Roommate App Development Blog » Uncategorized

Roommate App Is Secure!

Heff — Tue, 05 May 2009 04:12:01 +0000

Finally got the SSL certificate from Starfield set up. It was a bit trickier of a process because it’s a wild card cert, meaning it covers all subdomains of roommateapp.com as well. These are a lot more expensive too, though not too bad through my reseller account at Heffernan.biz. About $250.

If you want a cheap one, sign up for an account there, let me know, and I’ll set you up for discounted pricing.

Shhhhh! It’s launched.

Heff — Wed, 29 Apr 2009 05:15:32 +0000

Roommate App is live! Shhhhhh. : ) It’s available for anyone to sign up, but It’s not 100% tested, so I’m waiting to tell anybody. For those who find this and try it out, let me know if you find any bugs. Thanks!

Roommate App

Soooo close!

Heff — Mon, 20 Apr 2009 05:41:32 +0000

Very close to launching Roommate App. I’m at a point where I’m cutting back on features that don’t absolutely have to be there for launch, and making sure the ones that are there are solid. I even designed a super simple homepage.

I’ve been hearing/reading a lot about how if you aren’t embarrassed by your first release, you launched too late. I think I’m cresting that point right now. There’s a million things I want to add to the app and refine, but I need to just get it up and let my users (if any) decide what my next priorities are.

Also, I’ve been experiencing a lot of “unperfect” apps lately, so while I want the system to be totally solid, I’m letting go of my need for perfection a little. If that makes sense.

For instance I just wrote about my friend’s parenting website. It’s a great site, and looks totally clean, but there are a lot of kinks in the functionality. I know they’ll get worked out in time.

Also, we recently launched a video encoding app, and we still have a long way to go before I’d consider it a great user experience from beginning to end. But the point is it’s up and people are using it.

On the other hand, we just finished KDQOL-Complete, and while I’m cautious to say it’s perfect, I don’t think we could be happier with it.

Anyway, all that said, Roommate App is coming very very soon. I’ve told myself if I can’t get it launched before I leave Wisconsin in 2 weeks, I have to abandon the project entirely. Obviously I don’t want that to happen, so one way or another it’s going live. Can’t wait.

RESTful CSS Update

Heff — Tue, 07 Apr 2009 03:29:11 +0000

Quick update pertaining to RESTful CSS, a method for organizing CSS I wrote a little while ago.

I created a few extra helpers for the body tag. I’ve been adding IDs and classes to the body tag in the form of id=”controller_action” class=”controller action”, so I created these methods to both clean up the html, and to add the extra functionality of overwriting “create” with “new”, and “update” with “edit”. This is needed because of the way Rails handle form errors. Here are the methods.

def body_id controller.controller_name + "_" + { "create" => "new", "update" => "edit" }[controller.action_name] || controller.action_name end

def body_class controller.controller_name + " " + { "create" => "new", "update" => "edit" }[controller.action_name] || controller.action_name end

The body tag would then look like:

No more dashboard

Heff — Sun, 05 Apr 2009 05:45:56 +0000

I’m out in Wisconsin for the month of April working in the Sevenwire office, and getting a lot done. Including getting very close to launching Roommate App. Just cleaning some things up, and tying up some loose ends, though it feels like I’ve been doing that forever now.

My latest big change in the app has been replacing the “Dashboard” tab with a “Roommates” tab. After the last major overhaul I did with Nate where we added subdomains, some new concepts were added that I didn’t have a chance to think about within the scope of the entire app. This included the concept of Roommates playing a much bigger role in the structure of the app itself (who’da thunk?). In Rails terms, they got their own controller (though it kinda took over the Users controller). Anyway, now that I had chance to step back and look at all the concepts as a whole, I realized the whole dashboard section was just the missing index (listing) and show (details) pages from the roommates controller. So now they’re combined into one.

Also, I may be dropping the “Home settings” tab as well, which brings me down to two tabs, Bills and Roommates. How simple is that? It’s crazy how much thought has been required to make this app so simple. Hopefully other people think it is too. If not, they should see my first version.

ActiveForm part 2

Heff — Thu, 19 Feb 2009 08:46:58 +0000

Just following up on my own internal discussion from last night about my ActiveForm javascript library.

I started switching everything to extend the elements themselves instead of building the functionality into separate objects. One thing I found was there was no way extending the form element was going to work. There would be too many name space conflicts. One of them being “name”. The form already has a name, so if you were making a form for a user, it would possibly try to make “user.name”. That would be bad. So that was out.

As far as the inputs themselves, that went more smoothly. I’m not trying to add dynamically generated methods to those objects, so there’s little problem with name space. In fact, switching allowed me to drastically clean up my ActiveForm class. Before I had to use eval() to embed certain functionality into an input, but once I was extending the input directly this wasn’t necessary anymore.

I’m really liking how this is turning out, and I’m looking forward to using it on future projects.

ActiveForm

Heff — Wed, 18 Feb 2009 06:03:20 +0000

I’m writing this post mainly to think through a javascript library I’m working on.

The library is called ActiveForm, and the concept is similar to how ActiveRecord objects in Rails interact with the database. ActiveForm, or rather its children ActiveFormField objects get and set values from actual input elements, while utilizing railsish callbacks. Including:

before_validate()
validate()
before_set()
set()
after_set()

before_get()
get()

This allows me to do something like this, where I format a the value as an amount before setting actual element’s value:

before_set: function(value) {
return this.format_as_money(value);
}

Whether or not this is the best way to do things, I couldn’t tell you. I’m no javascript genius. But I’m having fun writing it and I’m enjoying they syntax it creates.

A question that I’m asking myself now, that I know will come up again, is why didn’t I just extend the form elements themselves? This would be similar to how Prototype does with “$(input).getValue();“. So I’m documenting it here so when I forget why later I can look it up.

For starters, it had to do with how I was thinking of the objects versus the actual elements. In my head, the elements were like the database in Rails and the objets were like ActiveRecord objects. ActiveRecord objects cache and process database values and only call and update the database when necessary. Forgive me if I’m explaining Rails all wrong, I’m still a relative newbie.

This metaphor loses a little when you know that the user edits the elements directly, but ignore that part. It doesn’t actually make that big of a difference since the ActiveFormField immediately grabs the value and processes it.

Another reason would be containment of the ActiveForm methods, so they don’t conflict with anything else. I don’t know if this is a huge problem though.

At the moment, I can’t think of any other reason, so I may have to try switching and see what I run into. I think it’ll be a lot of work though.

And now somehow hours have gone by in a blink, and I have to go to bed, where I’ll stare at the ceiling and think about this all night…

To subdomain or not to subdomain

Heff — Sun, 25 Jan 2009 18:17:14 +0000

I’ve taken a lot of inspiration for Roommate App from awesome apps like Blinksale and Basecamp. One feature of those apps that I hadn’t followed was creating personalized subdomains for each account. Examples would be “myaccount.blinksale.com”, or “myaccount.basecamphq.com”.

Besides not knowing exactly how to accomplish that, I didn’t really like that it limited any future interaction between homes. In Roommate App, this would mean if a roommate ever switched homes, they couldn’t transfer their info to the new home. They would have to create a brand new account and set all their preferences again, like avatars and payment preferences. So I decided to just keep accounts under the main domain.

Yesterday, Nate brought up the option again, and pointed out that there are security benefits to using subdomains, and that it simplifies the app a little. So we spent a little while discussing it, because it’s a super major decision that has a lot of implications.

One big downside to subdomains for Roommate App, is that it’s not as easy to come up with a memorable subdomain to represent a home that roommates live in. In the cases of Blinksale and Basecamp, it’s easy because businesses use those apps, and every business has a name. Roommate homes don’t have names.

So what do you call it? Do you name it by your address, like “123paperstreet.roommateapp.com”? You wouldn’t want to name it after yourself, like “steveshome.roommateapp.com”, because then that locks the home to you. If you ever move out, most likely the account will stay with your roommates and not you, so it wouldn’t make sense anymore. The problem with both of these is that there may be a lot of 123 Paper Streets, and I know there’s a lot of Steves. So there’s not one easy solution that everyone can go with.

The most important thing is that the subdomain is memorable, so roommates easily know where to go to log in. With subdomains they can’t just go to roommateapp.com to log in, they have to go to their subdomain. Fortunately, all the bill and payment emails that are sent out will have links to their subdomains, so that will help. It’s still a concerning issue though.

Benefits of using subdomains include:

Security – The subdomains are hidden, and require different cookies to keep a user logged in.
Simplicity – I no longer have to track if a user has multiple homes.
Scalability – Subdomains can be distributed to different servers as Roommate App grows

Those reasons were enough for us to choose the subdomain path. Hopefully it’s the right choice.

To accomplish this, we’ll be using subdomain-fu.

Rails is frustrating

Heff — Mon, 19 Jan 2009 01:26:32 +0000

It seems like anytime I start work on another part of my Rails app, I have to spend a long time fixing something else that’s not working. Mostly because I have no clue what’s going on within the Rails framework, and half the time I spend digging around to find the source of the error. It doesn’t help that I’m learning Ruby at the same time too.

I think PHP spoiled me. Whenever there was an error, it was right in my face and I could hunt it down quickly with no problem. And if I didn’t know the solution, there was always php.net or a wealth of discussions to read through. The Rails community is great, but for whatever reason I’m finding it harder to find answers. And once I do get it working, I rarely have done it the “correct” way.

I definitely developed some bad habits writing PHP. Never used a versioning system, worked directly on the server, never wrote tests, never even used the command line for anything. So I’m having to update myself on all of that at the same time. Pretty overwhelming.

Maybe it would be better if I were just doing Ruby, but throw Rails on top and I get totally lost. There’s so many moving parts involved that most of the time I feel like I shouldn’t even try to solve a problem myself, I should just look for how someone else did it. That takes a lot of the enjoyment of programming away for me.

Anyway, I can still see how Rails will be great once I understand it all better, but right at the moment it’s killin me.

Starting the new year

Heff — Fri, 02 Jan 2009 05:56:47 +0000

Well, obviously I didn’t get nearly as far as originally planned at this point, but it’s picking up momentum. Today Nate upgraded Rails and a lot of the plugins, and then helped me optimize my code a little. A big thanks to him. I also got this blog switched over from Movable Type to Word Press. It was surprisingly easy getting the old posts moved over.

You’ll notice in the side bar there’s a “Don’t Break The Chain” calendar. I’d like to try to work on Roommate App at least a little every day from here on out, so I’ll be tracking my success there.

Anyway, good start for the year. Hope I can keep it up.