<blockquote>
The <a href="http://www.asp.net/identity">ASP.NET Identity</a> system is designed to replace the previous ASP.NET Membership and Simple Membership systems. It includes profile support, OAuth integration, works with OWIN, and is included with the ASP.NET templates shipped with Visual Studio 2013.
</blockquote>
In and of itself, it is a very nice and pleasant library that really does integrate nicely into a WebAPI or MVC project. The template project that comes with VS2013 spins up very quickly and provides the illusion of something that the developer can just forget about. Were that the case, I would not have this post.
All I wanted was to change the default primary key from a <code>GUID</code> to an int. My initial thought was to change the default <code>ApplicationUser</code> from inheriting from the non-generic IdentityUser to the generic <code>IdentityUser&#x3C;int, VigilUserLogin, VigilUserRole, VigilUserClaim></code>. Compiling worked just fine, running caused all kinds of type casting problems. Why? Because all of the other classes surrounding the <code>IdentityUser</code> were all coded to the <code>string</code> type (which internally is a <code>GUID</code>).
Thus began the big hunt to find out all of the places where I needed to override, inherit, or otherwise manipulate the library to get the primary key to be an integer. This meant replacing the five main classes that create the authentication and authorization schema:
<ul>
<li>IdentityUser</li>
<li>IdentityRole</li>
<li>IdentityUserRole</li>
<li>IdentityUserClaim</li>
<li>IdentityUserLogin</li>
</ul>
Then I had to create a new <code>UserStore</code> and <code>UserManager</code> to utilize these new classes, and create a new <code>ClaimsPrincipal</code> to store and retrieve the new security identifier (Sid) for the user. Finally, a new <code>BaseController</code> class was created that all other controllers must inherit from, in order to utilize all of this new code.
<pre><code class="language-csharp">using System;
using System.Security.Claims;
using System.Web.Mvc;
using Vigil.Data;

namespace Vigil.Web.Controllers
{
 public abstract class BaseController : Controller, IDisposable
 {
 public VigilClaimsPrincipal CurrentUser
 {
 get { return new VigilClaimsPrincipal((ClaimsPrincipal)this.User); }
 }
 protected readonly IVigilContext VigilContext = new VigilContext();

 public BaseController() { }

 protected override void Dispose(bool disposing)
 {
 if (disposing &#x26;&#x26; VigilContext != null)
 VigilContext.Dispose();
 base.Dispose(disposing);
 }
 }
}
</code></pre>
<pre><code class="language-csharp">using System.Data.Entity;
using Microsoft.AspNet.Identity;
using Microsoft.AspNet.Identity.EntityFramework;
using Vigil.Data;
using Vigil.Data.System;

namespace Vigil.Web
{
 public interface IVigilUserStore : IUserStore&#x3C;VigilUser, int> { }

 public class VigilUserStore : UserStore&#x3C;VigilUser, VigilRole, int, VigilUserLogin, VigilUserRole, VigilUserClaim>, IVigilUserStore
 {
 public VigilUserStore() : base(new VigilContext()) { }
 public VigilUserStore(DbContext context) : base(context) { }
 }

 public class VigilUserManager : UserManager&#x3C;VigilUser, int>
 {
 public VigilUserManager() : this(new VigilUserStore(new VigilContext())) { }
 public VigilUserManager(IVigilUserStore store) : base(store) { }
 }
}
</code></pre>
<pre><code class="language-csharp">using System;
using System.Security.Claims;
using Vigil.Data;
using Vigil.Data.System;

namespace Vigil.Web
{
 public class VigilClaimsPrincipal : ClaimsPrincipal
 {
 public VigilClaimsPrincipal(ClaimsPrincipal principal) : base(principal) { }

 public int UserId
 {
 get { return Int32.Parse(this.FindFirst(ClaimTypes.Sid).Value); }
 }

 public VigilUser VigilUser
 {
 get { return (new VigilContext()).Set&#x3C;VigilUser>().Find(UserId); }
 }
 }
}
</code></pre>
<pre><code class="language-csharp">using System;
using System.Collections.Generic;
using System.Data.Entity;
using System.Data.Entity.Infrastructure;
using System.Data.Entity.ModelConfiguration.Conventions;
using System.Data.Entity.Validation;
using System.Threading;
using System.Threading.Tasks;
using Microsoft.AspNet.Identity.EntityFramework;
using Vigil.Data.System;

namespace Vigil.Data
{
 public interface IVigilContext : IDisposable
 {
 DbChangeTracker ChangeTracker { get; }
 DbContextConfiguration Configuration { get; }
 Database Database { get; }
 DbEntityEntry Entry(object entity);
 DbEntityEntry&#x3C;TEntity> Entry&#x3C;TEntity>(TEntity entity) where TEntity : class;
 IEnumerable&#x3C;DbEntityValidationResult> GetValidationErrors();
 int SaveChanges();
 Task&#x3C;int> SaveChangesAsync();
 Task&#x3C;int> SaveChangesAsync(CancellationToken cancellationToken);
 DbSet&#x3C;TEntity> Set&#x3C;TEntity>() where TEntity : class;
 DbSet Set(Type entityType);
 }
 public class VigilContext : IdentityDbContext&#x3C;VigilUser, VigilRole, int, VigilUserLogin, VigilUserRole, VigilUserClaim>, IVigilContext
 {
 public VigilContext()
 : base("VigilDb")
 {
 Database.SetInitializer&#x3C;VigilContext>(new VigilDbInitializer());
 }

 protected override void OnModelCreating(DbModelBuilder modelBuilder)
 {
 modelBuilder.HasDefaultSchema("vigil");
 modelBuilder.Conventions.Remove&#x3C;OneToManyCascadeDeleteConvention>();
 }
 }
}
</code></pre>
<pre><code class="language-csharp">using Microsoft.AspNet.Identity;
using Microsoft.AspNet.Identity.EntityFramework;

namespace Vigil.Data.System
{
 public class VigilUser : IdentityUser&#x3C;int, VigilUserLogin, VigilUserRole, VigilUserClaim>, IUser&#x3C;int>
 {
 }
 public class VigilRole : IdentityRole&#x3C;int, VigilUserRole>, IRole&#x3C;int>
 {
 }
 public class VigilUserRole : IdentityUserRole&#x3C;int>
 {
 }
 public class VigilUserLogin : IdentityUserLogin&#x3C;int>
 {
 }
 public class VigilUserClaim : IdentityUserClaim&#x3C;int>
 {
 }
}
</code></pre>
At this point, I am wrestling with the fact that now my domain model has a hard requirement on the <code>Microsoft.AspNet.Identity</code> library. I had initially hoped to have my <code>User</code> and <code>Role</code> domain models be completely separated from the actual authentication/authorization system. This creates a hard link that is going to be the death of me unless I can find a way to isolate the framework and run unit tests without requiring the database. For now, though, I am going to continue in this direction, and turn a blind eye to the technical debt accruing from this decision. Maybe before I ship this, I will find a better way to do it all, but for now at least I have my integers.

All I Wanted Was an Integer Primary Key

For any serious runner, looking through training plans inevitably leads to the discovery that they all want you to do hill repeats. Hal Higdon (basically the godfather of running in Chicago), recommends that <a href="http://www.halhigdon.com/training/51142/Marathon-Advanced-2-Training-Program">hills be a quarter-mile</a> or longer. If you have never been to Chicago, allow me to clue you in - the city is flat. The largest and longest "hills" we have are the on and off ramps for the expressways. Finding long stretches of elevation are difficult, and I cannot think of any that are a full quarter-mile long. However, here are a couple of spots that I have found that make for at least a somewhat noticeable incline.
<ul>
<li><a href="https://goo.gl/maps/j2txxh9e2oK2">Lincoln Park</a>, just north of North Pond, near the entrance to the Diversey Yacht Club. It is long enough to at least make you feel it, though it is not very steep. There are lots of trees around, so it is nice and shaded during hot days.</li>
<li><a href="https://goo.gl/maps/xVUTgSYVe3y">Cricket Hill</a>, between Lake Shore Drive and Montrose Beach. This staple of the Montrose Beach area has a gentle incline for the first half, and then a steep incline for the last half, rising about 45 feet from the base. The north west side of the hill is currently <a href="https://www.dnainfo.com/chicago/20140514/uptown/playing-fields-by-cricket-hill-getting-2-million-upgrade">under construction</a> while the city puts in a new turf field and an <a href="http://james46.org/wp-content/uploads/2014/04/Lincoln-Park-Wilson-Field-and-Track-Site-Plan.pdf">8-lane running track</a>. This is going to get a lot of runner attention once it is finished.</li>
<li><a href="https://goo.gl/maps/ZiogKX4rLK72">Solder Field South Lot</a>, south of Burnham Harbor and Solder field, accessible via the Lakefront Trail, this sledding hill is a short but steep incline. I would estimate that it is about 200 meters from bottom to top, and probably a 30 foot rise. Nothing spectacular, but we take what we can get.</li>
<li><a href="https://goo.gl/maps/WtgmjEG9AWE2">Horner Park</a>, bounded by Montrose, Irving Park, California, and the river. The incline is not very steep, and I would be surprised if the hill crested at more than 15 feet. However, it is about the longest stretch of incline you will find in a city park.</li>
<li><a href="https://goo.gl/maps/Lw5HyrTCFMK2">Humboldt Park</a>, next to Little Cubs Field - more of a grassy knoll than a hill, at least it is something. Also, there are lots of paths all around Humboldt Park, making for a nice run. While there, check out the gorgeous field house.</li>
<li>Parking garages - these are spread throughout the city, with lots of them being downtown or at movie theaters. I have used the one at the <a href="https://goo.gl/maps/Pz3JLRM45SU2">Regal City North</a> a couple of times. It is not as fulfilling as running on the grass hills, but I find the concrete and cars are still better than trying to run on a treadmill.</li>
<li>Treadmills - the ugliest of the suggestions, but sometimes it is the only way. Set the thing to the steepest it will go, and start out slow.</li>
</ul>
If I find any others, I will update the list. My runs are almost exclusively on the Lakefront Path and on the north side of the city, but if I hear of anything elsewhere in the city, I will gladly share it.

Hills In Chicago

It has been almost over ten years since I last bought a backpack, and now I need a new one. My three primary use cases are while commuting on a Divvy bike, riding on my motorcycle, and generally walking around. It needs to be comfortable and secure in all three scenarios. Other requirements (in no particular order):
| Requirements | Nice To Haves |
|-|-|
| Padding and ventilation on my back | Separate laptop sleeve/compartment |
| Enough space to hold 15" laptop | Two separate large compartments |
| Pocket with partitions for pens, cables and other small stuff | Exterior side pockets for waterbottle |
| Compression straps for lighter loads | High and low compression straps |
| Sternum strap | Waist strap |
| Rain proof | Built-in Backpack rain cover |
| Secure fit with heavy and/or bulky loads | Harness pocket for Divvy key, hip pocket for phone |
| Ability to secure all loose straps[^2] | Reflective portions for night riding |
After searching around the internet and asking a few friends, I have narrowed it down to three different brands - based on quality, price, reputation, and what I have seen on the backs of other riders. Timbuk2, Chrome, and Osprey. In alphabetical order, these are the leading contenders right now.
<ul>
<li><a href="http://www.chromeindustries.com/us/en/bags/backpacks/bravo-night">Chrome Bravo Night</a> - the cross compression straps on this back are extremely inviting, though its large roll-top opening is a turn-off.</li>
<li><a href="http://www.chromeindustries.com/us/en/bags/backpacks/cardiel-fortnight">Chrome Cardiel Fortnight</a> - I like the profession look of this bag, and the pockets seem well organized. One of the pictures shows a pair of shoes in the bag, but something tells me that my colossal size 15 running shoes will not fit quite as well. The bag certainly does not look as large as other 40 liter packs.</li>
<li><a href="http://www.chromeindustries.com/us/en/bags/backpacks/citadel">Chrome Citadel</a> - I really like the "large zip-around main compartment", in theory it makes getting items out of the bag easier. I am not sure about having straps that wide on the harness, though. Another point where I am going to have to try it on it determine if it is a pro or a con.</li>
<li><a href="http://www.ospreypacks.com/en/product/commute_1/momentum_30">Osprey Momentum 30</a> - same as the Radial, no pictures of the backing, though it seems to have a low profile on the back - good for unobstructed views over my shoulder</li>
<li><a href="http://www.ospreypacks.com/en/product/city__trail/quasar">Osprey Quasar</a> - This would not have even made the list if they did not happen to have it in stock for me to try at a store I went to. However, it was very comfortable with lots of pockets and secure fitting.</li>
<li><a href="http://www.ospreypacks.com/en/product/commute_1/radial_34">Osprey Radial 34</a> - the website does not show what the back padding/ventilation is, so this might get written off quickly once I get my hands on it.</li>
<li><a href="http://www.timbuk2.com/especial-medio-cycling-backpack/435-3-2001.html">Timbuk2 Especial Medio</a> - the entire Especial line looks really nice, and this is my favorite of the three. The feature list is extensive (the video shows them off), though I wonder if the accessory pocket is going to open enough to get to the bottom of the pocket without having to blindly dig around. Overall, this seems to be the just right in sizes between the Tres and Vuelo.</li>
<li><a href="http://www.timbuk2.com/tb2/products/especial-tres-waterproof-cycling-laptop-backpack">Timbuk2 Especial Tres</a> - the roll-top of the backpack could be too much of an inconvenience for me. With a 40.0 liter capacity, it may be too large for what I need.</li>
<li><a href="http://www.timbuk2.com/especial-vuelo-cycling-laptop-backpack/458-3-2182.html">Timbuk2 Especial Vuelo</a> - the smallest of the Especial line, I think its lack of features and smaller size might not suit my needs. I want to get a hold of it though, to find out if it is larger than it appears.</li>
</ul>
<h2>Late Additions</h2>
Two different sources suggested that I take a look at Deuter, a company that I had never heard of before.
<ul>
<li><a href="http://www.deuter.com/us/us/hiking/futura-28-34214-119.html">Deuter Futura 28</a> - A friend went through a similar exercise, settled on this, and has been gracious enough to let me borrow it for a little bit to see how it fits. Unfortunately, it turned out have a frame too high to comfortably wear on a bicycle.</li>
</ul>
<h2>Rejects</h2>
While most of these are not inherently bad backpacks, they just do not fit my purpose.
<ul>
<li><a href="http://www.ospreypacks.com/en/product/tech/pixel">Osprey Pixel</a> - main compartment is too small for my needs.</li>
</ul>
<h2>Honorable Mentions</h2>
There were some bags that I came across that had neat little features that are of no use to me, but I thought I would give them a little credit for creativity.
<ul>
<li><a href="http://www.ospreypacks.com/en/product/tech/pixel_port">Osprey Pixel Port</a> - contains a clear plastic pocket under the main flap to hold a 10" tablet (read: iPad), which allows use without removing it from the bag.</li>
</ul>
Some points of interest that I have found in my research. Osprey's website is tolerable, Chrome's is acceptable, and Timbuk2 blows them both out of the water. For no other reason than each of Timbuk2's bags include a video showing all of the different pockets, which is as close as I can get to touching and feeling the bag without it being in my hands. Before I went to a store to actually see how they felt on my back, I had ruled out all of the Osprey offerings. They would are only still on my list because a sales associate at the first place I went to suggested I try the Osprey Quasar bag. Furthermore, all of the sites need to show what the back looks like on tall models. <a href="http://www.cdc.gov/nchs/data/nhsr/nhsr010.pdf">5'9' may be average</a>, but their consumers are not going to all be average.
I will be updating these bags as I get some hands-on time with them, and then I will eventually post about which bag I chose and periodic updates.

I Need a New Backpack

After a month of having a Divvy membership, I have been on only nine trips. A slacker, I know; but with two weeks of vacation in there, I'm averaging one trip every other day. As the weather improves, I will certainly be upping that to my goal of seven out of ten trips in a week.
But here-in lies a problem that I have discovered. The Divvy rental is for a thirty minute period, and that is terribly close to the amount of time it takes me to make the trip <a href="https://goo.gl/maps/rX8KRLvEU6U2">between LaSalle &#x26; Washington and Racine &#x26; Belmont</a>.
<a href="https://goo.gl/maps/rX8KRLvEU6U2"><img src="/images/divvy-route.png" alt="Divvy Route" title="Divvy Route"></a>
For my first three trips I averaged 27:53, which is dangerously close to the 30 minute allotment. For the next six trips, I took advantage of the "unlimited" rides feature of my pass. In order to be able to take a more leisurely ride home (and be less worried about hitting every damn stop light along the way), I have been stopping at some point along the route, docking my bike, then borrowing the same bike again.
I feel a little like an urban Pony Express rider.
For these dock-and-swap rides, I had two go over the (combined) thirty minute threshold, three below it, and one at exactly thirty combined minutes. Considering that it only takes about 15 seconds for me to dock and undock the bicycle, I think this is the best way for me to get between home and the office without having to worry about going over that magic half-hour mark.
In chatting with other people around the office, it seems that most people think that thirty minutes is just barely too short, and they have shied away from getting memberships because of this. As a comparison, NYC's Citi Bike allows 45 minutes for members, and Phoenix and St. Paul / Minneapolis give a 60 minute grace period for members (both of which are also operated by Bixi). All other cities with 1,000 or more bicycles had a thirty minute duration. These include Boston, Long Beach, Miami Beach, San Diego, and Washington, D.C.
Taking a look at time estimates from Google Maps, for a conservative rider that wants to keep their ride under an estimated 25 minutes, Divvy is basically limiting their customers to only those south of Fullerton, north of 35th St (you can maybe make it to a Sox game!), and east of Garfield Park. By my estimates, just adding 15 minutes to the allotted time will open up a much larger segment of Chicago, even if that time is only offered to members. I really wish I had much better data on population segments, household incomes, commuter routes, and wear costs per mile on the bicycles. Oh the fun I could have with all of that data.
Alternatively, they could just advertise the #PonyExpress option (I'm going to make this happen).

A Month With Divvy

The short version of the trip:
<ul>
<li>Saturday, 4/12: ORD to TU</li>
<li>Sunday: Air show at Davis-Monthan AFB, "what's in your fridge?" ad-hoc dinner</li>
<li>Monday: bums on a couch, grocery shopping, BNP&#x26;C</li>
<li>Tuesday: bums on a couch, hospital trip</li>
<li>Wednesday: bums on a couch, failed grilled chicken</li>
<li>Thursday: bums on a couch, bookstore, "gnocchi"</li>
<li>Friday: bowling, Arizona Pizza, and Jeremy's baseball game</li>
<li>Saturday: Jeremy's baseball game 2, Pima County Fair, and Wings on Broadway</li>
<li>Sunday: visit to Phoenix, Easter dinner by Alyssa</li>
<li>Monday: bums on a couch, grocery shopping, brats on the grill</li>
<li>Tuesday: Tucson Zoo, Poco &#x26; Mom's</li>
<li>Wednesday, 4/23: TUS to ORD</li>
</ul>

April 2014 Arizona Trip

In terms of reaction to a request for work, there are several responses. In order from most desirable to least desirable:
<ol>
<li>Successful action with a message of success.</li>
<li>Failed action with a message of failure.</li>
<li>Successful action without a message.</li>
<li>Successful action with a message of failure.</li>
<li>Failed action without a message.</li>
<li>Failed action with a message of success.</li>
</ol>
I ran into an issue this week where users were saying that an action they were attempt to perform was reporting success. However, the later steps in their workflow were showing that the original task had failed. This was a disaster for the user experience. I dug into the code and found the offending switch. It was an entirely acceptable failure and a very common occurrence. However, because it was reporting success, users were obviously confused and overall productivity quickly ground to a halt. When my success is determined by the productivity gains that my software provides to others, this is clearly a very serious concern of mine.
From the time the issue was first report to me to the time it was fixed was approximately four minutes. No one in the organization really gave it a second thought, and merrily they went about their day. I, however, was thoroughly distraught. How did this pass through automated tests, manual testing, and user testing, and find its way into the production environment?
Today, I finally rearranged the entire testing suite, streamlining both the ability to add a new test and to determine code coverage. In reporting back to my boss about what I did this week ("re-factored the testing procedures"), he was thoroughly confused why I spent more than twenty hours on something that was fixed in four minutes. Even the explanation of "so it won't happen again" only slightly satisfied his question. However, as soon as I took the deep breath to fully explain why I did what I did and how it will help, he just cut me off say "it won't happen again? it's fine?" and he moved onto other things. It is really comforting knowing my boss has this level of trust. That he knows that whatever I have decided to spend my time on is almost assuredly results in the best service to our organization.
Or maybe he has no trust in me, but I am reading it as a message of success.

False Errors Are Better Than Silent Failures

<img src="/images/data-layers.png" alt="Data Workflow" title="Data Workflow">
This should be a footnote, but I have always wanted an excuse to use the stencils from the Visio Work Flow Diagram. They just seem like a fun way to represent something.
What does it take, if almost everything is created by hand, to do something as simple as add a new field to a table in the database:
<ol>
<li>UPDATE TABLE script</li>
<li>Edit DBML xml file (let Visual Studio regenerate the designer file)</li>
<li>Edit the POCO class and any mapping methods</li>
<li>Edit the JSON serialization method</li>
<li>Edit the Knockout.js Model or ViewModel object</li>
<li>Edit the HTML template</li>
</ol>
This is just for displaying the content. Additional work is required to allow for editing values and pushing them all the way back to the database. This lead me down the road of finding how I can eliminate steps. I reasoned that there had to be solutions already created that would took care of any least one of the various steps. Thankfully, I was able to find solutions for most of it.
Entity Framework Code First (ef-cf) has been the most promising tool for me to use to cut out the step 1 and roll it into a replacement for step 2. AutoMapper took care of replacing mapping methods in step 3 (though I still have to update the POCO class). Json.Net has been my library of choice for serialization. Knockoutjs.Mapping plugin too care of Step 5.
At this point, updating the data model looks like this:
<ol>
<li>Edit EF model.</li>
<li>Edit POCO class.</li>
<li>Edit HTML template.</li>
</ol>
This is clearly more manageable. I am putting together a tutorial all about taking a project from scratch and implementing all of the pieces needed to get this configured. If I stopped having so many false starts, I am sure the post (and the project) would be finished by now.

Rovani's Sandbox

Blog Posts - Page 13