In my ever-changing world, there interview questions that are well placed and worth my time and then there are ‘trick’ questions that are (to me) more of a personality test than a technical test (based on the technology actually used in the workplace). It is better to ask about personality directly in an interview rather than indirectly inquiring (either deliberately or out of ignorance).

Adam is a Hash Table

These are the questions I’ve encountered (that I could not answer) over the last few weeks that are worth my time:

  • What is the difference between the Logical Tree and the Visual Tree? This question can be considered a fundamental WPF question—but it is the kind of question that only comes from experience with advanced scenarios—or by reading some extensive prose meditating on how WPF works. In 2007, Josh Smith does exactly that in “Understanding the Visual Tree and Logical Tree in WPF.” As Jason Dolinger of lab49 suggests in “Exploring the WPF Logical and Visual Trees,” the logical tree closely represents programmers intent (what is declared by the programmer in XAML) while the Visual Tree represents what WPF or Silverlight is doing under the hood to actually render what is intended.
  • What is _ViewStart.cshtml? I actually was very warm on this one. This is a great question to test for MVC3-level intimacy. I have not been working with MVC on a daily basis over the last three months so I’m not too upset for not getting this one.
  • What is the purpose of the yield keyword? This question opens a whole new subject of computer science under the heading “state machine”—according to “Yield is implemented by the compiler as a custom class that implements a state machine.” Erik Forbes: “Yield is used to create implementations of the Enumerable pattern—a software pattern that allows you to treat a collection of things as an enumeration, over which you can perform some process.” MSDN: “The yield statement can only appear inside an iterator block, which can be implemented as the body of a method, operator, or accessor.” I really should know about this topic. No “excuses” here.
  • What is the difference between ref and out parameter modifiers? “…semantically, ref provides both ‘in’ and ‘out’ functionality, whereas out only provides ‘out’ functionality.” Where in refers to initializing the variable/parameter (with the expectation that it will be read inside the method accepting the parameter). This implies that out suggests that the parameter is write-only. I cannot think of a situation (outside of Interop and TryParse patterns) where using ref or out parameters are required. I prefer to use a struct or a class (stacking complexity vertically) over using multiple parameters (stacking complexity horizontally).
  • How can custom validation be centralized/reused by several view models? The CustomValidationAttribute accepts a Type and a method-name string where the Type points to a central validation class. I have actually written code using this kind of centralized custom validation—but clearly not enough to memorize this level of detail during a job interview.
  • What is the volatile keyword? “The volatile modifier is usually used for a field that is accessed by multiple threads without using the lock statement to serialize access. Using the volatile modifier ensures that one thread retrieves the most up-to-date value written by another thread.” My limited understanding of threading is a known issue—I was actually relieved to find out that this question was related to threading.

Here are some parlor-trick questions that are not really worth my time (but they are apparently worth serious money):

  • How would you swap the contents of two variables? I am unable to think of any real-world situation where I would need to do this. I actually got angry when I was asked this question—which prevented me from answering the question properly.
  • Given a LinkedList, with a head and a tail, how would you make the head a tail and the tail a head? I assume that the short answer here is to use the Reverse() method (see MSDN). When you can’t use Reverse(), then you can (ha, ha) swap the contents of the Next and Previous properties of LinkedListNode<T> in a state machine with null checks—null implying a head/tail. When this question was presented to me, I assumed that the head was at the top of a hierarchy (like a parent node) and the tail was the last child in this hierarchy—this caused me to shut down and short circuit at the whiteboard in more inappropriate fits of anger. (This question, by the way, was not asked of me within the last few weeks but about four months ago—before I took my last job.)
  • What is the difference between a Dictionary and a HashTable? As far as I’m concerned, a Dictionary supports generics and HashTable does not (it boxes values in object)—that’s the difference. The lack of generics support implies that HashTable is irrelevant to me for most of my work (but this may imply that I’m ignorant of the use of HashTable in distributed cache systems—used in the real world of high-traffic Web sites).

So the personality test that I’m failing is that I get angry when people f around with me—such that my anger prevents me from protecting myself. My anger causes me to go into a form a paralysis that renders me useless. When sadists sense this, they attempt to press my buttons with abandon.

Beyond the anger, some of these parlor trick questions are really tests that look for computer science training over, say, physics training. So here are some ‘pure’ computer science subjects I’m sensing out there:

  • The technical history of collections. (Who invented the LinkedList and why?)
  • Memorializing sort algorithms. (What is a bubble sort?)
  • Bitwise operations. (What is the [Flags] attribute used for?)

I do admit that these subjects are of little interest to me and I maintain that it is unfair to refuse to hire a person based almost entirely on whether they are knowledgeable in these areas. To me, you don’t hire a person that is not aware of, say, how MVC3 works intimately (which I have been guilty of). I am also guilty of being afraid of confrontation in these interviews (turning my anger against me)—what I should do is ask this question:

I do not understand this question, is there another one?

In the real world there should always be another question. The real world always operates in holistic diversity. There is no perfect, one way in real life. When my tester says no, there are no other questions, I should have the courage to politely conclude the interview, stand up and leave the building.

Related Resources

@elijahmanor Elijah Manor

“Modernizr.js : Feature Detection” by @OdeToCode #webdev

@elijahmanor Elijah Manor

“KnockoutJS Custom Binding for Invoking an Action with Enter Key” by @smichelotti #javascript

@jongalloway Jon Galloway

JSON Formatter & Validator—can take a URL and return validated / formatted JSON –

@elijahmanor Elijah Manor

“Setting up ASP.NET MVC 3 to work with jQuery UI 1.9” #jquery

@elijahmanor Elijah Manor

“Replacing Text Nodes With jQuery” by @bennadel #jquery

@mtaulty Mike Taulty

I missed this “Was .NET all a mistake”—the background is well written and it’s a good (provocative) discussion

@elijahmanor Elijah Manor

“SassAndCoffee 1.1 Released for ASP.NET (MVC & WebForms)” by @xpaulbettsx #aspnetmvc

@pluralsight Pluralsight

Free Webcast—Improving Web Performance with Robert Boedigheimer—starts in 30 minutes! RT

@codinghorror Jeff Atwood

“HTML5 standardizes every terrible practice inflicted on us by browser authors, making them part of the web forever.”…

@pluralsight Pluralsight

Steve Michelotti @smichelotti has just published a new course: ASP.NET MVC Scaffolding

@kazuhito Kazuhito Kidachi

Don’t use IDs in CSS selectors? ❧ (@boblet) –

@jongalloway Jon Galloway

Strange to see IE Conditional Comments go—handy for IE6 specific stylesheets, hopefully no longer as necessary.

@LeaVerou Lea Verou

Microsoft removes conditional comments, behaviors and XML Data Islands from IE10, due to incompatibilities with HTML5…

@JohnBristowe John Bristowe

Awesome scrolling experience is awesome:

@KyleMcClellan Kyle McClellan

Survey: Tell me what you think of the RIA DomainService Wizard #RIAServices

@martinfowler Martin Fowler

I’ve published an article on eradicating non-determinism in tests:

@mattpodwysocki Matthew Podwysocki

Erik Meijer’s CoSQL paper on the duality of SQL and NoSQL got Slashdotted: #cosql

@maryjofoley Mary Jo Foley

Microsoft readies three new Office service packs:

@JesseLiberty Jesse Liberty

Posted: Linq: SelectMany

@mtaulty Mike Taulty

Building with WCF RIA Services?

@davewiner Dave Winer ☮

I’ve never understood what the browsers guys do with RSS.

@amyhoy Amy Hoy

i hate watching smart, capable people defend crap—or refuse to express strong opinions—in the name of “professionalism”

I have tried and failed to include a simple Web Parts sample for ASP.NET Web Forms Basics and Esoterics in my studio. The error is:

Exception information: 
    Exception type: SqlException
    Exception message: Failed to generate a user instance of SQL Server due to a failure in starting the process for the user instance. The connection will be closed.
   at System.Data.SqlClient.SqlInternalConnection.OnError(SqlException exception, Boolean breakConnection)
   at System.Data.SqlClient.TdsParser.ThrowExceptionAndWarning()
   at System.Data.SqlClient.TdsParser.Run(RunBehavior runBehavior, SqlCommand cmdHandler, SqlDataReader dataStream, BulkCopySimpleResultSet bulkCopyHandler, TdsParserStateObject stateObj)

It concerns me that System.Data.SqlClient is in use here instead of System.Data.SqlServerCe.4.0, which is specified in the connection string. This setting comes from installing the NuGet package, System.Web.Providers based on Scott Hanselman’s “Introducing System.Web.Providers—ASP.NET Universal Providers for Session, Membership, Roles and User Profile on SQL Compact and SQL Azure.” I have yet to find an article about using Web Parts with SQL CE.

Related Resources

MVP Chalermvong and his session - ASP.NET Performance Up!

Learning how to work with ASP.NET Web Forms is the other thing of “New Things Learned in the Labor Camp”… I’ve avoided confronting ASP.NET “classic” directly for years. Instead, I made the right decision to learn things about Web programming and Web design that Microsoft notoriously undervalued in its professional communication for years until the rise of Scott Guthrie from ASP.NET “classic” to ASP.NET MVC.

Sadly, there is a job market full of IT shops that are deeply invested (often emotionally invested) in ASP.NET “1.9”—the way of seeing the Web through the Microsoft looking glass just before ASP.NET 2.0 was released in 2005.

For years I’ve “run away” from confronting ASP.NET Web Forms directly. The following interview questions I’ve slapped together “finally” engage—directly. By the way to show that I’m super serious about this matter, I’ve literally spent hours over about two weeks building samples for ASP.NET for Web Forms in my studio.

What are ASP.NET Web Forms?

ASP.NET Web Forms is a technology that breaks down a plain-old HTTP Request into a series of clearly defined events (classified broadly under initialization, load and render)—these events describe the “life cycle” of a familiar model, the data-entry form in a Web page. ASP.NET is based on the event-driven programming paradigm made popular and traditional by Microsoft in the 1990s.

What is a ‘Post Back’?

The ‘Post Back’ is the concept developed by Microsoft that stands between the post HTTP verb and the first event recognized by an ASP.NET Page. Typically, the Page.IsPostBack Boolean is checked in the Page.OnLoad event handler. When Page.IsPostBack == false then the get HTTP verb is the request. Microsoft has defined a conventional Jscript function, __doPostBack(), that invokes the ‘Post Back.’ ASP.NET Web Forms seem crude and extravagant in the 21st century when a page-level post action is regarded as expensive and disruptive (ASP.NET AJAX was supposed to remedy this).

Why do ASP.NET controls require a parent form?

Without a parent form element, the ‘Post Back’ concept is not implemented. Conventional hidden fields in the form store event metadata (__EVENTTARGET, __EVENTARGUMENT, __VIEWSTATE).

See also:TRULY Understanding ViewState

What is the AutoPostBack property?

The AutoPostBack property is a Boolean that is apparently defined independently and conventionally in the System.Web.UI.WebControls.ListControl, System.Web.UI.WebControls.CheckBox and System.Web.UI.WebControls.TextBox. When AutoPostBack == true an event associated with these controls (select, check, blur) will cause a Post Back.

See also:
System.Web.UI.PostBackOptions and System.Web.UI.Page.AutoPostBackControl.

super smart msft peeps

What is a Cross-Page Post Back?

By default, an ASP.NET page “posts back” to itself (frequently). A Cross-Page Post Back declares that one Web Form should post to another Web Form—not back to itself. Input captured in the posting page can be captured in the receiving page via the Page.PreviousPage property.

Tip: The Page.IsCrossPagePostBack property is always null. During a cross page Post Back Page.PreviousPage.IsCrossPagePostBack will be true.

See also:

What is a Server Transfer?

An ASP.NET Server Transfer allows a Page executing on the Server to Transfer execution to another Page. It can be thought of as a call to another procedure from the currently executing procedure. In terms of a POCO class, a Server Transfer is a method call in a procedure of one class with a member from another class—the catch is that no arguments can be passed in the method call. After a Server Transfer, the new page running has access to all of the Request, Session and Application-scope data in the old page. Importantly, there is a Page.PreviousPage property which can be used with FindControl() to locate controls on the old page.

According to Microsoft MVP Karl Moore, “So, in brief: Response.Redirect simply tells the browser to visit another page. Server.Transfer helps reduce server requests, keeps the URL the same and, with a little bug-bashing, allows you to transfer the query string and form variables. …Don’t confuse Server.Transfer with Server.Execute, which executes the page and returns the results. It was useful in the past, but, with ASP.NET, it’s been replaced with fresher methods of development. Ignore it.”

See also:ASP.NET Server.Transfer Method

What is a Master Page?

The ASP.NET Master helps keep Web Forms DRY. They share visual elements among several ASP.NET pages providing the “consistent look and feel” Web designers expect. ASP.NET Master pages also centralize HTML page “meta” information, much of the angle-bracketed ceremony that make Web page markup noisy.

How does a Master Page affect the Control.FindControl function?

The ASP.NET Master adds one or more ContentPlaceHolder controls to the visual tree. These in turn demand Content controls, which implement INamingContainer—a “naming container” affects the ID attributes of child controls “inside” the container. So: a control named foo inside a naming container, bar, can be located by Control.FindControl("bar$foo") where Control would be an instance of the Master page in our example.

The ASP.NET Dream Team

How do Web Forms handle Validation?

ASP.NET handles Validation with PostBack-dependent server controls. These server controls share a ControlToValidate property, pointing to the input-related server control to Validate. There are validation controls for marking input as required, checking input for a data type, testing input with a Regular Expression and testing input with custom code.

One little annoyance: custom validators don’t work with blank input—so they must be paired with the validation control for marking input as required (RequiredFieldValidator).

What are Validation groups?

Validation groups allow more than one “validation context” to exist in a Web Form. For example, a layout with a GridView might have an EditItemTemplate with one Validation Group for editing while another form used for inserting new rows into the grid will have another Validation Group. Validation groups are declared with the ValidationGroup property.

What is a “Partial Page Update”?

The “Partial Page Update” is made possible with the UpdatePanel and the ScriptManager, which is the heart of ASP.NET AJAX. This extension to ASP.NET introduced in version 2.0 dramatically reduces the need to do the “classic” PostBack by implementing the “asynchronous PostBack” a.k.a. the “Partial Page Update.”

K Scott Allen

What I notice about ASP.NET HttpContext:

  • The ApplicationInstance property has some of the events that have seemed to me to appear from nowhere in Global.asax.cs—by the way, Application_Start is not there—what is there are a bunch of authentication events (leading me quite ‘naturally’ to the subject of authentication).
  • For the Items collection, the MSDN documentation explains that the collection “…can be used to organize and share data between an IHttpModule interface and an IHttpHandler interface during an HTTP request…” My opinion: this collection is meant to be used for Module programming—not for passing object-boxed data between pages (especially when these data can be accessed with the Page.PreviousPage property).
  • The Profile property sitting in HTTP Context helps me distinguish between Profiles and Membership—it’s the Membership processes that result in User Profile data.
  • The Cache property refers to the second type of out-of-the-box ASP.NET caching, application data caching (output caching, by the way, is the first). Based on my current level of experience with the innards of ASP.NET I assume that the Cache property is still useful when more advanced distributed caching solutions, like Velocity, are in use.

Related Links