Should I expose a "computed" value as a property or a method?

Question

I have a C# class that represents a content type in a web content management system.

We have a field that allows a web content editor to enter an HTML template for how the object is displayed. It basically uses the handlebars syntax for substituting object property values into the HTML string:

<h1>{{Title}}</h1><p>{{Message}}</p>

From a class design perspective, should I expose the formatted HTML string (with substitution) as a property or method?

Example as property:

public class Example
{
  private string _template;
  public string Title { get; set; }
  public string Message { get; set; }
  public string Html 
  {
    get
    {
      return this.ToHtml();
    }
    protected set { }
  }

  public Example(Content content)
  {
    this.Title = content.GetValue("title") as string;
    this.Message = content.GetValue("message") as string;
    _template = content.GetValue("template") as string;
  }

  private string ToHtml()
  {
    // Perform substitution and return formatted string.
  }  
}

Example as method:

public class Example
{
  private string _template;
  public string Title { get; set; }
  public string Message { get; set; }

  public Example(Content content)
  {
    this.Title = content.GetValue("title") as string;
    this.Message = content.GetValue("message") as string;
    _template = content.GetValue("template") as string;
  }

  public string ToHtml()
  {
    // Perform substitution and return formatted string.
  }  
}

I'm not sure from a design standpoint does it make a difference or are there reasons why one approach is better than the other?

Answer 1

UPDATE: This question was the subject of my blog in May 2014. Thanks for the great question!

---

To add to Robert Harvey's answer: a property should be:

• logically a property of the class, the way that say its color or year or model are the properties of a car.

• not more than, let's say, ten times slower to compute than fetching from a field.

• something you don't mind being computed while debugging. The VS debugger automatically computes properties.

• unable to fail. Getters should always return a value no matter what the state of the object is.

I don't think your proposed Html property hits any of those. Don't make it a property unless it hits all of them.

Answer 2

ToHtml is correctly a method, as you have written it in both cases. Just expose it publicly.

Knerd makes a good point: properties can be serialized. You would never deserialize from the HTML, so it doesn't make much sense to make it a property from that perspective.

Consistent with the way ORM's and repository objects work: fields in a record or tuple are represented with properties, but you retrieve the record (or some form of it) using a method.