A fanatic is one who redoubles his effort when he has forgotten his aim

— George Santayana

Recently I had a weird rendering problem that only happened on Chrome for iOS and needed access to developer tools (debugger, DOM inspector, etc).

I knew about Safari for Mac allowing you to debug via USB and I had read that this also worked with any app that used any UIWebView .

But for some reason it didn’t work for Chrome. So I kept looking and found weinre :

weinre is a debugger for web pages, like FireBug (for FireFox) and Web Inspector (for WebKit-based browsers), except it’s designed to work remotely, and in particular, to allow you debug web pages on a mobile device such as a phone.

The basic steps to get it working:

1. Install it via node package manager:

   sudo npm -g install weinre

2. Run it on the console:

   weinre --boundHost <IP ADDRESS OF THE WEB SERVER>

3. Link the website you want to debug to weinre by adding this to the page

   <script src="http://<IP ADDRESS OF THE WEB SERVER>:8080/target/target-script-min.js"></script>

4. Load weinre’s remote debugger on your browser at

   http://<IP ADDRESS OF THE WEB SERVER>:8080/client/

5. Reload your website.

I used<IP ADDRESS OF THE WEB SERVER> everywhere, in my case it was the local IP of my development machine.

weinre worked nicely for me, and fast too. Can’t complain.

Now about the name:

weinre is WEb INspector REmote. Pronounced like the word “winery”. Or maybe like the word “weiner”. Who knows, really.

Yeah… naming things is hard.

I’ve been buying my Calgary public transit passes via the online store and recently they launched a new version called CITYonline. It looks ok, and it’s definitely a step up from the previous version. But in the 10 minutes that took me to register and buy next month’s pass I found some glaring usability and security problems .

First: you broke my browser

Disabling copy and paste is the new “right click is disabled” (so, 1997).

Dear Calgary: people use copy and paste, they like it and rely on it. When the iPhone didn’t have copy and paste people complained.

So, why do you disable it?

If you are worried that your database will get filled with incorrect emails because of typos then check the SMTP response headers. Or just accept errors as a part of life, but don’t make your lives easier by dumping shit on your users.

Second: so many options

On any listing of products, Calgary chose to use a select input element (marked with the red arrows).

Select elements (combo boxes in corporate parlance) are meant to give the usersmore than one option to choose from.

Each of those select elements have only one option. There’s nothing to choose!

This is a small problem and can be solved easily, but still shows that the people who made the site didn’t know what they were doing.

Third: I say “Yes”, but I may mean “No”

When you create your account and confirm your email (more on that on the next problem) you are required to select a password, and a helpful message tells you that you have to use_at least_ 7 characters.

Then, I tried to use 44 using 1Password and noticed that it looked shorter than 44 characters.

We’ll it turns out they tell you to use at least 7 characters, but don’t tell you that you have a maximum of 20 available.

That’s a very bad usability problem, it’s like saying to your kid “yeah, go out and party all night and come whenever” and “forgetting” to tell them that you won’t open the door when they arrive at 3 in the morning =).

BTW: they do the same thing with emails, limiting them to 50 characters. The maximum length of an email address is actually 254 characters .

Fourth: here are the keys to my kingdom

Registering is a two step process in CITYonline:

1. You write your name and email (no copy/pasting!)
2. You receive a temporary password to your email.

When you go and login with said temporary password, it forces you to set a new password. Very safe.

And then they send you FRIGGIN’ PASSWORD TO YOUR EMAIL!!!

Never, ever send a password to the user’s email EVER!

1. Email is not a secure enclosure.
2. It potentially means that the City of calgary is storing the passwords in a readable way.

Good thing CITYonline doesn’t store your credit card information (or does it?).

Fouth plus one: the emails suck

This one is just a pet peeve.

The email I received from CITYonline has header image of 1029 pixels wide, this wreaks havoc with many email clients as you are forced to scroll back and forth to read the whole thing.

Just use an email template from the pro’s .

I’ll just say it: I barely worked on the app this week. I’ve been feeling ill and most of my evenings were spent either watching TV or studying for the citizenship test (I’m writing this as I wait for the test to begin).

Week 2: map/reduce in Swift

I did work on it a bit. Mostly reading and trying to figure out the syntax for closures in Swift to do map/reduce.

There are many different ways to write an Array.reduce; I like the succinct syntax of Swift but I think it allows succinctness to be taken to an extreme. For example doing a reduce to add up all elements can be this:

myArray.reduce(0, {$0 + $1})

That’s something you’ll need to re-understand every time you read it mostly because of those$ signs^[1]. It also feels dirty coming from ObjectiveC where everything had names. Long names. I liked that.

You can also give names to$0 and$1 which is nicer:

myArray.reduce(0, {x,y in x + y;})

I also like that because operators in Swift are just functions you can pass it like this:myArray.reduce(0,+).

So, there are many levels of succinctness depending on the amount of hair you want to lose when it’s time to read the code again after some time.

Personally, I like the last one for one-liners but for more complex transformations I’m gonna go with this syntax:

myArray.reduce(0) { x,y in
    return x + y;
}

I’ve also used named functions instead of closures just to keep code readable.

myArray.reduce(0, addUpElements);

All this was explained with more details by Silviu Pop on this article: Higher Order Functions: Map, Filter, Reduce and more – Part 1 .

Back to work

I’m behind schedule by a week and need to catch up. I’ll be alone at home next week while Mary goes to The Grace Hopper conference so I’ll probably will get more time into this. Hopefully. Unless Netflix wins again^[2].

Most of the work is run as automated unit tests which is good because it has allowed me to quickly prototype and change the way I structure and massage the data.

Week 3?

Today (Oct 7) is already the middle of the week 3 so things are still piling up. Week 3 is about starting to do some UI work for basic things like shopping list creation and going shopping with it to start looking at real life data point (event if just for one user: me). I’m planning to go shopping today so that’s probably not gonna happen.

I’m sad and I’m hitting publish now.

When Apple showed us the Apple watch I decided I wanted to build an app for two reasons:

1. I want to make something for it and see how it fares in an environment not as overcrowded as the iPhone App Store.

   I don’t think the Apple Watch will produce the same gold rush era like the one the iPhone produced. Or at least it won’t be as long. That’s why I want ot try to get an app in before it gets overcrowded.

2. I want one.

   I’ve been considering getting a Fitbit for a while to track steps and see if I can get myself to walk and exercise more. But with the recall of the Fitbit force I put a hold on the purchase and waited to see what Apple did.

The app I’m building is not a new concept, especially not in the App Store: a shopping list app. But the idea I have on how to integrate it into the Apple Watch may be innovative and if I have it ready for launch day I’m sure it will be interesting.

Exactly the day I started working, Justin Jackson posted this tweet . It was very good timing!

The cold months are coming:

So that leaves three months. If you want to build your side-project, this is your window.

Here are a few steps you can take:

1. It’s September 20th today, that means launch day would be December 20th. Envision what you’d like to launch on that day.

2. Now open up your calendar. Work backwards from December 20th and schedule all the times you’ll work on your project. It’s best if you can make it regular weekly times: ie. Saturday mornings from 6am-10am, and Friday nights from 8pm-11pm.

3. Create a landing page. Put only 4 things on it: a headline (that says who this is for, and what their pain is), a sub-headline (describing what you’re offering), an email sign-up box (for folks to get updates), and a launch date (December 20th!).

I’ve done the calendar part and and have split the work over the three months and I plan to publish a post a week describing some of the things I’m learning and doing. Which brings me to the next point…

I haven’t done the landing page yet and to be frank it intimidates me to do that. Maybe I will create it just for the sake of having a reference for people reading this posts documenting my progress.

For now, I’m sorry if the following is too vague.

Week 1

Week 1 was all about two things:

1. Setting up the data model for creating lists, registering products and then marking products as picked in the store. I used Marco Arment’s FCModel which made things super simple. A lot easier than Core Data.
2. Analyzing that data to estimate which products you should be buying soon. I didn’t get too far into this. Partly because I got busy mid week^[1] and mostly because I’m still developing the idea of what data I want exactly.

There’s also a non product related goals I wanted to hit which was starting to write some of the code in Apple’s new Swift language.

I’d say I hit 80% of what I wanted to achieve.

I’m not sure if the app will be successful^[2] or not but I’m committed to getting it done.