Recently I started contributing to VsVim, a Visual Studio plugin that emulates Vim. When he was starting the project, Jared Parsons decided to write the bulk of it in F#. He did this mostly as a chance to learn a new language but also because it's a solid first class alternative to C#. For instance, F#'s features like pattern matching and discriminated unions are a natural fit for state machines like Vim.
This is my first experience with a truly functional language. For those who aren't familiar with F#, it's essentially OCaml.NET (the F# book uses OCaml for it's markup syntax), but also draws roots from Haskell. It's a big mind shift from imperative and pure object oriented languages, but one I'd definitely recommend to any developer who wants to be better.
Since I've been working on VsVim, I've been using F# in my spare time but C# in my regular day job. The longer I use F# the more I want C# to do what F# does. The biggest example is how F# handles nulls.
In C# (and Ruby, Python, and any imperative language) most values can be null, and null is a natural state for a variable to be in. In fact (partly due to SQL), null is used whenever a value is empty or doesn't exist yet. In C# and Java, null is the default value for any member reference, you don't even need to explicitly initialize it. As a result, you often end up with a lot of null pointer exceptions due to sloppy programming. After all, it's kind of hard to remember to check for null every time you use a variable.
In F#, nothing is null (that's not entirely true, but in it's natural state it's true enough). Typically you'll use options instead of null. For instance, if you have a function that fails to find or calculate something you might return null in imperative languages (and the actual value if successful). However, in F# you use an option type and return None on failure and Some value on success.
Here, every time you call find(kittens) you get back an option type. This type isn't a string, so you can't just start using string methods and get a null pointer exception. Instead, you have to extract the string value from the option type before it can be used.
At this point you might be thinking, "why would I want to do that? It looks like a lot of extra code". However, I challenge you to find a crashing bug in VsVim. Every time we have an instance of an invalid state we are forced to deal with it on the spot. Every invalid state is dealt with in a way that makes sense.
If we wrote it in C# it would be incredibly easy to get lazy while working late at night and forget to check for null and cause the plugin to crash. Instead, the only bugs we have are behavior quirks. If we ever have a crashing bug, the chances are the null value originated in C# code from Visual Studio or the .NET Framework and we forgot to check.
Discussion on HN
Wednesday, February 29, 2012
Friday, February 10, 2012
C# Reflection Performance And Ruby
I've always known that reflection method invocations C# are slower than regular invocations, but I've never never known to what extent. So I set out to make an experiment to demonstrate the performance of several ways to invoke a method. Frameworks like NHibernate or the mongoDB driver are known to serialize and deserialize objects. In order to do either of these activities they have to scan the properties of an object and dynamically invoke them to get or set the values. Normally this is done via reflection. However, I want to know if the possibility of memoizing a method call as an expression tree or delegate could offer significant performance benefits. On the side, I also want to see how C# reflection compares to Ruby method invocations.
I posted the full source to a public github repo. To quickly summarize, I wrote code that sets a property on an object 100 million times in a loop. Any setup (like finding a PropertyInfo or MethodInfo) is not included in the timings. I also checked the generated IL to make sure the compiler wasn't optimizing the loops. Please browse the code there if you need the gritty details.
Before I get into the implementation details, here are the results:
You can see that a reflection invoke is on the order of a hundred times slower than a normal property (set) invocation.
Here's the same chart but without the reflection invocation. It does a better job of showing the scale between the other tests.
Obviously, the lesson here is to directly invoke methods and properties when possible. However, there are times when you don't know what a type looks like at compile time. Again, object serialization/deserialization would be one of those use cases.
Here's an explanation of each of the tests:
Reflection Invoke (link)
This is essentially methodInfo.Invoke(obj, new[]{ value } on the setter method of the property. It is by far the slowest approach to the problem. It's also the most common way to solve the problem of insufficient pre-compile time knowledge.
Direct Invoke (link)
This is nothing other than obj.Property = value. Its as fast as it gets, but impractical for use cases where you don't have pre-compile time knowledge of the type.
Closure (link)
This isn't much more flexible than a direct invoke, but I thought it would be interesting to see how the performance degraded. This is where you create a function/closure ((x,y) => x.Property = y ) prior to the loop and just invoke the function inside the loop (action(obj, value)). At first sight it appears to be half as fast as a direct invoke, but there are actually two method calls involved here, so it's actually not any slower than a direct invoke.
Dynamic Dispatch (link)
This uses the C# 4.0 dynamic feature directly. To do this, I declared the variable as dynamic and assigned it using the same syntax as a direct invoke. Interestingly, this performs only 6x slower than direct invoke and about 20x faster than reflection invoke. Take note, if you need reflection, use dynamic as often as possible since it can really speed up method invocation.
Expression Tree (link)
The shortcoming of most of the previous approaches is that they require pre-compile time knowledge of the type. This time I tried building an expression tree (a C# 3.0 feature) and compiled a delegate that invokes the setter. This makes it flexible enough that you can call any property of an object without compile-time knowledge of the name, as long as you know the return type. In this example, like the closure, we're indirectly setting the property, so two method calls. With this in mind, it took almost 2.5 times as long as the closure example, even though they should be functionally equivalent operations. It must be that expression trees compiled to delegates aren't actually as simple as they appear.
Expression Tree with Dynamic Dispatch (link)
Since the expression tree approach requires compile-time knowledge of the return type, it isn't as flexible. Ideally you could use C# 4.0's covariance feature and cast it to Action
As expected, it's the slowest approach. However, its still 16 times faster than direct reflection. Perhaps, memoizing method calls, like property sets and gets, like this would actually yield a significant performance improvement.
Compared To Ruby
I thought I'd compare these results to Ruby where all method calls are dynamic. In Ruby, a method call looks first in the object's immediate class and then climbs the ladder of parent classes until it finds a suitable method to invoke. Because of this behavior I thought I would be interesting to also try a worst-case scenario with a deep level of inheritance.
To do this fairly, I initially wrote a while loop in Ruby that counted to 100 million. I rewrote the while loop in n.each syntax and saw the execution time get cut in half. Since I'm really just trying to measure method invocation time, I stuck with the n.each syntax.
I honestly thought C# Reflection would be significantly faster than the Ruby with 5 layers of in inheritance. While C# already holds a reference to the method (MethodInfo), Ruby has to search up the ladder for the method each time. I suppose Ruby's performance could be due to the fact that it's written in C and specializes in dynamic method invocation.
Also, it interests me why C# dynamic is so much faster than Ruby or reflection. I took a look at the IL code where the dynamic invoke was happening and was surprised to find a callvirt instruction. I guess I was expecting some sort of specialized calldynamic instruction (Java 7 has one). The answer is actually a little more complicated. There seems to be several calls - most are call instructions to set the stage (CSharpArgumentInfo.Create) and one callvirt instruction to actually invoke the method.
Conclusion
Since the trend of C# is going towards using more Linq, I find it interesting how much of a performance hit developers are willing to exchange for more readable and compact code. In the grand scheme of things, the performance of even a slow reflection invoke is probably insignificant compared to other bottlenecks like database, HTTP, filesystem, etc.
It seems that I've proved the point that I set out to prove. There is quite a bit of performance to be gained by memoizing method calls into expression trees. The application would obviously be best in JSON serialization, ORM, or anywhere when you have to get/set lots of properties on an object with no compile-time knowledge of the type. Very few people, if any, are doing this - probably because of the added complexity. The next step will be to (hopefully) build a working prototype.
I posted the full source to a public github repo. To quickly summarize, I wrote code that sets a property on an object 100 million times in a loop. Any setup (like finding a PropertyInfo or MethodInfo) is not included in the timings. I also checked the generated IL to make sure the compiler wasn't optimizing the loops. Please browse the code there if you need the gritty details.
Before I get into the implementation details, here are the results:
You can see that a reflection invoke is on the order of a hundred times slower than a normal property (set) invocation.
Here's the same chart but without the reflection invocation. It does a better job of showing the scale between the other tests.
Obviously, the lesson here is to directly invoke methods and properties when possible. However, there are times when you don't know what a type looks like at compile time. Again, object serialization/deserialization would be one of those use cases.
Here's an explanation of each of the tests:
Reflection Invoke (link)
This is essentially methodInfo.Invoke(obj, new[]{ value } on the setter method of the property. It is by far the slowest approach to the problem. It's also the most common way to solve the problem of insufficient pre-compile time knowledge.
Direct Invoke (link)
This is nothing other than obj.Property = value. Its as fast as it gets, but impractical for use cases where you don't have pre-compile time knowledge of the type.
Closure (link)
This isn't much more flexible than a direct invoke, but I thought it would be interesting to see how the performance degraded. This is where you create a function/closure (
Dynamic Dispatch (link)
This uses the C# 4.0 dynamic feature directly. To do this, I declared the variable as dynamic and assigned it using the same syntax as a direct invoke. Interestingly, this performs only 6x slower than direct invoke and about 20x faster than reflection invoke. Take note, if you need reflection, use dynamic as often as possible since it can really speed up method invocation.
Expression Tree (link)
The shortcoming of most of the previous approaches is that they require pre-compile time knowledge of the type. This time I tried building an expression tree (a C# 3.0 feature) and compiled a delegate that invokes the setter. This makes it flexible enough that you can call any property of an object without compile-time knowledge of the name, as long as you know the return type. In this example, like the closure, we're indirectly setting the property, so two method calls. With this in mind, it took almost 2.5 times as long as the closure example, even though they should be functionally equivalent operations. It must be that expression trees compiled to delegates aren't actually as simple as they appear.
Expression Tree with Dynamic Dispatch (link)
Since the expression tree approach requires compile-time knowledge of the return type, it isn't as flexible. Ideally you could use C# 4.0's covariance feature and cast it to Action
As expected, it's the slowest approach. However, its still 16 times faster than direct reflection. Perhaps, memoizing method calls, like property sets and gets, like this would actually yield a significant performance improvement.
Compared To Ruby
I thought I'd compare these results to Ruby where all method calls are dynamic. In Ruby, a method call looks first in the object's immediate class and then climbs the ladder of parent classes until it finds a suitable method to invoke. Because of this behavior I thought I would be interesting to also try a worst-case scenario with a deep level of inheritance.
To do this fairly, I initially wrote a while loop in Ruby that counted to 100 million. I rewrote the while loop in n.each syntax and saw the execution time get cut in half. Since I'm really just trying to measure method invocation time, I stuck with the n.each syntax.
I honestly thought C# Reflection would be significantly faster than the Ruby with 5 layers of in inheritance. While C# already holds a reference to the method (MethodInfo), Ruby has to search up the ladder for the method each time. I suppose Ruby's performance could be due to the fact that it's written in C and specializes in dynamic method invocation.
Also, it interests me why C# dynamic is so much faster than Ruby or reflection. I took a look at the IL code where the dynamic invoke was happening and was surprised to find a callvirt instruction. I guess I was expecting some sort of specialized calldynamic instruction (Java 7 has one). The answer is actually a little more complicated. There seems to be several calls - most are call instructions to set the stage (CSharpArgumentInfo.Create) and one callvirt instruction to actually invoke the method.
Conclusion
Since the trend of C# is going towards using more Linq, I find it interesting how much of a performance hit developers are willing to exchange for more readable and compact code. In the grand scheme of things, the performance of even a slow reflection invoke is probably insignificant compared to other bottlenecks like database, HTTP, filesystem, etc.
It seems that I've proved the point that I set out to prove. There is quite a bit of performance to be gained by memoizing method calls into expression trees. The application would obviously be best in JSON serialization, ORM, or anywhere when you have to get/set lots of properties on an object with no compile-time knowledge of the type. Very few people, if any, are doing this - probably because of the added complexity. The next step will be to (hopefully) build a working prototype.
Friday, February 3, 2012
Thoughts on the C# driver for MongoDB
I recently started a new job with a software company in Boulder. Our project this year is rewriting the existing product (not a clean rewrite, more like rewrite & evolve). One of the changes we're making is using MongoDB instead of T-SQL. Since we're going to be investing pretty heavily in Mongo we all attended the mongo conference in Boulder on Wednesday. The information was great and now I'm ready to dig into my first app. Today I played around with some test code and made some notes about features/shortcomings of the C# driver.
First of all, the so-called "driver" is much full featured than a typical SQL driver. It includes features to map documents directly to CLR objects (from here on I'll just say document if I mean Mongo BSON document and object for CLR object). There's plans to support Linq directly from the driver. So right off I'm impressed with the richness of the driver. However, I noticed some shortcomings.
For instance, all properties in the document must be present (and of the right type) in the object. I perceived this as a shortcoming because this is unlike regular JSON serialization where missing properties are ignored. After thinking a little further, this is probably what most C# developers would want since the behavior caters toward strongly typed languages that prefer fail-fast behavior. If you know a particular document might have extraneous properties that aren't in the object, you can use the BsonIgnoreExtraElements attribute.
Thinking about this behavior, refactor renaming properties could be less trivial. You would have to run a data migration script to rename the property (mongo does have an operation for renaming fields). It would be great if the driver had a [BsonAlias("OldValue")] attribute to avoid migration scripts (maybe I'll make a pull request).
Something I liked was that I could use object for the type of the _id property instead of BsonObjectId. This will keep the models less coupled to the Mongo driver API. Also, the driver already has a bi-directional alias for _id as Id. I don't know any C# developers who wouldn't squirm at creating a public property named _id.
This brings me to my biggest issue with the C# mongo driver. All properties must be public. This breaks the encapsulation and SRP principles. For instance, most of the time I have no reason to expose my Id (or _id) property as public. NHibernate solves this by hydrating protected fields. I would like this to be solved very soon (but there are some issues with this since there isn't any mappings).
Last, it has poor support for C# 4.0 types. Tuple doesn't fail, but it's serialized as an empty object ({ }). There is also zero support AFAIK for dynamic.
In conclusion, there's some room for improvement with Mongo's integration with .NET but overall I have to say I'm impressed. Supposedly Linq support is due out very soon, which will make it unstoppable (imo). Also, we haven't started using this in a full production environment yet, so there will most likely be more posts coming on this topic.
First of all, the so-called "driver" is much full featured than a typical SQL driver. It includes features to map documents directly to CLR objects (from here on I'll just say document if I mean Mongo BSON document and object for CLR object). There's plans to support Linq directly from the driver. So right off I'm impressed with the richness of the driver. However, I noticed some shortcomings.
For instance, all properties in the document must be present (and of the right type) in the object. I perceived this as a shortcoming because this is unlike regular JSON serialization where missing properties are ignored. After thinking a little further, this is probably what most C# developers would want since the behavior caters toward strongly typed languages that prefer fail-fast behavior. If you know a particular document might have extraneous properties that aren't in the object, you can use the BsonIgnoreExtraElements attribute.
Thinking about this behavior, refactor renaming properties could be less trivial. You would have to run a data migration script to rename the property (mongo does have an operation for renaming fields). It would be great if the driver had a [BsonAlias("OldValue")] attribute to avoid migration scripts (maybe I'll make a pull request).
Something I liked was that I could use object for the type of the _id property instead of BsonObjectId. This will keep the models less coupled to the Mongo driver API. Also, the driver already has a bi-directional alias for _id as Id. I don't know any C# developers who wouldn't squirm at creating a public property named _id.
This brings me to my biggest issue with the C# mongo driver. All properties must be public. This breaks the encapsulation and SRP principles. For instance, most of the time I have no reason to expose my Id (or _id) property as public. NHibernate solves this by hydrating protected fields. I would like this to be solved very soon (but there are some issues with this since there isn't any mappings).
Last, it has poor support for C# 4.0 types. Tuple doesn't fail, but it's serialized as an empty object ({ }). There is also zero support AFAIK for dynamic.
In conclusion, there's some room for improvement with Mongo's integration with .NET but overall I have to say I'm impressed. Supposedly Linq support is due out very soon, which will make it unstoppable (imo). Also, we haven't started using this in a full production environment yet, so there will most likely be more posts coming on this topic.
Monday, January 2, 2012
BDD ideas for structuring tests
Lately I've been thinking a lot about the best way to do BDD in C#. So when I saw Phil Haack's post about structuring unit tests, I think I had a joyful thought. Earlier I had been thinking in terms of using my Behavioral NUnit experimental project to hash out Haack's structuring idea with better BDD integration.
In short, his idea is to use nested classes. There is the normal one-to-one class-to-test-class mapping, but each method under test gets it's own inner class. To use his example:
In this example the Titleify and Knightify methods (imo two terrible uses of the -ify suffix) have corresponding test classes dedicating to testing only one method. Each method in the class (or Fact, in the case of xUnit. I actually haven't used xUnit but it seems to encourage a somewhat BDD readability) test one aspect of the method, much like the it method is used in rspec.
I generally like Haack's test structure. For example, he points out how it plays nicely with Visual Studio's natural class/method navigation which makes the tests even more navigable. The only issue I have with it is that I dislike having 1000+ SLOC classes - tests or regular. If I were to adopt this method, I would probably break each of those inner classes into separate files (and use partial classes to break up the top class).
My practice for a long time was to have one whole namespace per class under test. Consider my tests for objectflow. I actually picked up this practice from Garfield Moore, objectflow's original developer. Each class (or significant concept) has a namespace (e.g. objectflow.stateful.tests.unit.PossibleTransitions or PossibleTransitionTests). Each class in that namespace is names according to essentially what the Setup does. Some examples: WhenGivenOnlyBranches, WhenGivenOnlyYields, etc.
I like the way these tests read. It's very easy to find a particular test or to read up on how a particular method is supposed to operate. But in practice this has led to very deep hierarchies, often with single class namespaces. Further, I find that creating a whole new class for each setup tends to create too much extra code. As a result, I have a hard time sticking closely to this practice.
More recently I've felt a little overwhelmed with my original practice so I've evolved it slightly. Now I've started doing the one-to-one class to test mapping like commonly practiced. But each test has it's own method that does setup. For instance
I also sometimes use this small variation of that structure where I keep the BDD sentence-style naming scheme but use TestCase attributes to quickly cover edge cases.
I often use some hybrid of the last two approaches, especially if I would be using a TestCase attribute that breaks the BDD readability, I'll break the setup code into one of those Given_* support setup methods and reuse it between two different test methods.
I generally like my most recent ways of structuring tests because of it's readability and ability to gain excellent edge case coverage by adding additional test cases. But I do really like Haack's structuring, so I may find myself adopting part of his suggestion and further evolving my tests.
As far as this applies to Behavioral NUnit, I want to explore the possibility of a Describe attribute that mimics the usage of rspec's describe method. One idea is to make the new attribute generate another hierarchical level of test cases
In short, his idea is to use nested classes. There is the normal one-to-one class-to-test-class mapping, but each method under test gets it's own inner class. To use his example:
In this example the Titleify and Knightify methods (imo two terrible uses of the -ify suffix) have corresponding test classes dedicating to testing only one method. Each method in the class (or Fact, in the case of xUnit. I actually haven't used xUnit but it seems to encourage a somewhat BDD readability) test one aspect of the method, much like the it method is used in rspec.
I generally like Haack's test structure. For example, he points out how it plays nicely with Visual Studio's natural class/method navigation which makes the tests even more navigable. The only issue I have with it is that I dislike having 1000+ SLOC classes - tests or regular. If I were to adopt this method, I would probably break each of those inner classes into separate files (and use partial classes to break up the top class).
My practice for a long time was to have one whole namespace per class under test. Consider my tests for objectflow. I actually picked up this practice from Garfield Moore, objectflow's original developer. Each class (or significant concept) has a namespace (e.g. objectflow.stateful.tests.unit.PossibleTransitions or PossibleTransitionTests). Each class in that namespace is names according to essentially what the Setup does. Some examples: WhenGivenOnlyBranches, WhenGivenOnlyYields, etc.
I like the way these tests read. It's very easy to find a particular test or to read up on how a particular method is supposed to operate. But in practice this has led to very deep hierarchies, often with single class namespaces. Further, I find that creating a whole new class for each setup tends to create too much extra code. As a result, I have a hard time sticking closely to this practice.
More recently I've felt a little overwhelmed with my original practice so I've evolved it slightly. Now I've started doing the one-to-one class to test mapping like commonly practiced. But each test has it's own method that does setup. For instance
I also sometimes use this small variation of that structure where I keep the BDD sentence-style naming scheme but use TestCase attributes to quickly cover edge cases.
I often use some hybrid of the last two approaches, especially if I would be using a TestCase attribute that breaks the BDD readability, I'll break the setup code into one of those Given_* support setup methods and reuse it between two different test methods.
I generally like my most recent ways of structuring tests because of it's readability and ability to gain excellent edge case coverage by adding additional test cases. But I do really like Haack's structuring, so I may find myself adopting part of his suggestion and further evolving my tests.
As far as this applies to Behavioral NUnit, I want to explore the possibility of a Describe attribute that mimics the usage of rspec's describe method. One idea is to make the new attribute generate another hierarchical level of test cases
Friday, December 30, 2011
Can Bad Code Ruin Your Career?
I started writing this post over a year ago. I was working at a large company where I was stuck in a mouse wheel - always running to keep up but never getting anywhere. The code I had to work with was downright terrible. This, among other things, prodded me into looking for another job. While I was starting my job search I was pondering this post and decided to not finish it because I wasn't sure if some prospective employer would hold it against me.
I just finished reading through a messy Java file. It was the usual mess of a class with a 500 line god-method (similar to the god-object) and hundreds of counts of copy and pasted code. Besides the redundant code and lack of structure the coder also used nested loops through ArrayLists when they could have used a HashSet and didn't once use generic collections, using the un-type checked versions instead. After several hours of refactoring and renaming variables I finally got to a point where I could begin fixing the bug I was after. There were absolutely no unit tests - all this code was written inline with HTML in a JSP.
I spend so much time reading bad code that sometimes I wonder if I am beginning to specialize in hacks. Is it possible to read so much bad code that you forget what good code looks like? Humans are an especially adaptive species, and I think it's definitely possible that a great programmer can be forced to work in the muck so long that they forget what good code looks like.
I've seen several situations where good developers produced bad code. These situations are almost always a product of an environment where features are more important than bug fixes. These companies typically invest heavily in sales and neglect IT and development costs. Or sometimes the problem is just that product management knows nothing of software development.
A recent coworker likened our job of working with brittle, badly designed code to the 5 stages of grief. While we were uneasily laughing about it I silently decided that this was more realistic than I wanted to believe.
For instance, imagine starting a new job. In the interview process you were interviewed by intelligent, enthusiastic developers and were led to believe you were going to be working on cutting edge technologies - a dream right? When you actually get to the job you find out that the code is so backwardly complicated that its nearly impossible to touch anything without bringing the proverbial house of cards crashing down.
Grief Stage 1: Denial and Isolation
Obviously the code isn't the problem, you just weren't careful enough. They probably have specific guidelines and strategies that help them be more productive. It's probably just something wrong with me...
Grief Stage 2: Anger
Dammit! Who the hell even thinks of this crap? [more cursing...] Is this a god-object?? [hair gets thinner...]
Grief Stage 3: Bargaining
This is typically when you start plotting potential strategies to hide the ugliness of the code. Creativity and hopeful thoughts abound. Many IT managers will talk like they are very supportive of you at this stage.
Grief Stage 4: Depression
This is where the reality strikes that this stage is bad for the business plan because it involves spending less time on revenue-producing features. The IT managers that seemed so supportive now flip flop to the CEO's side and deny you the ability to cope with your problems
Grief Stage 5: Acceptance
There are only two outcomes of this stage. Either (1) you accept that you can never fix the code so you decide to move on to another job or (2) you accept that you can never fix the code so you give up on trying. This is what separates good coders from bad.
Again, I started this post over a year ago. I've seen a lot of bad code. At my most recent job I almost took the "give up on trying" path in the acceptance stage. Luckily we hired a great older developer who snapped me out of it. I just started my new job today, I think I will be much happier.
So can bad code ruin your career? My answer is a resounding YES! But it doesn't have to. Honestly, stage 5 can have better endings, but that inevitably requires understanding on behalf of management - a scarce resource.
With that said...
I just finished reading through a messy Java file. It was the usual mess of a class with a 500 line god-method (similar to the god-object) and hundreds of counts of copy and pasted code. Besides the redundant code and lack of structure the coder also used nested loops through ArrayLists when they could have used a HashSet and didn't once use generic collections, using the un-type checked versions instead. After several hours of refactoring and renaming variables I finally got to a point where I could begin fixing the bug I was after. There were absolutely no unit tests - all this code was written inline with HTML in a JSP.
I spend so much time reading bad code that sometimes I wonder if I am beginning to specialize in hacks. Is it possible to read so much bad code that you forget what good code looks like? Humans are an especially adaptive species, and I think it's definitely possible that a great programmer can be forced to work in the muck so long that they forget what good code looks like.
I've seen several situations where good developers produced bad code. These situations are almost always a product of an environment where features are more important than bug fixes. These companies typically invest heavily in sales and neglect IT and development costs. Or sometimes the problem is just that product management knows nothing of software development.
The 5 stages of grief
A recent coworker likened our job of working with brittle, badly designed code to the 5 stages of grief. While we were uneasily laughing about it I silently decided that this was more realistic than I wanted to believe.
For instance, imagine starting a new job. In the interview process you were interviewed by intelligent, enthusiastic developers and were led to believe you were going to be working on cutting edge technologies - a dream right? When you actually get to the job you find out that the code is so backwardly complicated that its nearly impossible to touch anything without bringing the proverbial house of cards crashing down.
Grief Stage 1: Denial and Isolation
Obviously the code isn't the problem, you just weren't careful enough. They probably have specific guidelines and strategies that help them be more productive. It's probably just something wrong with me...
Grief Stage 2: Anger
Dammit! Who the hell even thinks of this crap? [more cursing...] Is this a god-object?? [hair gets thinner...]
Grief Stage 3: Bargaining
This is typically when you start plotting potential strategies to hide the ugliness of the code. Creativity and hopeful thoughts abound. Many IT managers will talk like they are very supportive of you at this stage.
Grief Stage 4: Depression
This is where the reality strikes that this stage is bad for the business plan because it involves spending less time on revenue-producing features. The IT managers that seemed so supportive now flip flop to the CEO's side and deny you the ability to cope with your problems
Grief Stage 5: Acceptance
There are only two outcomes of this stage. Either (1) you accept that you can never fix the code so you decide to move on to another job or (2) you accept that you can never fix the code so you give up on trying. This is what separates good coders from bad.
Conclusion
Again, I started this post over a year ago. I've seen a lot of bad code. At my most recent job I almost took the "give up on trying" path in the acceptance stage. Luckily we hired a great older developer who snapped me out of it. I just started my new job today, I think I will be much happier.
So can bad code ruin your career? My answer is a resounding YES! But it doesn't have to. Honestly, stage 5 can have better endings, but that inevitably requires understanding on behalf of management - a scarce resource.
Wednesday, December 28, 2011
Behavior Driven Development in C#
I've been a fan of Test Driven Development since I worked in an XP shop. But every time the work starts getting bigger and more complex I always struggle to not get lost in the magnitudes of tests. I remember many early-on conversations with my elders about unit test naming conventions. The [method]_[input]_[output] convention starts to break down badly when your inputs become things like mocks, or if there ends up being more than 1 or 2 inputs; same with outputs.
When a coworker introduced me to BDD earlier this year, it really clicked and flowed naturally. The idea of writing tests so they read like sentences out of a book or spec seems like the answer to all my questions. The ruby rspec is beautiful:
The organization of the tests forces you to focus on the expectations of your test and highlight descriptive assertions. This is especially useful for complicated setups with lots of mocks, etc. I put as much of my setup code in one of those before :each blocks, so that way the assertions are limited to simple inputs and one or two observations about the outputs.
There's been a number of people in the .NET community that have attempted BDD but [imo] failed to grasp the simplicity. NBehave is a complete overhaul of unit testing that uses attributes like xUnit. As a result, NBehave doesn't really look at all like rspec - which really isn't a bad thing, necessarily. However, the thing I like about rspec is it's ability to describe things of arbitrary depth, which is handy when testing complex code:
This spec is able to describe possible modes that the object under test can be in (complex inputs). This is made possible by rspec's arbitrary nesting depth. This is definitely a language feature that is much harder to implement in C#.
My current approach to BDD in C# usually looks like
I think this is the simplest BDD layer I can slap on top of NUnit. And simple is important to me because (a) I do a lot of open source projects and I want to keep the barrier to entry for contributions low and (b) the people I work with tend to resist change. When people are resistant to change, it's hard to rationalize using something other than NUnit or introducing lots of nested lambdas.
NUnit remains the most popular unit testing framework and has excellent support with a GUI runner, console runner, and IDE integration with R#, TestDriven.NET, and others. Given all that support, I would really rather not abandon NUnit if possible.
FluentAssertions is a nice simple BDD layer on top of NUnit (or whatever you use). It doesn't change the structure of our spec above, but it does change the structure of our assertion to
This assertion is [imo] very clean and succinct. I like how it reads even clearer than NUnit's fluent syntax. Last weekend I was thinking about this and I decided to explore an idea to make a BDD extension to NUnit that is even clearer than FluentAssertions. The project, BehavioralNUnit for now, is hosted at github. The earliest goal for the project was simply to use operator overloading to make the assertions even more like rspec. For instance, I want to be make the previous assertion:
I was able to do this, but I realized that the C# compiler was insisting that this expression needed to be assigned to something, so I [haven't yet] added another concept somewhat analogous to "it" in rspec:
This is most similar to NSpec's approach by using an indexer instead of a method. This appeals to me because I sometimes find matching parentheses to be a pain (I guess I just like ruby & coffeescript). Then again, I don't like NSpec because it feels like it was written by one of those whining .NET developers that wishes dearly he could get a RoR job - it doesn't abide to .NET conventions at all.
I still have a ton of ideas to hash out with Behavioral NUnit. I'm convinced that BDD in C# can be simpler and more beautiful than it currently is. If you have input or ideas, please fork the repository & try out your ideas (pull requests are welcome).
When a coworker introduced me to BDD earlier this year, it really clicked and flowed naturally. The idea of writing tests so they read like sentences out of a book or spec seems like the answer to all my questions. The ruby rspec is beautiful:
The organization of the tests forces you to focus on the expectations of your test and highlight descriptive assertions. This is especially useful for complicated setups with lots of mocks, etc. I put as much of my setup code in one of those before :each blocks, so that way the assertions are limited to simple inputs and one or two observations about the outputs.
There's been a number of people in the .NET community that have attempted BDD but [imo] failed to grasp the simplicity. NBehave is a complete overhaul of unit testing that uses attributes like xUnit. As a result, NBehave doesn't really look at all like rspec - which really isn't a bad thing, necessarily. However, the thing I like about rspec is it's ability to describe things of arbitrary depth, which is handy when testing complex code:
This spec is able to describe possible modes that the object under test can be in (complex inputs). This is made possible by rspec's arbitrary nesting depth. This is definitely a language feature that is much harder to implement in C#.
My current approach to BDD in C# usually looks like
I think this is the simplest BDD layer I can slap on top of NUnit. And simple is important to me because (a) I do a lot of open source projects and I want to keep the barrier to entry for contributions low and (b) the people I work with tend to resist change. When people are resistant to change, it's hard to rationalize using something other than NUnit or introducing lots of nested lambdas.
NUnit remains the most popular unit testing framework and has excellent support with a GUI runner, console runner, and IDE integration with R#, TestDriven.NET, and others. Given all that support, I would really rather not abandon NUnit if possible.
FluentAssertions is a nice simple BDD layer on top of NUnit (or whatever you use). It doesn't change the structure of our spec above, but it does change the structure of our assertion to
This assertion is [imo] very clean and succinct. I like how it reads even clearer than NUnit's fluent syntax. Last weekend I was thinking about this and I decided to explore an idea to make a BDD extension to NUnit that is even clearer than FluentAssertions. The project, BehavioralNUnit for now, is hosted at github. The earliest goal for the project was simply to use operator overloading to make the assertions even more like rspec. For instance, I want to be make the previous assertion:
I was able to do this, but I realized that the C# compiler was insisting that this expression needed to be assigned to something, so I [haven't yet] added another concept somewhat analogous to "it" in rspec:
This is most similar to NSpec's approach by using an indexer instead of a method. This appeals to me because I sometimes find matching parentheses to be a pain (I guess I just like ruby & coffeescript). Then again, I don't like NSpec because it feels like it was written by one of those whining .NET developers that wishes dearly he could get a RoR job - it doesn't abide to .NET conventions at all.
I still have a ton of ideas to hash out with Behavioral NUnit. I'm convinced that BDD in C# can be simpler and more beautiful than it currently is. If you have input or ideas, please fork the repository & try out your ideas (pull requests are welcome).
Monday, December 26, 2011
Why I hate generated code
If you've worked with me for any amount of time you'll soon figure out that I often profess that "I hate generated code". This position comes from years of experience with badly generated code. Let me explain.
The interesting thing about this work is how little of the generated code was actually being used. I'm sure, in the beginning, the developer's thoughts were along the lines "oh look at all this code I don't have to write manually :D". However, after some time, subsequent developer's thoughts were along the lines of "with all this dead code, it's hard to find real problems". It's funny how some exciting breakthroughs turn into headaches down the road. The table is always used, but some entities are created & read but never modified, others are only created during migrations and only read from during run time.
Code generators often produce code you don't need. Since all code requires maintenance, dead code is just a liability because it doesn't provide any benefit. I always delete dead code and commented out code (it'll live on in version control, no need to release it into production).
There are several professional developer communities that generate code as a way of life. Ruby on Rails comes prepackaged with scripts to generate models, views, and controllers in a single command. ASP.NET MVC will generate controllers and views with a couple clicks. And if you've ever used either of these frameworks, you'll probably find yourself deleting a lot of generated code.
While I hate code generators, it's hard to argue how I could possibly hate compilers. They allow me to write code once and compile it several different ways and achieve different goals. Therefore, I have to introduce my first caveat - I don't hate all generated code, I only hate generated source code.
This problem of hating generated code is complicated further by the fact that NHibernate generates source code too. You don't ever check in the code that NHibernate generates because it's done at run time. The most obvious way NHibernate generates code is the SQL that is written in the background to query & perform DML operations. (For those questioning if SQL is source code, consider how SQL is compiled into an execution plan prior to execution). It's also hard to argue that I hate this kind of code generation because it doesn't suffer from the same problems of the CodeSmith generated code. It only generates code just-in-time meaning that it's only generated when needed, so there isn't any extra code generated.
Since NHibernate and compilers do code generation in a way that I like, I'm going to refine my statement to "I hate generated persistent code". This generally means, I still hate generated code when the resulting code sticks around long enough for a fellow developer to have to deal with it.
Since C# doesn't yet (and probably won't ever) include variadic templates or variadic generic types, writers of .NET API's often write some really redundant code to account for all combinations of generic methods or types. I know I've done it. This example uses a T4 template to produce a C# file with a *.generated.cs extension. The T4 template is executed on build but not ignored from version control.
I do like this approach because it takes a DRY approach to a redundant problem without much complication. Another thing I really like about this approach is that T4 templates are a standard part of Visual Studio and are executable from Mono as well. As such, they can be considered a free tool that is openly available (important for open source projects) and, more importantly, are executed as part of the build process.
Another thing I like about this approach is the usage of partial classes to separate the generated portion of the class from the non-generated portion. This minimizes the amount of code that is sheltered from refactoring tools (code inside the *.tt file).
The thing I hate about this particular iSynaptic.Commons example is that the generated file is included in version control. I think, perhaps, this is reduced to a small pet peeve of mine since the generated code isn't wasteful and is updated on every build. Still, I would like a mechanism to (a) have the file ignored from the IDE's perspective and (b) ignored from version control. I wouldn't want anyone to mistakenly edit the file when they should be editing the T4 template.
The baby comes with a lot of bathwater
In the past year I had an experience with a generated data layer where CodeSmith was used to generate a table, 5 stored procedures, an entity class, a data source class, and a factory class for each entity that was generated. My task was to convert this code into NHibernate mappings.The interesting thing about this work is how little of the generated code was actually being used. I'm sure, in the beginning, the developer's thoughts were along the lines "oh look at all this code I don't have to write manually :D". However, after some time, subsequent developer's thoughts were along the lines of "with all this dead code, it's hard to find real problems". It's funny how some exciting breakthroughs turn into headaches down the road. The table is always used, but some entities are created & read but never modified, others are only created during migrations and only read from during run time.
Code generators often produce code you don't need. Since all code requires maintenance, dead code is just a liability because it doesn't provide any benefit. I always delete dead code and commented out code (it'll live on in version control, no need to release it into production).
There are several professional developer communities that generate code as a way of life. Ruby on Rails comes prepackaged with scripts to generate models, views, and controllers in a single command. ASP.NET MVC will generate controllers and views with a couple clicks. And if you've ever used either of these frameworks, you'll probably find yourself deleting a lot of generated code.
The problem of transient code generation
The issue that I keep running into with my policy of hating code generation is that it's nearly impossible to be a professional software engineer and not generate code. The most fundamental problem is compilers. When you run a compiler over your source code, it generates some sort of machine readable code that is optimized for various goals like speed or debugging or different platform targets.While I hate code generators, it's hard to argue how I could possibly hate compilers. They allow me to write code once and compile it several different ways and achieve different goals. Therefore, I have to introduce my first caveat - I don't hate all generated code, I only hate generated source code.
This problem of hating generated code is complicated further by the fact that NHibernate generates source code too. You don't ever check in the code that NHibernate generates because it's done at run time. The most obvious way NHibernate generates code is the SQL that is written in the background to query & perform DML operations. (For those questioning if SQL is source code, consider how SQL is compiled into an execution plan prior to execution). It's also hard to argue that I hate this kind of code generation because it doesn't suffer from the same problems of the CodeSmith generated code. It only generates code just-in-time meaning that it's only generated when needed, so there isn't any extra code generated.
Since NHibernate and compilers do code generation in a way that I like, I'm going to refine my statement to "I hate generated persistent code". This generally means, I still hate generated code when the resulting code sticks around long enough for a fellow developer to have to deal with it.
The thin line between good and bad code generation
When is generated code persistent and when is it transient? We already decided that code generation isn't so bad when it happens during of after the compilation process. But my statement is that I hate persistent code. There are other cases of code generators generating transient source code. One such example is in iSynaptic.Commons.Since C# doesn't yet (and probably won't ever) include variadic templates or variadic generic types, writers of .NET API's often write some really redundant code to account for all combinations of generic methods or types. I know I've done it. This example uses a T4 template to produce a C# file with a *.generated.cs extension. The T4 template is executed on build but not ignored from version control.
I do like this approach because it takes a DRY approach to a redundant problem without much complication. Another thing I really like about this approach is that T4 templates are a standard part of Visual Studio and are executable from Mono as well. As such, they can be considered a free tool that is openly available (important for open source projects) and, more importantly, are executed as part of the build process.
Another thing I like about this approach is the usage of partial classes to separate the generated portion of the class from the non-generated portion. This minimizes the amount of code that is sheltered from refactoring tools (code inside the *.tt file).
The thing I hate about this particular iSynaptic.Commons example is that the generated file is included in version control. I think, perhaps, this is reduced to a small pet peeve of mine since the generated code isn't wasteful and is updated on every build. Still, I would like a mechanism to (a) have the file ignored from the IDE's perspective and (b) ignored from version control. I wouldn't want anyone to mistakenly edit the file when they should be editing the T4 template.
Summary
The end result of my thought is "I hate source code that is generated prior to the build process". I want to further say that I also hate generated code that is checked into version control, but this is a bit of a lesser point. However, code generation can be a useful tool; as seen in the cases of NHibernate and T4 templates. But even still, code generation should be used wisely and with care. Generating excess code can become a liability that detracts from the overall value of a product.
Subscribe to:
Posts (Atom)