Acceptability of C# Partial Classes
-
Clifford Nelson wrote:
It seems to make working on the code easier since functionality is separated by files,
I do the same thing if a class has a large number of methods, but this is quite rare. In most cases, my top level facade class will expose a number (e.g. half a dozen) of services, each of which offer specific functionality. This allows me to swap/evolve service provider implementations over time. This pattern has proved useful when maintaining this[^] app over several versions. /ravi
My new year resolution: 2048 x 1536 Home | Articles | My .NET bits | Freeware ravib(at)ravib(dot)com
Hi Ravi-ji, Assuming you have not published in depth about this strategy for app structure, I'd really appreciate seeing an article by you on this. cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
-
I have heard people say that only use partial classes when there is auto-generated code. I have found that I like to use them to separate different parts of the code that are pretty much independent, but not totally. It seems to make working on the code easier since functionality is separated by files, and this reduces the length of each file, which makes it easy to find what I am looking. If I created separate classes for the same thing, I would need to pass information to allow them to change access elements. I have also used, but did not implement, a case in an adapter where all the calls to get database data were through the same class, and each table had its own file which was a partial class. I thought this worked well, and kept code conflicts down. I just followed the pattern.
I use them extensively. Particularly for library routines. .net 1 should not have been released without them.
-
I have heard people say that only use partial classes when there is auto-generated code. I have found that I like to use them to separate different parts of the code that are pretty much independent, but not totally. It seems to make working on the code easier since functionality is separated by files, and this reduces the length of each file, which makes it easy to find what I am looking. If I created separate classes for the same thing, I would need to pass information to allow them to change access elements. I have also used, but did not implement, a case in an adapter where all the calls to get database data were through the same class, and each table had its own file which was a partial class. I thought this worked well, and kept code conflicts down. I just followed the pattern.
Never having used one I now see an immediate use for them. I have a dashboard with multiple tabs, all the data is related but I got annoyed by the size of the code base and split the 3rd tab out to a UC. Now I see I should have create a partial class.
Never underestimate the power of human stupidity RAH
-
But it's true. :^) Maintaining a fairly large app (with a large active user base) over 12+ years is no easy feat. I need the benefit of a carefully thought out design to help me do that. I was just sharing my experience. /ravi
My new year resolution: 2048 x 1536 Home | Articles | My .NET bits | Freeware ravib(at)ravib(dot)com
I was just teasing you, Ravi. I agree with your comments on the topic. :)
-
I have heard people say that only use partial classes when there is auto-generated code. I have found that I like to use them to separate different parts of the code that are pretty much independent, but not totally. It seems to make working on the code easier since functionality is separated by files, and this reduces the length of each file, which makes it easy to find what I am looking. If I created separate classes for the same thing, I would need to pass information to allow them to change access elements. I have also used, but did not implement, a case in an adapter where all the calls to get database data were through the same class, and each table had its own file which was a partial class. I thought this worked well, and kept code conflicts down. I just followed the pattern.
Partial classes is a good thing, but like anything it can be abused. I have inherited an application with thousands of partial classes spread over just a few files partitioned by funtionality. If I ever meet the programmer I'll teach him generics with a drill and a cactus.
Wrong is evil and must be defeated. - Jeff Ello
-
I have heard people say that only use partial classes when there is auto-generated code. I have found that I like to use them to separate different parts of the code that are pretty much independent, but not totally. It seems to make working on the code easier since functionality is separated by files, and this reduces the length of each file, which makes it easy to find what I am looking. If I created separate classes for the same thing, I would need to pass information to allow them to change access elements. I have also used, but did not implement, a case in an adapter where all the calls to get database data were through the same class, and each table had its own file which was a partial class. I thought this worked well, and kept code conflicts down. I just followed the pattern.
For a particular WinForm app, I have a lot of controls on the main page. I use partial classes to break up the handling of the control events into their logical groups. Otherwise, I never use partial classes, and instead use a publisher/subscriber pattern to communicate between instances, which also has the advantages of: 1) letting me hook in logging so I can see what the heck is going on 2) process intercommunication asynchronously, as the pub/sub I use can make message handler call on a separate thread 3) better exception handling, as the pub/sub will wrap the message handler call in a try-catch and log exceptions 4) because the exception handler uses the pub/sub itself to log the exception, I get can wire up other services, like an email notification, when errors occur 5) and the pub/sub instantiates the receiver class, so I'm enforcing completely isolated processing, which is great for thread safety. So effectively, all the things that actually do non-UI things become services, and I often write them as runtime loaded modules that register themselves in the pub/sub, which is cool because I can then easily mock the services, change the business logic by loading a different module that implements different behaviors, extend the behaviors simply by adding new modules that handle the same messages, and so forth. So there, in a nutshell, you have The Clifton Method[^]. Marc
Imperative to Functional Programming Succinctly Contributors Wanted for Higher Order Programming Project! Learning to code with python is like learning to swim with those little arm floaties. It gives you undeserved confidence and will eventually drown you. - DangerBunny
-
Hi Ravi-ji, Assuming you have not published in depth about this strategy for app structure, I'd really appreciate seeing an article by you on this. cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
Dunno if it's worthy of an article, Bill-ji. Maybe a tip, at best? It isn't something I designed - just a pattern I employ. But I'll tipify it if you think it's worth doing. /ravi
My new year resolution: 2048 x 1536 Home | Articles | My .NET bits | Freeware ravib(at)ravib(dot)com
-
Dunno if it's worthy of an article, Bill-ji. Maybe a tip, at best? It isn't something I designed - just a pattern I employ. But I'll tipify it if you think it's worth doing. /ravi
My new year resolution: 2048 x 1536 Home | Articles | My .NET bits | Freeware ravib(at)ravib(dot)com
Sukria, I accept tips :) cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
-
Sukria, I accept tips :) cheers, Bill
«There is a spectrum, from "clearly desirable behaviour," to "possibly dodgy behavior that still makes some sense," to "clearly undesirable behavior." We try to make the latter into warnings or, better, errors. But stuff that is in the middle category you don’t want to restrict unless there is a clear way to work around it.» Eric Lippert, May 14, 2008
OK. But you're giving me more credit for my knowledge of Hindi (and software engineering) than I'm due. :) /ravi
My new year resolution: 2048 x 1536 Home | Articles | My .NET bits | Freeware ravib(at)ravib(dot)com
-
Never having used one I now see an immediate use for them. I have a dashboard with multiple tabs, all the data is related but I got annoyed by the size of the code base and split the 3rd tab out to a UC. Now I see I should have create a partial class.
Never underestimate the power of human stupidity RAH
Usually when I have tabs I split out into separate classes, but partial classes seems like I good option