|
Perhaps you should have told us what the requirements were.
What would he say to a system of plug-ins?
|
|
|
|
|
Where does that "unknown" come from? Unknown at compile time? (that would eliminate switch as option..)
|
|
|
|
|
It relates to an "unknown" GUI requirement as we have not been told what options we will provide the user with so options 1 and 2 are there for show.
|
|
|
|
|
Well if it's just that you could just go back and add some cases.. right?
|
|
|
|
|
Ok thanks for all the comments, it appears that a switch is all that is needed
|
|
|
|
|
If you need that logic in multiple places, there might be some value in putting it in a dictionary structure (e.g., if you need a combination of the values and the logic for selecting which keys depends on where the dictionary is used). That way, you can pass the dictionary around. Then again, you could just wrap a switch statement in a delegate and pass that around (may not work well if there is custom logic to access several keys). In this specific case, I see no reason a switch statement should not be used. It is certainly simpler and I recommend choosing the simpler approach unless there is a real chance you will need the more complex approach later.
|
|
|
|
|
You might want to move some of this code out of your main method and into other methods.
For e.g.
thebuzzwright wrote: methodList.Add("1", (() =>
_person.ForEach(s => Console.WriteLine("{0} {1}", s.Name, s.Title))));
might go into a readInput() method.
thebuzzwright wrote: person.Add(new Person( "James","Prog" )); _person.Add(new Person( "Claire","Illu"));
could go into loadData() or something like that.
thebuzzwright wrote: string input = Console.ReadLine(); if (methodList.ContainsKey(input)) methodList[input]();
could go into a method called userResponse().
If any of these methods are modified later, these logical groups might make it easier for others to understand what is done where in the program.
modified on Saturday, May 1, 2010 12:39 AM
|
|
|
|
|
I see the point your Lecturer is trying to convey. You are at a learning stage. Hence there will always be possibilities that the code you write will need to be improved/modified/tweaked.
If this is the final code piece and there cannot be any more changes to be made, this code is fine.
However if there is a possibility that someone else may be using your code (as in the case of Open Source/ or in a Production Environment), this code may cause misunderstandings.
Consider this, what happens if you join a company and are given this piece of code and are asked to make some changes. How comfortable would you be? (You understand this code. But any other such code example?)
|
|
|
|
|
As a switch zealot (only kidding), I don't see anything wrong with you using the dictionary - as long as you were doing it for one particular case (you aren't in this case so a switch would be a better choice). If your application was using Dependency Injection with plugins then it could make sense to use the dictionary because you may have more and more plugins so your code would constantly need to be modified.
In this case, it looks like the lecturer is complaining about the lambda. I like Expressions in their place but there comes a time when you need to step back and ask yourself if your code is clean and easy to read.
"WPF has many lovers. It's a veritable porn star!" - Josh Smith As Braveheart once said, "You can take our freedom but you'll never take our Hobnobs!" - Martin Hughes.
My blog | My articles | MoXAML PowerToys | Onyx
|
|
|
|
|
Thanks for the replies its quite a mixed view from what im getting so far, from what I understand I should use a switch but if it were to be like a component "plug in and play" style I would use the approach I am using so far.
Thanks again !
|
|
|
|
|
There are two concerns really, and perhaps your lecturer didn't clearly state which he's concerned about (or both?):
0) Switch vs. Dictionary of delegates
1) Regular Method vs. Anonymous Method vs. Anonymous Method with lambda
As written, I would go with the switch, as that would also eliminate both concerns.
However, when you consider growth, the Dictionary may be the way to go. But before you get there you should make your code more modular (as suggested by Abhinav S).
In a command-line-oriented application I'm working on I have an OperationMap class that is a Dictionary of Dictionaries of delegates. The delegates all refer to "regular" methods in the API.
Here's most of the Run method of my command line interpreter:
Command cmd = null ;
do
{
try
{
cmd = GetCommand ( interactive ) ;
OperationMap [ cmd.Operation ] [ cmd.Entity ]
(
this.api
,
cmd
) ;
}
catch ( System.Exception err )
{
System.Console.WriteLine ( err ) ;
}
}
while ( cmd.Operation != Operation.Exit ) ;
Notice how things like getting the command from the user is done by another method rather than muddying up this method. Modularity, separation of responsibility, encapsulation, etc. are things you should master before trying to use anonymous methods and lambdas.
Also, start thinking of your applications in layers. Generally, your interface should call methods of your API, your API should call methods in your business layer, your business layer should call methods in your data access layer, etc.
In the "real world", rather than using anonymous methods and lambdas, your Dictionary will more likely contain references to methods in the API.
As an excercise, write a class that implements a collection of your Person class and offers a DumpAllToConsole() method, then add a delegate to DumpAllToConsole to your Dictionary.
Oh, and the Person class should probably have an override of ToString that formats itself properly, and maybe a FullName property.
|
|
|
|
|
That is an awesome answer thank you for that !
I was not completely sure of what the exercise meant but I gave it a try:
class Person
{
private string _forName;
private string _surName;
private string _title;
public Person(string forname, string surname, string title)
{
ForName = forname;
SurName = surname;
Title = title;
}
public string ForName
{
get { return _forName; }
set { _forName = value; }
}
public string SurName
{
get { return _surName; }
set { _surName = value; }
}
public string Title
{
get { return _title; }
set { _title = value; }
}
public override string ToString()
{
string fullName = string.Format("{0} {1} {2}", ForName, SurName,Title);
return fullName; ;
}
}
class PersonManager
{
private List<Person> _personList;
public PersonManager()
{
_personList = new List<Person>(10);
_personList.Add(new Person("James", "x", "Prog"));
_personList.Add(new Person("Claire", "y", "Illu"));
}
public void DumpAllToConsole()
{
foreach (Person p in _personList)
{
Console.WriteLine(p.ToString());
}
}
}
class Program
{
delegate void DictionaryDel();
static DictionaryDel _invokeAction;
static Dictionary<PersonManager, DictionaryDel> _action;
static PersonManager _personManager;
static void Main()
{
_action = new Dictionary<PersonManager, DictionaryDel>();
_personManager = new PersonManager();
_invokeAction += _personManager.DumpAllToConsole;
_action.Add(_personManager, _invokeAction);
_action[_personManager]();
Console.Read();
}
}
?
|
|
|
|
|
Hmmm... just about.
I'm not much for needless local variables, so I wouldn't use the fullName in the ToString and the _invokeAction in Main.
I also thought you'd still want to do the menu and have the same Dictionary as before, but use the _personManager.DumpAllToConsole rather than the anonymous method.
Also usually += is used with events, not "regular" delegates.
(And it still looks like your generic parameters got eaten by the forum -- check the Encode "<" (and other HTML) characters when pasting option.)
|
|
|
|
|
Thanks for the reply I left the Person class as it is but changed Program:
class Program
{
delegate void DictionaryDel();
static Dictionary<string, DictionaryDel> _action;
static PersonManager _personManager;
static void Main()
{
_personManager = new PersonManager();
LoadActions();
_action["1"]();
Console.Read();
}
static void LoadActions()
{
_action = new Dictionary<string, DictionaryDel>();
_action.Add("1", _personManager.DumpAllToConsole);
}
}
modified on Sunday, May 2, 2010 8:01 AM
|
|
|
|
|
Good call on writing LoadActions -- now can you make a class to encapsulate the menu options?
Menu menu = new Menu ( _personManager ) ;
|
|
|
|
|
This is the code that I have created so far, your comments are very helpful!
class Program
{
static void Main()
{
PersonManager personManager = new PersonManager();
Menu mainMenu = new Menu(personManager);
mainMenu.LoadActions();
mainMenu.Display();
Console.Read();
}
}
class Person
{
public Person(string forname, string surname, string title)
{
ForName = forname;
SurName = surname;
Title = title;
}
public string ForName { get; set; }
public string SurName { get; set; }
public string Title { get; set; }
public override string ToString()
{
string fullName = string.Format("Name :{0} {1}\nTitle {2}", ForName, SurName, Title);
return fullName;
}
}
class PersonManager
{
private readonly List<Person> _personList;
public PersonManager()
{
_personList = new List<Person>(10)
{
new Person("James", "x", "Prog"),
new Person("Claire", "y", "Illu")
};
}
public void DumpAllToConsole()
{
foreach (Person p in _personList)
{
Console.WriteLine(p.ToString());
}
}
}
class Menu
{
private readonly PersonManager _personManager;
private delegate void DictionaryDel();
private Dictionary<string, DictionaryDel> _action;
public Menu(PersonManager personManager)
{
_personManager = personManager;
}
void GetInput(string input)
{
if (_action.Count == 0) return;
if (!_action.ContainsKey(input)) return;
_action[input]();
}
public void LoadActions()
{
_action = new Dictionary<string, DictionaryDel>
{
{"1", _personManager.DumpAllToConsole}
};
}
public void Display()
{
Console.WriteLine("Press 1 to display all students or 2 to quit");
string input = Console.ReadLine();
GetInput(input);
}
}
Is there anything else I should think about? Thanks !
...For some reason turning "Encode "<" (and other HTML) characters when pasting" on or off does not allow me to show the dictionary generic structure
modified on Sunday, May 2, 2010 12:46 PM
|
|
|
|
|
(Especially for the automatic properties (or whatever they're called).)
Not much else:
Maybe GetInput should be named DoAction instead?
I would likely call the LoadActions from the constructor.
I would have a while loop in Display.
You may also consider creating an interface from the PersonManager so the Menu can be used with other types of manager.
Come to think of it, maybe the Manager should be generic -- Manager<Person>.
|
|
|
|
|
Thanks again, I have implemented your advise and I can see that it is a much more manageable way that what it originally was, even though there is allot more work involved it does pay off!
class Program
{
static void Main()
{
Manager<Person> personManager = new Manager<Person>();
personManager.Add(new Person("James","x","Soft"));
personManager.Add(new Person("Claire", "y", "Illu"));
Menu mainMenu = new Menu(personManager);
mainMenu.Display();
}
}
class Manager<T>
{
private readonly List<T> _managedList;
public Manager()
{
_managedList = new List<T>(10);
}
public void DumpAllToConsole()
{
foreach (T p in _managedList)
{
Console.WriteLine(p + "\n");
}
}
public void Add(T input)
{
if (input.Equals(null)) return;
_managedList.Add(input);
}
public void Remove(T item)
{
if (!_managedList.Contains(item)) return;
_managedList.Remove(item);
}
public List<T> GetCopyOfList()
{
List<T> list = new List<T>(_managedList);
return list;
}
}
class Menu
{
private readonly Manager<Person> _personManager;
private delegate void DictionaryDel();
private Dictionary<string, DictionaryDel> _action;
private bool _running = true;
public Menu(Manager<Person> personManager)
{
_personManager = personManager;
LoadActions();
}
void DoAction(string input)
{
if (_action.Count == 0) return;
if (!_action.ContainsKey(input)) return;
_action[input]();
}
void LoadActions()
{
_action = new Dictionary<string, DictionaryDel>
{
{"1", _personManager.DumpAllToConsole},
{"2",(() => _running = false)}
};
}
public void Display()
{
while (_running)
{
Console.WriteLine("Press 1 to display all students or 2 to quit");
string input = Console.ReadLine();
DoAction(input);
}
}
}
Thanks again !
Also on a side note, I have just finished reading "The pragmatic programmer, from journeyman to master" and it talks about "orthogonality" which I am positive is the end result from the help I have received !
|
|
|
|
|
It's turning out better than I thought.
In Add, I would use a simple if ( input == null ) (this isn't Java).
Yeah, the lambda seems good there.
GetCopyOfList is a good idea, but look into list.AsReadOnly() to see if it might be what you want there.
I would still argue for an Interface that the Menu class requires -- that way it can (theoretically) be used with different types of Manager.
So far I see no need to make Menu be generic, but you may consider it.
thebuzzwright wrote: "orthogonality"
That's one of the (few) things I remember from college.
thebuzzwright wrote: even though there is allot more work involved it does pay off
It certainly can. It may not in this case, but having experience in how to do this sort of abstraction should pay off when it's really needed later.
|
|
|
|
|
Thanks again "PIEBALDconsult" you have helped me allot !
I don't feel the need to show the code as its just minor changes but:
1 - I added a IMenu interface which can "Display" so that I can derive new menu's for other situations.
2 - Menu inherits IMenu and I replaced " Menu mainMenu = new Menu(personManager);" with " IMenu mainMenu = new Menu(personManager);" so that I can use different menu's.
3 - I decided to switch the List<t> with an IList<t> so that I can use the "AsReadOnly()" method as I don't need to modify the contents.
4 - For the "Add" method: "if (input != null) ManagedList.Add(input);" looked better for me.
Final thoughts: I could apply the strategy pattern to IMenu so that it encapsulates its behaviour if needed, but this project was to show manageable code, and I think that would make it more complicated and unnecessary
Thanks again!
|
|
|
|
|
|
Just out of curiosity - Where do you attend and what level of student are you currently?
"I need build Skynet. Plz send code"
|
|
|
|
|
Alaric_ wrote: Where do you attend and what level of student are you currently?
I am a college (UK) student on a final year of a BTEC National Diploma in Computing. University this year to study computing hopefully
Why the curiousness ?
|
|
|
|
|
I have some code that creates an image, and renders some text to it. In a winform app, this results in a reasonably attractive image. However, in Asp.Net, the very same code results in a very ugly image (the stroke width is uneven, and the anti-aliasing is just barely noticable). Can anyone explain why this happens, and even better, a provide solution?
If you want to try it yourself, here's the code:
Bitmap bmp = new Bitmap(60, 30, PixelFormat.Format32bppArgb);
Graphics g = Graphics.FromImage(bmp);
g.Clear(Color.Transparent);
g.SmoothingMode = System.Drawing.Drawing2D.SmoothingMode.Default;
g.CompositingMode = CompositingMode.SourceOver;
g.TextRenderingHint = System.Drawing.Text.TextRenderingHint.SystemDefault;
Font font = new Font("MS San Serif", 17, FontStyle.Regular, GraphicsUnit.Pixel);
Size textSz = TextRenderer.MeasureText(title, font);
int x = (int)((bmp.Width - textSz.Width) * 0.5);
int y = (int)((bmp.Height - textSz.Height) * 0.5);
TextRenderer.DrawText(g, title, font, new Point(x, y), Color.Black );
bmp.RotateFlip(RotateFlipType.Rotate270FlipNone);
MemoryStream ms = new MemoryStream();
bmp.Save(ms,System.Drawing.Imaging.ImageFormat.Png);
Byte[] buffer = new Byte[ms.Length];
ms.Position = 0;
ms.Read(buffer, 0, (int)(ms.Length));
Response.ContentType = "image/png";
Response.BinaryWrite(buffer);
Response.Write(title);
bmp.Dispose();
g.Dispose();
On the first line of code above, I had to fully qualify the last parameter in the WinForms app
(System.Drawing.Imaging.PixelFormat.Format32bppArgb )
.45 ACP - because shooting twice is just silly ----- "Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997 ----- "The staggering layers of obscenity in your statement make it a work of art on so many levels." - J. Jystad, 2001
|
|
|
|
|
Hi John, I think I have a partial answer for you: within a WinForm app, when you create a bitmap and then a Graphics from that, you get a Graphics that is compatible with your screen, so e.g. you could use MeasureString and get (pretty) accurate predictions for when you are going to draw some text in the Form.Paint handler.
AFAIK it does so by guessing which monitor you are on (when you have more than one that is). I never did any such thing in ASP.NET, I guess it does not have a (relevant) monitor around for you, so its properties could be way off; you might have a look at DpiX and DpiY (they are getters only!).
I don't know the solution right now; a little Google action pointed me here[^], maybe that helps.
|
|
|
|
|