OO Software design epiphany - it might not matter

[Greg Utas]

My initial reaction is to agree, but the term is nonetheless useful.

Robust Services Core | Software Techniques for Lemmings | Articles
The fox knows many things, but the hedgehog knows one big thing.

[Rage]

charlieg wrote:

Are you from Stack Overflow

Since he did not ask you if you to first go get a CS degree before posting questions, he is probably not.

Do not escape reality : improve reality !

[Rage]

den2k88 wrote:

AutoSAR

:omg: Someone else on CP who knows Autosar !

Do not escape reality : improve reality !

[raddevus]

My theory is that there are very few examples around that bring OOP all together. OOPs main purpose should be reuse and making the code that you write smaller (less code to deal with when adding enhancements or fixing bugs). I believe that with one very focused example you would see all of PIE-A (Polymorphism, Inheritance, Encapsulation and Abstraction) come together. But most of the time you don't need this type of architecture until things get large. And, most projects don't get large -- especially samples you see. Here's My Attempt This should be an article but I'll do that later. The Entire Premise Imagine you want to save data to three different data stores: 1) file 2) database 3) web location Save instantly becomes our main verb(functionality). Requirements: We Want Four Things 1. Any dev must be able to include the Save() functionality on their class in the future. (interface) 2. Any dev must be able to call the Save() functionality on any class in the future and easily know that it is named Save() -- this is self-documenting code 3. There must be an easy way for dev to configure where the data will be stored (file, db, url) 4. A dev must be able to create a list of various types (classes in the architecture) and iterate through them, calling Save() and knowing that they will save to their appropriate destination. This is Polymorphism -- all objects implement the Interface which provides Save(). Here is the smallest sample I can come up with and it really works. Get LINQPad - The .NET Programmer's Playground[^] and run the code below. You will see the following output:

I'm saving into a FILE : super.txt
I'm saving into a FILE : extra.txt
I'm saving into a DATABASE : connection=superdb;integrated security=true
I'm saving into a WEB LOCATION : http://test.com/saveData?
I'm saving into a FILE : super.txt
I'm saving into a FILE : extra.txt
I'm saving into a DATABASE : connection=superdb;integrated security=true
I'm saving into a WEB LOCATION : http://test.com/saveData?

Now a dev can 1. create an IPersistable object. 2. pass in an IConfigurable object (which determines which data store the Save() will write to 3. call Save() on the object Dev Only Needs To Know Two Things: Abstraction 1. create a configurable object -- select which data store 2. call Save()

void Main()
{

[Greg Utas]

I also worked in what you could call the embedded world, though not so close to the hardware, and soft rather than hard real-time. An OO rewrite saved the product I was working on, and it's still seeing development over 20 years later. We used all three (encapsulation, inheritance, and polymorphism) extensively.

Robust Services Core | Software Techniques for Lemmings | Articles
The fox knows many things, but the hedgehog knows one big thing.

[den2k88]

Got thrown into it about a year and a half ago. It was bound to happen, working in a company that has 80% automotive contracts.

GCS d--(d+) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++*      Weapons extension: ma- k++ F+2 X

[den2k88]

Firmware developement follow different rules than software, there's no circling around it. Software should not depend on the underlying implementation details, firmware is the underlying implementation and it's all about details. Some things should be kept as agnostic as possible, e.g. the main state machine should not depend on the exact make of the various hardware components so interfaces to control hardware should be generic (i.e. peripheral_On, peripheral_Off, peripheral_Sleep, peripheral_Send...) but all the rest can not. One component may be turned on/off via the combination of two pins while another, identical on every aspect, may require timed pulses on a single pin and follow a protocol based on several outputs.

GCS d--(d+) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++*      Weapons extension: ma- k++ F+2 X

[den2k88]

The farther you get from the hardware the more generalized approach are useful / a necessity. I worked in a product with similar specifications ("not so close to the hardware, and soft rather than hard real-time.") and OOP was a huge benefit, when we started adopting it there has been a significant improvement in quality, developement time, customization time and stability.

GCS d--(d+) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++*      Weapons extension: ma- k++ F+2 X

[Marc Clifton]

You are quite correct, IMO. First, forget about the promise of re-use when it comes to objects. Nobody ever does anything in the real world twice exactly the same way to merit any re-use benefit. In order of importance, to me: 1. Encapsulation - keeps stuff organized 2. Interfaces - defines what the class is expected to implement. I also use empty interfaces simply to indicate that the class supports some other behavior. Could use attributes for that as well, but interfaces are sometimes more convenient when dealing with a collection of classes that all support the same thing and there are methods that operate on that, hence I can pass in "IAuditable", for example. 3. Inheritance/Abstraction - mostly useless, but there are times when I want to pull out common properties among a set of logical classes. Note that I don't consider this to be true abstraction, it's using inheritance to defined common properties and behaviors. 4. Polymorphism - useful, but less so now with optional default parameters that do the work of what one often used polymorphic methods for. IMO, the reality of "how useful is OO" falls quite short of the promise of OO.

Latest Articles:
Client-Side Type-Based Publisher/Subscriber, Exploring Synchronous, "Event-ed", and Worker Thread Subscriptions

[Nelek]

I was almost thrown in a project where AutoSar was involved. Luckily for me, another project started to burn down and I was sent there to extiguish the fire in the last moment and I could remain in my blessed ignorance.

M.D.V. ;) If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about? Help me to understand what I'm saying, and I'll explain it better to you Rating helpful answers is nice, but saying thanks can be even nicer.

[charlieg]

lol, true. I'm an EE, written lots of code but I've always wanted to take an algorithms class. Don't know why, well, at least my analyst asks me.

Charlie Gilley <italic>Stuck in a dysfunctional matrix from which I must escape... "Where liberty dwells, there is my country." B. Franklin, 1783 “They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759

[User 13269747]

I tend to take the view that isolation is a good target. Code isolated from other code is both maintainable and reusable, regardless of whether it is via an OO design or not. Reuse is limited when using any OO language, because then any program that wants to reuse that code has to use the same language. If you're going after reuse, you have to give up OO because they are mutually exclusive. All the re-used code are written in a non-OO language. Sqlite, for example, is one of the most widely deployed pieces of code in the world. All media format readers, as well, are widely deployed and non-OO. If you write something really new and novel that does not exist (a new image format, a new protocol, new encryption algo, new compression format, interface to any of the above, or interface to existing daemons (RDBMSs, etc)), and you write it in Java or C#, the only way that it can become popular is if someone re-implements it in C so that Python, C, C++, Java, C#, Delphi, Lazarus, Perl, Rust, Go, Lisps, Php, Ruby, Tcl (and more) programs can use it. The upside of producing library files (.so or .dll) that can be used by any language is that the result is also quite isolated and loosely coupled from anything else:

• It can be be easily extended by anyone, but not easily enhanced.
• It can be easily swapped out and replaced with a different implementation without needing the programs using that library to be recompiled, redeployed or changed in any way.
• Because it is a library, it will only be for a single type of task (no one would even think of putting unrelated functionality into a compression library, but I've seen devs happily put in unrelated stuff into a compression class).

Ironically, you can more easily achieve SOLID principles writing plain C libraries (.so or .dll) than you can with actual OO languages, because of the limitations of the call interface in dynamic libraries.

[NelsonGoncalves]

fwiw, the code I am modifying has not changed in 10 years. So, why bother making it general?

Well, if it has not changed in 10 years I would say it is general enough :-D I am in sort of the same place as you. Mostly embedded development, and whenever I tried using OO I mostly failed. Usually because I decide to make a class for something that will only have one object instance.

[giulicard]

NelsonGoncalves wrote:

Usually because I decide to make a class for something that will only have one object instance.

Usually, when I see that a class is instantiated only once, like some sort of singleton, I always think it might be wise to drop the class and put everything in a namespace.

[NelsonGoncalves]

Yes, I agree. But I start with the "everything is an object" mentality, and only latter do I recognize that, for my applications, most of the times that is the wrong way to think. I could spend 1hr thinking on the design, but that would just prevent me from latter spending 2 days fixing bad design decisions ;P

[_WinBase_]

I know where you are coming from but would think *it depends* on what you are working on. some projects and cases may merit it and some not, and although i dont work in embedded would assume that it's less important. however as a 'business app guy' i find it great, however if i really looked at it i probably don't need it as much as i think i do, but it's the way i roll now and i quite like it :) GL

[Rage]

I work in the company that kind of created it, so avoiding it completely is not possible. :-D

Do not escape reality : improve reality !

[Ed Korsberg]

I do embedded firmware full time. We use all of the OO features you have mentioned. For the most part this has been done as a positive effort, ie it was worth the effort. But lately I think we have taken it too far and the code is too difficult to follow and debug and also our performance seems really bad. We are now in the process of going back (yet again) to profile, study and do performance measurements. But I think it is a case of having too much of a good thing. We have overused such features on an embedded design with limited memory and tight timing requirements.

[DumpsterJuice]

My Opinion (Based on experience), Every implementation of Polymorphism, on every team that used it, goes sideways pretty quickly, and becomes quite chore to maintain. OOP principles are generally sound, but I have not even considered using Polymorphism in 20 years. Mainly because its harder to debug. BUG: "The Investor Customer is disturbing the customerChecking balance because they customer is also an investor" - Resulting in the loss of his mortgage. Keep It Simple, keep it moving.

[PhilipOakley]

Yeah. Reality bites. At some point we try to abstract away all the hardware bit stuffing, but it's still there ruining all the nice plans - all it needs is an auto resetting status bit when other bits in the byte are read to ruin everything.