I believe you should keep the amount of channels abstracted:
The amount of channels may vary in the future, e.g. hardware up/downgrade (new devices)
You may want to reserve some channels in the future, e.g. for out-of-band communications and notifications
The exactness should be kept and ensured in initialization.
You should expose the channels with the least specific type possible, such as IEnumerable<ReceivingChannel>
and IEnumerable<TransmittingChannel>
.
The MainDevice
should expose a way to initialize or manage these collections, you have several non-exclusive options:
private List<IReceivingChannel> receivingChannels = new List<IReceivingChannel>();
private List<IReceivingChannel> transmittingChannels = new List<ITransmittingChannel>();
public MainDevice(IEnumerable<IReceivingChannel> receivingChannels,
IEnumerable<ITransmittingChannel> transmittingChannels)
{
this.receivingChannels.AddRange(receivingChannels);
this.transmittingChannels.AddRange(transmittingChannels);
}
- Fluid initialization methods
public MainDevice WithReceivingChannels(IEnumerable<IReceivingChannel> receivingChannels)
{
this.receivingChannels.AddRange(receivingChannels);
return this;
}
public MainDevice WithReceivingChannels(params IReceivingChannel[] receivingChannels)
{
return WithReceivingChannel((IEnumerable<IReceivingChannel>)receivingChannels);
}
public MainDevice WithTransmittingChannels(IEnumerable<ITransmittingChannel> transmittingChannels)
{
this.transmittingChannels.AddRange(transmittingChannels);
return this;
}
public MainDevice WithTransmittingChannels(params ITransmittingChannel[] transmittingChannels)
{
return WithTransmittingChannels((IEnumerable<ITransmittingChannel>)transmittingChannels);
}
Perhaps throw an exception (e.g. InvalidOperationException
) in case you use these methods after the object has been first used by actual worker methods. Or better yet, have these methods in a builder class, where you finally call GetDevice
which will create and initialize the device, possibly with dependency injection (I will not discuss DI here.)
Methods
AddReceivingChannel
, RemoveReceivingChannel
, AddTransmittingChannel
, RemoveTransmittingChannel
GetReceivingChannels
returning IEnumerable<IReceivingChannel>
, or a similar ReceivingChannels
property
GetTransmittingChannels
returning IEnumerable<ITransmittingChannel>
, or a similar TransmittingChannels
property
GetReceivingChannelById
, GetReceivingChannelByAddress
, GetTransmittingChannelById
, GetTransmittingChannelByAddress
Don't use settable properties, as that is asking for trouble. The MainDevice
should be the only entity responsible for managing its actual composition.
This way, you may start by using arrays or lists to store the channels, but switch to dictionaries or trees (e.g. SortedDictionary<TKey, TValue>
) if accessing them by some key becomes a bottleneck. And the users of MainDevice
are just as happy.
You can define a base interface IChannel
for the common things that all channels do, such as opening and/or closing with an optional timeout, and have IReceivingChannel
and ITransmittingChannel
hold the specific operations, such as receiving and transmitting with an optional timeout:
public interface IChannel
{
void Open(TimeSpan timeout);
void Close(TimeSpan timeout);
}
// I'm using generics, but you don't have to if you'll always
// deal with the same data type, e.g. IEnumerable<byte>
//
// I'm also over-simplifying by not stating end-of-transmission
// and other usual channel state.
public interface IReceivingChannel<out T> : IChannel
{
T Receive(TimeSpan timeout);
}
public interface ITransmittingChannel<in T> : IChannel
{
void Transmit(T obj, TimeSpan timeout);
}
In case you need to access some inner object from a channel, define specific interfaces on which you may use the as
operator instead of the actual class. For instance, a IFrobberContainer
interface with a GetNativeFrobber
method (or similar NativeFrobber
property), which does not extend IReceivingChannel
.
This way, any object can contain a frobber:
public interface IFrobberContainer
{
IFrobber NativeFrobber { get; }
}
internal class FrobberReceivingChannel : IReceivingChannel<Thing>, IFrobberContainer
{
public void Open(TimeSpan timeout) { /* ... */ }
public void Close(TimeSpan timeout) { /* ... */ }
public Thing Receive(TimeSpan timeout) { /* ... */ }
public IFrobber NativeFrobber { get; private set; }
}
internal class FrobberTransmittingChannel : ITransmittingChannel<Thing>, IFrobberContainer
{
public void Open(TimeSpan timeout) { /* ... */ }
public void Close(TimeSpan timeout) { /* ... */ }
public void Transmit(Thing obj, TimeSpan timeout) { /* ... */ }
public IFrobber NativeFrobber { get; private set; }
}
In case you have something extra to do with frobber channels, you can ask if the channel has a frobber using the as
operator.
With the new C# 6 null-conditional operator, you can do this:
(channel as IFrobberContainer)?.NativeFrobber?.Frobulate();
instead of this:
var frobberContainer = channel as IFrobberContainer;
if (frobberContainer != null)
{
var frobber = frobberContainer.NativeFrobber;
if (frobber != null)
{
frobber.Frobulate();
}
}
By segregating these distinct aspects into interfaces, you make your actual objects more flexible, more easily replaceable and mockable (good for unit testing).
If you insist on having 10 receiving channels and 4 transmitting channels, then you probably shouldn't use collections, but conversely use very good names for 14 properties. For generic iteration, you can keep internal collections initialized once, to avoid obvious code repetition.
You should still use interfaces and document which ones each channel implements, instead of using actual classes as the types for these properties, to retain the replaceability benefits for refactoring, maintenance and testing purposes.
Best Answer
Basically the problem is blocking on a call that will never return. This is a solved problem in that every network call is highly likely to fail so we have things like sockets that have timeouts built-in to them. In one sense, the problem is that your code assumes this won't fail to respond when it can. The only way to get around that is to stop blocking on that call. You'll need some sort of asynchronous capability. Whether you use some sort of non-blocking IO or spawn a thread, the solution will have the same kind of high-level design.
The solution depends a lot on the overall architecture. If you are structured around a work-queue, then the nicest solution is when the API returns (or times out) you push the response to the work queue and deal with it. On the other hand, if your application is highly sequential, you probably have to block the main thread until the response comes back (or times out.)
In either approach, one issue is how long the API can be expected to take. If the response time is highly regular (low variance) then you can simply accept the timeout as a failure. If you can't put a reasonable upper-limit on the response time, you might want to take one more step on timeout to send another request to the API (preferably something fast) to check whether you are still connected. If you are still able to connect, the cord isn't cut and you can wait some more and repeat as necessary.