone’s different). I started by using a Google Docs spreadsheet, but that got old fast. Then I tried a couple web apps and they all tried too hard. I didn’t want them to attempt to tell me how many calories were in my breakfast. I wanted to calculate it once, record it with a useful name, and then log it by that name the next time I ate it. It was during WWDC 2008 that the idea for Count It Off, a calorie counter for the iPhone, came together.

The week after WWDC (Jun 16th-ish, 2008) I was flying to Spain to visit the MakaluMedia guys who I’ve been (and still am) doing contract work for. The extra long flight from San Francisco to Spain gave me plenty of time to think about Count It Off. Here are the sketches along with some commentary and the final 1.0 screens:

Food Log

This is the first page the user sees when they open the app. A good bit changed here in 1.0. I didn’t like the Today: 1100 calories header that I sketched out. I played with a few ideas and settled on a Today - XXXX left header along with a progress bar as the first cell in the ‘today’ entries. I also summed up previous days’ consumption and apply a green check mark if you were within 200 calories of your goal. This makes it easy to see how good/bad you’ve been.

Weight Log

The weight page in 1.0 is a simple log that shows how many lbs or kg are left to lose before reaching your goal. This is an area I want to improve on.

New Food Item

The odometer style picker I sketched wasn’t going to work. The number pad entry is much better. In the sketch, you might have noticed that I was going to allow users to capture images of the food they are eating. This isn’t in 1.0 because I’m not really convinced it’s useful. I just wanted to have a reason to use the camera on the phone.

New Log Entry

The “Pick | New” segmented control was also a bad idea. The assumption with Count It Off is that you eat a lot of the same foods over and over. So by default it brings up the searchable ‘pick food’ view shown above:

Progress

This progress page doesn’t exist in 1.0 but it is something I still want to do.

Somewhat surprisingly, the sketches from June are pretty similar to version 1.0 of the iPhone application. Those sketches have a few good ideas that didn’t make it into the first version and some bad ideas that didn’t make it in.

Now, those sketches are four months old and I didn’t submit the application to the App Store until Nov. 8th, 2008. Did it take me over four months to build this app? Nope. I played with the iPhone SDK for a few nights back in June, but then things got busy and I shelved the idea when I got home. Then in early October I unexpectadly found some free time (that’s another story ). On October 14th I was still larger than I wanted to be, so I decided to make Count It Off a reality.

I first completed the web app and expecting to announce that before the IPhone version (they sync of course). But before I got the web version to a point where I liked it enough to talk about it, the iPhone version was well under way. I resisted announcing these either project until they were both ready. That day is today!

I’ll blog more about this journey later this week and will let you know when the iPhone app is available in the store.

Why did I do all this? The main reason was because the one dedicated server was real cheap and scared me. It was a single disk setup with zero redundancy and only nightly backups of critical data that I had to setup manually. I had gotten away with running in that setup for about a year and it was really only a matter of time before that machine failed me. If that hard drive or any part of the machine failed I’d have to rebuild the server from scratch and manually restore the critical data.

After moving to EC2 I have a custom image that I boot with all my configuration setup and an elastic block storage volume with all the database and other changing data stored on it. The volume is easy to take snapshots of and I feel pretty good about the whole setup.

The only gotcha I hit was when I tried to run two SSL virtual hosts off the same IP. Turns out you can’t do that. The recommended solution is to use separate IP addresses to differentiate the hosts. However, EC2 currently only allows you to bind one public IP per instance. I went with the solution to keep one vhost on the standard https port, 443, and the other on a non-standard port. Not pretty, but it’s working for now.

Some light details about the setup:

• Apache virtual hosts are serving the 8 domains
• mod_rails is taking care of the 4 domains that are rails applications. So far I’m very pleased with mod_rails - add a new vhost and bam, it just works. No monkeying with mongrel_cluster yaml files.
• Sending of mail is being taken care of by authsmtp.com. This is new for me but so far I’m happy. I always hate setting up sending or anything related to email on a server. Was happy to outsource this.
• DNS is now being hosted by https://www.dnsmadeeasy.com/. Not a pretty app, but it does the job at a decent price and it’s something I’ll rarely have to tweak.

Thanks to Jonathan for helping with the move and recommending the DNS host. It was nice to know someone who also did a similar migration in case things blew up (he recently moved Round Haus to EC2).

In summary I’m really happy with the new setup. We’ll see how long I stay happy.

note: I was fairly light on the details of the move (post was already long enough). If anyones interested, feel free to email me or comment here.

Enough of my crying. It only took a short amount of googling to find how to do this with Leopard’s terminal:

http://blog.macromates.com/2006/word-movement-in-terminal/ also has a tip for mapping control-left to beginning of line and control-right to end of line. Making the terminal act more editor like.

Glad that’s over with.

GoogleAuth.setup_captcha("some@email.com")

After that runs, the next time I try ClientLogin with that email address it’s almost guaranteed that the CAPTCHA requirement will happen.

The script is available at http://gist.github.com/11820

As you can see, the script just hammers away trying to log into the given account until Google returns with the CaptchaRequired response.

No this isn’t nice or pretty. Ideally Google would let us pass a parameter that would force the CAPTCHA requirement to happen. Until then, this works.

Recently, 37 Signals added a journal feature to Backpack. I wrote a little dashboard widget that uses their API and can post your status and create new journal entries.

Download BackpackJournal.wdgt

To use this, you’ll need your Backpack API Key. To get that, go to ‘my info’ and click the ‘Show your API key’

After that, enter your text in whichever field you want and hit enter. Only the field you changed will be submitted.

Why didn’t I just use a Web Clip (the safari feature where you can take a section of a page and make it a widget)? I tried this first and was unhappy with the results.

I’m not sure I’ll do much more work on this. I think I like that it’s write only. However, one could see updating it to read everyone’s status and growl notify when someone changes their status or writes a journal entry. But for now, I’m not planning on doing that.

Please let me know about any bugs you run into. Also, feel free to fork it and do with it what you will. The Dashcode project is in github as Backpack Journal.

Enjoy

gwd()
{
  $*
  growlnotify “${1} is done” -m ‘…’
}


Drop that in your .bashrc or .bash_profile and you can now do things like:

gwd sleep 1
gwd some time consuming command

Assumptions: * You have growl and the command line growlnotify installed * You use bash

Update: The friend I was talking with didn’t have growlnotify installed. His google search for ‘growlnotify install’ turned up this, much more complete solution: http://www.simplicidade.org/notes/archives/2007/08/tip_use_growlno.html.

What’s next? For the short term I’ll be doing contract work (already lined up) mixed with working on my own apps. Only time will tell where that goes.

A couple weeks ago I downloaded Potion Store for fun and started playing with the admin interface in a development environment. It was way better than anything I had and I wanted it. The only downside was that I would have to sign up for PayPal pro and would lose $30 each month. I justified that as motivation to do more frequent updates and release new apps. Hope it helps.

It took me 7 days of very part time work to finish deploying the store. The biggest pain was adding support for state taxes. I’m a resident of Idaho and am required to collect sales tax from anyone buying rooSwitch that also lives in Idaho. PayPal’s express checkout is not setup to handle this very well. If you’re interested in the details, I started a thread at pdncommunity. I still can’t figure out why more people aren’t talking about this issue. It makes me feel like I’m doing something wrong. (Feel free to educate me in the comments if you know what I’m doing wrong )

Apart from that little sales tax glitch, I’m happy with the results.

Happy 2008!

Here’s some quick status on the rooApps:

rooSwitch

The 1.2 release was started many months ago and took way too long to finish. I have ideas and user submitted feature requests that could either be rolled into a 1.3, but it’s more likely the next major release will be a 2.0. However, I don’t expect to see this happen before Fall 2008. This would likely be a Leopard only release and an almost complete rewrite (the are some things in 1.x that need rewriting )

rooVid

I finally abondoned hope of ever realesing a rooVid non-lite version on the 1.0 label. During WWDC 2007 I sketched out rooVid 2.0. In my mind it was three things:

• A video convertor featuring profiles (just like lite 1.0 is today)
• A video uploader (to YouTube, your .mac account, your personal website, etc)
• A video organizer

The feature I was most excited about was the organizer. Then iLife ‘08 with the new iMovie was announced and I decided to shelf rooVid 2 again. It seemed the organizer part and a lot of the upload and converting was taken care of by Apple in the new iMovie. Now, months after using iMovie ‘08 I realize I still want a better organizer and a better convertor and a better everything. I enjoy creating movies in iMovie ‘08 (I think I’m one of the only ones that likes the new interface) but the organizing and exporting leave room for rooVid.

rooVid 2 will definitely be a Leopard only app and I hope to have it available in late Spring 2008.

Misc

I know I’ve never been good at estimating when I’ll be able to finish an app or even point releases. I blame this poor scheduling on lack of time and poor time management of that limited time by me. Having a full time job, a family (Married with 2 small girls) and trying to do this work when possible can be hard, and very hard to do consistently without burning out. If all goes well things will be changing around the end of January/early February that will help me with this time issue and remove an excuse or two for me not getting things done. More news on that when it’s official.

Merry Christmas!

1. Turn off auto-refresh. I tell NNW when I’m ready for news.
2. Every morning I set aside 20 minutes, open NNW and fetch all.
3. During those 20 minutes I look at all the new entries and answer a question: Do I want to read or think about this more?
   1. Yes? I open the article in a new tab in NNW, in the background.
   2. No? I move on.
4. After each entry has been processed I move on with my day.
5. Throughout the day I come back and read through the tabs casually. When I’m done reading a tab, I close it.
6. Next morning I rinse and repeat.

NNW Zero: I think it’s important for me to always close all my tabs before the end of the day. I used to leave some open thinking they were important and I’d like to deal with them more later. I never do. If it’s really important, I write it down in my todo list.

Use Folders: A while back I grouped my feeds into folders. Some examples are: cocoa devs, apple, photos, life, design, … When I click on a folder to process the contents, the amount of attention I give to the contents varies from folder to folder. I’ve figured out which folders are more important to me and know which can be given less attention. I feel like this helps me process the incoming items even faster.

Feel free to leave your own tips in the comments (or links to your tips).