We are working on a bigger project with Ruby on Rails. As you might know, Rails works with MVC pattern and typically enforces the following class types for Domain objects:
Controllers (+ Helpers)
Models
Views (e.g. erb)
Serializers
Mailers
We experienced - like many other teams - that this is not feasible for bigger, long-running projects.
So in our current project, we added some layers, so we typically have
Controllers (+ Helpers)
Policies (access control per controller)
Presenters (a.k.a ViewModel)
Forms (containing all attributes available in an HTML form)
Views
Models
Queries (static ActiveRecord queries)
Serializers
Mailers
Services (Domain Service which extracts business code from controller)
It is much more testable, maintainable and understandable.
However, our boss disagrees with "These are too many classes".
As he is typically a proficient architect (doing consulting), I don't want to simply reject this response. However, his time as active Java developer are gone and he often explains such sentences with his improved "gut feeling".
Of course, there are metrics that count classes - but is there a general approach to classify something like 'Classes per Domain object' or 'Class per Business case', which can be taken for an argument?
I have never heard of a rule or best practice which limits a number of classes per project, domain or whatever. One should create as many classes as he REALLY needs. The key word here is "REALLY" because quite often developers over complicate things and create unnecessary classes, layers or even tiers. Without the details about your project I suggest you do the following:
Ask your boss to elaborate on the "Too many classes" phrase. Ask for his concerns. If he is a good architect and a team member he understands his responsibility of coaching people.
Make sure you clearly explained to him why you needed those classes. Provide examples.
If your boss is stubborn, his best days as a developer are gone and he just wants to show his significance, remind him about his responsibility for a successful project. For example, tell him he is an architect and a person who makes the final decision and you will follow it because he is the one who is responsible for the project success but on the other hand it's your responsibility to be proactive and express any concerns you have to help your boss make the right decision. This should make your boss think again as he doesn't want to fail the project because of his "gut feeling".
Hope it helps!
Related
I have the following folder structure for my application
app
core
features
feature1
domain
entities
entity1
entity2
entity3
entity4
entity5
entity6
data
models
model1
model2
model3
model4
model5
model6
presentation
feature2
domain
entities
entity1
entity2
entity3
entity4
entity5
entity6
data
models
model1
model2
model3
model4
model5
model6
presentation
Model 1 to 6 for both features are exactly the same and more are coming as the application scales. This is becoming hard to maintain. Does clean architecture allow for sharing models and entities across the multiple features? Would that be done through the core folder?
Yes you can and you maybe should not do it.
Yes, because the clean architecture doesn't make a statement about this. As long as the dependency rule is honored you are compliant in terms of the clean architecture.
But you maybe should not do it, because there are other principles and considerations you should take into account.
First, you should ask your self if it would be a violation of the single responsiblility principle. Sometimes things look the same but are not really douplicated code. It's more an accident that they look the same. The question is "Who is the change trigger?". Usually the features change for different reasons and thus the features use cases and entities change for different reasons. If so you should not share the models between them.
Second, from a DDD (domain driven design) perspective the features can be in different bounded contexts. In this case you can have two entities with the same name, but they have different meanings in the differnt context. Therefore, the models have different properties and different methods.
If you decide to share the modles you should take a close look at them in intervalls and scan the properties and methods. Are there properties and methods that are exclusively dedicated to one feature? If so you are maybe mixing concerns.
Finally if you are facing problems like "Uhh, we changed someting here and broke something there", you should rethink if the sharing of some code is a good way to go. Sharing of code always couples code, because all clients depend on the shared code. It's a trade off between duplication (maintenance costs) and dependencies. Principles like the SRP help you to make an educated guess.
Does clean architecture allow for sharing models and entities across the multiple features?
I think DRY principles should be applied whether you use clean architecture or not.
As for the answer:
I think you could abstract your shared models and entities into separate modules or packages. If it's all dart code, I suggest choosing packages. You can place it inside the root project (Monorepo) or separate it to another repository, this way you could achieve modularity by abstracting the all layer of shared dependency (abstract class, interface, clients, or maybe repositories) out of the main application
There is a good video about this topic on Google I/O'19, it is about Android but you can get great insight and applied to general mobile development. I suggest you give it a try
I'm assuming you're following the ResoCoder course. I did too. I used that design in our team for a few weeks before realizing there soon become problems with it (which ResoCoder himself realizes if you check the Github issues and responses for his repo):
Sharing between features doesn't work well.
It's hard to have functionality split between multiple features.
Not well organized with a lot of code.
Some more which I forget off the top of my head (He wrote some article on the failing of it somewhere).
Hence, for my team's app (getting quite complex now), we've adapted this which seems to work:
(ignore the 2 \core directories, that's by mistake. There's meant to be only 1)
Then, inside all of these top-level directories (excluding /core - that's specifically for things like API clients, routers, etc.), there's folders for each feature. Ex: authentication, settings, posting, etc.
Then, here's the important bit, in each of these feature-split directories (ex: /domain, /presentation, etc.) we have a sub-directory called /shared that resembles what each folder needs to look like, except it just contains functionality that's categorized as (example) domain or data. This stuff is then split between all features.
For example, if I have an app that allows users to post content, I'm going to create the post entity (using ResoCoder terminology) inside the /posts feature. Except, UH OH, I need to have it displayed inside the /feed feature as well! This is then a perfect case for /shared inside the general /domain directory.
Let me know if this helps, or if you have any further questions!
This thread introduced two different approaches to a fundamental need of every web-app - email-based interactions with a site's users - the uses the term MailManager. I'm currently suffering the pain of a poorly designed, first gen attempt at generating (and sending, and logging, and searching) things like invoices, notifications, followups and tracking. Stuff that wasn't in the original spec. I understand the need to refine things to use a more generalized controller.
So i'd like to study how this functionality has been implemented in other projects. Is this role generic or well defined enough that perhaps one of the codeplex or git projects (or an MVC-based tutorials) has a fleshed out example?
I have finally decided to hop up on the train of MVC 2.
Now I have been doing a lot of reading lately and following is the architecture which I think will be good enough for most Business Web Applications.
Layered Architecture:-
Model (layer which communicates with Database). EF4
Repository (Layer which communicates with Model and includes all the queries)
Business Layer (Validations, Helper Functions, Calls to repository)
Controllers (Controls the flow of the application and is responsible for providing data to the view from the Business Layer.)
Views (UI)
Now I have decided to create a separate project for each layer (Just to respect the separation of concerns dilemma. Although I know it's not necessary but I think it makes the project look more professional :-)
I am using AutoMetaData t4 template for Validation. I also came across FluentValidation but cant find much on it. Which one should I go with?
Which View Engine to go for?
Razor View Engine Was Love at first sight. But it's still in beta and I think it won't be easy to find examples of it. Am I right?
Spark .. I can't find much on it either and don't want to get stuck somewhere in the middle crying for help when there is no one to listen...:-(
T4 templates auto generate views and I can customize them to generate the views the way I want? Will this be possible with razor and spark or do I have to create them manually?
Is there any way to Auto generate the repositories?
I would really appreciate it if I can see a project based on the architecture above.
Kindly to let me know if it's a good architecture to follow.
I have some confusion on the business layer like is it really necessary?
This is a very broad question. I decided to use Fluent NHibernate's autoconfig feature for a greenfield application, and was quite impressed. A lot of my colleagues use CakePHP, and it needed very little configuration to get it to generate a database schema compatible with the default conventions cake uses, which is great for us.
I highly suggest the book ASP.NET MVC2 in Action. This book does a good job at covering the ecosystem of libraries that are used in making a maintainable ASP.NET MVC application.
As for the choice of view engines, that can depend on your background. I personally prefer my view to look as much like the HTML as possible, so I would choose Spark. On the other hand if you are used to working with ASP.NET classic, the WebForms view engine may get you up and running fastest.
Kindly to let me know if its a good architecture to follow?
It's a fine start - the only thing I would suggest you add is a layer of abstraction between your Business Logic and Data Access (i.e: Dependency Inversion / Injection) - see this: An Introduction to Dependency Inversion.
i know its not necessary but i think it makes the project looks more professional :-)
Ha! Usually you'll find that a lot of "stuff" isn't necessary - right up until the moment when it is, at which point it's usually too late.
Re View Engines: I'm still a newbie to ASP.NET MVC myself and so aren't familiar with the view engines your talking about; if I were you I'd dream up some test scenarios and then try tackling them with each product so you can directly compare them. Often, you need to take things for a test drive to be more comfortable - although this might take time, but it's usually worth it.
Edit:
If i suggest this layer to my PM and give him the above two reasons then i don't think he will accept it
Firstly, PM's are not tech leads (usually); you have responsibility for the design of the solution - not the PM. This isn't uncommon, in my experience most of the time the PM isn't even aware they are encroaching on your turf that isn't theres. It's not that I'm a "political land grabber" but I just tend to think of "separation of concerns" and, well, I'm sure you understand.
As the designer / architect it's up to you to interpret requirements and (taking business priorities into account) come up with solution that provides the best 'platform' going forward.
(Regarding DI) My question is , is it really worth it?
If you put a gun to my head I would say yes, however the real world is a little more complex.
If you answer yes to any of these questions then its more likely using DI would be a good idea:
The system is non-trivial
The expected life of the system is more than (not sure what the right figure is here, there probably isn't one, so I'm going to put a stake in the ground at) 2 years.
The system and/or its requirements are fluid.
Splitting up the work (BL / DAL) into different teams would be advantageous to the project (perhaps you're part of a distributed team).
The system is intended for a market with a diverse technical landscape (e.g: not everyone will want to use MS SQL).
You want to perform quality testing (this would make it easier).
The system is large / complex, so splitting up functionality and putting it into other systems is a possibility.
You want to offer more than one way to store data (say a file based repository for free, and a database driven repository for a fee).
Business drivers / environment are volatile - what if they came to you and said "this is excellent but now we want to offer a cloud-based version, can you put it on Azure?"
Id also like to point out that whilst there's definitely a learning curve involved it's not that huge, and once you're up-to-speed you'll still be at least as fast as you are now; or at worst you;ll take a little longer but you'll be providing much more value (with relatively less effort).
In terms of how much effort is involved...
One-Off Tasks (beyond getting the team up to speed):
Writting a Provider Loader or picking DI Framework. Once you've done this it will be reusable in all your projects.
'New' Common Tasks (assuming you're following the approach taken in the article):
Defining interface (on paper) - this is something you'll be doing right now anyway, except that you might not realise it. Basically it's OO Design, but as it's going to be the formal interface between two or more packages you need to give it some thought (and yes you can still refactor it - but ideally the interface should be "stable" and not change a lot; if it does change it's better to 'add' than to 'remove or change' existing members).
Writting interface code. This is very fast (minutes not hours), as you're not writting any implementation; and when you go to implement you can use tools provided by your IDE to generate code-stubs based on the interface.
Things you do now that you'd do differently:
Instantiating a variable (in your BL classes) to hold the provider, probably via a factory. Writting this shouldn't take long (again, minutes not hours) and it's fairly simple code to copy, paste & refactor where required.
Writing the DAL code: should be the same as before.
Sometimes it is way more easy to learn patterns from code : Sharp Architecture is a concrete implementation of good practices in MVC, using DDD.
What is CSLA Framework and Its use ?
My Opinions From My Experience w/ a 1.7M LOC code base:
CSLA is intended for a distributed application/database environment. This is why the basic business object is and does everything, for example it's own data persistence. An object (and everything remotely associated w/ its state) is intended to be serialized, sent to a different application and/or data server and work.
If the above is not a problem you need to solve, CSLA is overkill, big time. Our development team regrets having committed to CSLA.
Juggling all the CSLA balls in a complex Windowed UI is tough. We have multi-tabbed screens (which may in turn open sub-screens) that, unless you follow the "left to right, top to bottom" flow of data entry, and click save often, ends up putting and/or fetching incomplete data to/from the database; or dropping data altogether that you just entered. Yes, our original coders are at fault, but so is CSLA... It just seems that there are so many moving parts to enable, control, and coordinate CSLA features. It's like having to deal with all the dials & switches of a fighter jet when all you really need is something more like a Cessna 152.
You will write lots of custom code to enable the CSLA features. For example CSLA will never be confused with object relational mapper (ORM) tools like Hibernate and Entity Framework. Our SAVE() methods are non trivial, so are the trivial ones.
Encouraging the use of code generators compounds problems. We used CodeSMith to generate classes from data tables. So we end up with code that has a 1-1 correspondence of table to c# class. So you must write all the code to handle dataStore to your "real" objects.
Data store/ and fetch is very inefficient w/ CSLA. Because of the Behemoth, monolithic BusinessObject-does-all-and-knows-all centric paradigm, objects end up doing a one-object-at-a-time data fetch and instantiate. Collections of composite objects significantly compound the problem. A single "get this object" always results in a cascade of separate data fetches (one or more for each individual object) to instantiate the entire inheritance & composite relationship chains. Its known as the "N+1 query problem." Oh, and fetching data ALWAYS results in a new object being created, even if we're only updating an existing one. No wonder our more complex screens are FUBAR.
It allows you to architect your application with solid object oriented principals and a good seperation of concerns.
Yes and no. Mostly no.
The BusinessObject handles it's own data storing. That is anti separation of concerns.
"It allows you..." well, yeah - so does a blank text editor screen, but does not force or encourage you like the MVC.NET framework does, for example. IMHO, CLSA provides absolutely zero benefit for ensuring that the code you develop with it follows "solid OO principles". In fact coders w/ weak OO skills (the majority, in my experience) will really stand out when using CSLA! Woe betide the maintenance programmer.
CSLA is the poster child for the solid object oriented principle favor composition over inheritance. CLSA code is untestable. Because an inherited framework BusinessObject is, does, and needs everything, all at once and every time, it's not likely that you will be able to get much test coverage. You can't get at the pieces because everything is tightly coupled.The framework is not amenable to dependency injection. It is an iron curtain of code.
Your code will be difficult to debug. Call stacks get very deep and as you get near the center of the sun so to speak, everything turns into reflection - "what *&^# methods just got called???" And you simply get lost. period.
EDIT 7 Mar 2016
What more insight can I add after the original post? Two things, perhaps:
First, It feels like CSLA has some promise. If we knew how to juggle all those moving parts together. But CSLA is so enigmatic that even things we have done right are corrupted over time. IMHO without a very strong team-wide CSLA wherewithal, any implementation is doomed. Without a vibrant "open source" of technical references, training, and community it's hopeless. In almost a decade our CSLA code, in my final analysis, is just compounding technical debt.
Second, here is a recent comment I made, below:
Our complexity often does not seem to fit in the CSLA infrastructure
so we write outside of the framework. This and cheap labor results in
rampant SRP violations and has me hitting brick walls managing dynamic
rule application, for example. Then, CSLA parent/child infrastructure
propagates composite object validation but we don't always want c/p
relationships, so we write more validation and store code. So today
our CSLA implementation is inconsistent & confusing. Refactoring to
more-better CSLA will have profound domino effects. So after that
initial injection CSLA is essentially abandoned.
end Edit
CSLA is business object framework that allows you to easily create business objects on top of a data layer. It allows you to architect your application with solid object oriented principals and a good seperation of concerns.
I would highly recommend you read the CSLA book by Rocky Lhotka called Expert C# 2008 Business Objects. That will not only teach you about the framework but also teach good software architecture principals.
You can grab the book here on Amazon
I suggest reading the What is CSLA? page, and browse through the CSLA .NET FAQ site.
For the latest published information check out the Using CSLA 4 ebook series.
In reply to #radarbob https://stackoverflow.com/a/10922373/261363, hope I won't regret this and start a flame war.
Our team has been developing a couple of LOB applications with CSLA. From my experience on writing green field apps with CSLA and maintaining existing code here are my replies to your points.
The BO is not suppose to do it's own data persistence, you will have a Factory that will handle all data persistance, for example using a ORM to map to Models that are later on saved.
Sorry to hear that, I make sure I study the framework documentation and write at least one toy application before committing to a existing code database. Furthermore you can even download and browse the CSLA code.
You have BO -> Portal -> Factories that should not be very complicated the existing CSLA examples go a long way on explaining what is happening on each level.
CLSA should never be confused with a ORM
As you should, business object are rarely mapped to one table and thus require a bit of work when saving. In the case they are mapped to one table to you use something like AutoMapper to map your BO to your POCO in 1 line.
Look into CSLA Commands, also is nothing stopping you from keeping your BO as small or big as you want as long as you keep in mind that they are not the same as the POCO's that you will persist.
In a project we worked on we where able to easily test BO to ensure that the business logic was correct. Because of the nice separation of concerns we tested our Factories in isolation to make sure that business objects will be persisted accordingly.
At one point I was able to easily persist part of my BO's in MongoDB so the application was running on a hybrid database MSSQL and MongoDB without having to even change one line of code in my business objects, all I had to do was to update the factories to use Mongo instead of the current ORM.
Hope this addresses all your points in a fair manner,
Regards
CSLA: Component-based Scalable Logical Architecture
A paragraph in a nutshell that described CSLA to me from the website was this:
CSLA .NET enables you to create an object-oriented business layer that abstracts and encapsulates your business logic and data. The framework ensures your business objects work seamlessly with all .NET interface technologies, including WinRT XAML, WPF, ASP.NET MVC, ASP.NET Web Forms, WCF, asmx services, Windows Phone 7, Silverlight, Windows Workflow and Windows Forms.
Why you might use it:
Business rule management. Once you learn the business rule system, it provides a way to enforce business logic in a tidy package. If you have an object with child objects that need to report Validation to the parent most level, there is a way to handle that. (see http://www.lhotka.net/weblog/CSLA4BusinessRulesSubsystem.aspx for more information on the rule system)
You have business objects that you need to support N-level undo? A CSLA BusinessBase has a baked in property management system (like dependency properties) for functionality like N-Level Undo (it is actually completely implemented.) (This also ties into the business rule management. You can fire validation when a primary property changes, or you can change a property based on the value of another property.)
Data portal management. This one was an interesting concept. If I need to execute data operations locally, CSLA is configured for this out of the box. I can also stand up a WCF service that references my business object libraries, and use a few lines of configuration to make a WCF endpoint to manage data operations. The WCF service is a part of the CSLA framework. It was neat to see this in action. Other scenarios? Sure! Your business object library doesn't need to change, the DataPortal class determines if it needs to execute remotely or locally, according to your configuration.
CSLA does force you to use a few mechanisms that may not feel natural at first. I think that is somewhat true of any pattern you choose to implement. However, when it comes to the challenges of implementing Service Oriented Architecture, CSLA offers a lot. Yes, you are going to have an architect level developer authoring some of your libraries. However, if you're building an enterprise class application, shouldn't you be doing that already?
CSLA, when architected correctly, is testable. We use the repository pattern to replace out the actual dal with a mock layer and test both by specification (using NUnit/SpecFlow) and in a unit fashion when appropriate.
As far as support, including Rocky himself, there is a community of contributors that ensure things like CSLA.Net for Xamarin become a reality. There are consultancies that know CSLA and use it on a regular basis depending on the scope of work (and no, not just the consultancy for which I work.)
All things considered, CSLA may not be for you. Like others have indicated, read the site and the books (especially Expert C# 2008 Business Objects.) Ask questions, as the CSLA community tends to give quality advice.
CSLA is described in detail here. The new book is a great starting point. As a great compliment to the book I would recommend checking out our CSLA 3.8 templates. Rocky recommends using a Code Generator, and we have the leading set of templates, that will get you up and running in no time.
Thanks
-Blake Niemyjski (Author of the CodeSmith CSLA Templates)
Update: This question was inspired by my larger quest for mapping ontologically the whole software systems architecture enchilada. I've written a blog post about it, and hopefully it will help clarify what I'm after.
Many, many, many frameworks and stacks that's event-driven have too much variation for my little head to get around. Is there somewhere some resources that defines the outline of a reasonable Application Event Model, what events there are, and what triggers are most common?
I've got my own framework with a plugin and event-driven architecture, but I want to open-source it, and as such would like to make it closer to some common ground as not to alienate people.
So to clarify; this is for an application, meaning setting up the environment, the dependencies, the data sources (like databases), and being a MVC framework setting up the model, the view, launching controllers / actions, and in the GUI various stages of the interface (header, content, columns, etc.).
Ideas? Thoughts? Pointers? (And I've made it language and platform neutral at this point)
I read your blog entry, which btw I found an extremely interesting read, but... this question does not seem to reflect the broadness of the issue you are presenting there.
What you are after is very abstract and theoretical. What I mean to say is that if you tie any of those ideas to actual technology you will find yourself 'stuck' with it. This is why many of us are reluctant to use any framework. Especially the 'relabeled' products suddenly claiming to conform to the trend. We choose mainly on the basis of what appears to be needed to reach a predetermined result.
Frameworks (or tools in general) that target the application architecture domain distinguish themselves primarily by the amount of responsibility they are designed to take on. Spring for example only deals with the concept of decoupling and is therefore easily adopted and useable in many situations. The quality of any framework is expressed in terms of how well the designers of such frameworks were able to keep their products within the boundaries of that responsibility. Some front-to-end products will do exactly the opposite, code generators being among the 'worst' of them.
To answer your question at the top of this page, I do not think there is a framework that does what you want at this time and I do not think there is a single model of how applications (should) work. Keep in mind though that the application architecture domain deals with technology more than it does with concepts. In other words: If it works and meets the requirements, then you're pretty much done.
That said, you might find something of value in agent-based systems.
Heh. Most developers pick the major framework they like the tools for and stick with it. That's usually the winning strategy. I sympathize with your desire not to marry a single vendor.
Keep in mind however, that in developing your own framework, you're going to end up tied to a single vendor anyway. :-)
Is there somewhere some resources that defines the outline of a reasonable
Application Event Model, what events there are, and what triggers are most common?
I don't think so.
From what I see, there are two kinds of models out there: those with a real framework with which you can make a working data entry dialog, and abstract meta-meta-models that are optimized for modeling themselves.
Try surveying a few current frameworks that have good documentation online and cross-reference the major terminology in a spreadsheet. It's an interesting exercise.
I'd have a look at Spring for Java, and the XT Framework Spring module (http://springmodules.dev.java.net/docs/reference/0.9/html/xt.html), which apparently supports event-driven architecture, as starting points. Spring has an MVC framework (inc. convention-based routing to controllers), db configuration (for Hibernate, particularly), plus full dependency injection support. There's also a mechanism in Spring for modularising your web apps, called Spring Slices. And it can be integrated with Jersey for building RESTful apps.
(Unfortunately, I tried to provide links to everything, but this place only lets new users post a single link. So you'll have to do some googling :) )