Can't Driven Development
rm4n0s.github.io
rm4n0s.github.io
Static constraints like types can also be also good, where types are good.
I applaud anyone who advocates to focus on the unhappy path over the happy path.
On a side note 25 years ago was the end of the dot.com and the height of the y2k scam and I have never seen developers more burnt out than then; but that’s because I was just starting my career then. I don’t know about previous eras like the 89 crash and recession.
I am not sure why the OP has rose coloured glasses of the past. But it’s always good to treat history with a little more interest.
php apache mysql linux
it was so small we could put it into an acronym.
Meanwhile in 2024 you could barely begin to even describe any individual layer of the stack with so few words. You'd be out of breath before even beginning to finish describing React.
(btw it's silicon, not silicone).
Because the option to parse stack traces from the programming language itself, it will destroy what ever built on or for microservices
Vectorization, more cache levels, branch prediction, multiple cores and processors, gpu, virtual machines, (operating systems, but they didn't change much, I only let the placeholder for the layer), containers, but all that in the cloud, distributed databases, frameworks written in a language that need to be compiled to a language that is actually implemented by a web browser in the client (we are far better on compatibility between browsers, but much libraries or practices keep stuck in the past).
I haven't touched shift in practices that make it far worse.
That make it more unreasonable.
Secondly, companies need to realize that they request multiple programming languages because the current programming languages can't parse stack traces, and they push everything to microservices, so they can parse their errors from DevOps services. Which means they don't request only to know multiple languages, but also DevOps services.
I have written about all these in a previous article of mine https://rm4n0s.github.io/posts/5-it-will-take-your-job
in the article I wrote how detrimental it is to the developer's social life, to know multiple languages.
We learned this in high school and it was super fun: https://en.wikipedia.org/wiki/Busy_beaver
Do I expect everyone to be an instant expert in any language? Or able to create the busiest beaver? No of course not! But I do expect any decent programmer to pick up any programming language and work in it.
Do we all have preferences? Heck yes! I absolutely dislike Python and would not take a job where I have to program in it if I wasn't forced to by circumstances. I've never done C# or Lua professionally but it was super easy to pick both up when I dabbled with them for game development in private (each respective game engine used these languages for scripting).
Secondly, companies need to realize that they request multiple programming languages because the current programming languages can't parse stack traces
FWIW, yes, I've actually looked at previous stack frames (e.g. `e.getCause()` in Java) for error handling. It always felt very very dirty but it was my only choice under the circumstances.
they push everything to microservices
Some companies (meaning the people within those companies) because they read about how Netflix has x-many of them and it's the thing to do today. Stay away from those companies. They will also make you do other things just because someone read an article of forbes.com.
And then there's companies that are actually trying to solve similar problems as the ones that Netflix was trying to solve with microservices and it actually makes sense to model their internal application service landscape that way! We're in that space for example. We have one big legacy "macroservice" aka very monolithic core and then we have a bunch of newer microservices (some larger, some smaller) that actually solve problems. All of these microservices are written in exactly two languages: Java and Kotlin. Newer ones in Kotlin, older ones in Java (like the legacy monolith). The FE is written in exactly one language by now: Typescript (converted from quite a large javascript code base over multiple years).
so they can parse their errors from DevOps services
I'm really sorry to say this, but it seems like your articles are really just ranting about things that you personally don't like for one reason or another but there's no actual real reasoning or universal truth to it.
When you try to implement it, then everything on what I said will make sense to you.
Also, we don't need log systems when there is a programming language like Odin to parse stack traces with type checking (not just string like you gave me as an example from Java).
In microservices you are the error handlers. For example, if in the logs you see a stack trace, then you will go to the code and fix the error.
In Odin I don't need to go in that trouble, because ALL the stack traces can be handled. There will be no undefined behavior in software or unexpected input that caused an unexpected stack trace, so there is no reason to have logs.
Let's say you have the call stack as this (from your example):
f4()->
f3()->
f1()->
ErrInvestmentLost
But nobody would ever want to do this because it's super brittle. I change f4 so that it first calls f17, which then calls f3, which then calls f17 again, which calls f1 and your error handling based on the call path is suddenly broken.
What is it that you are trying to even achieve by doing this? Proper error handling doesn't depend on the call path. Proper error handling depends on the type of error that occurred and whether you can actually handle it at all or if you just have to give up and throw the error all the way to where it will get logged for a programmer to take a look at why it happened and why we couldn't handle it.
Your claim about being able to handle "all stack traces" makes no sense to me. You don't handle stack traces. You handle error types.
A real world example of the above (taking Java as an example again) might be a REST resource. My error handling should not depend on nor suddenly break, just because someone configured a new filter in the filter chain that sits above the actual resource method. Say someone added in a `AuthenticationFilter` that checks if some auth token is present and valid and that didn't used to be the case. Now any error handling in my resource method that was based on the exact stack trace combination that existed before that filter was added will break horribly.
Your system with Java will break if someone else add an AuthenticationFilter, but my system in Odin will not even compile until I have handled all the stack trace paths that include AuthenticationFilter.
Do you see the differences between handling stack traces with union types rather than string?
See, the `AuthenticationFilter` sits outside of my REST resource. I could not care less that someone configured it and that at runtime, based on some configuration that can change without even needing a recompilation, this filter will either be there on the stack or it won't.
My resource does not interact with this filter in any way and when an error happens somewhere down in another method I call, then I don't care that I was called with or without the auth token having been checked by said filter. I might care whether the method I called threw a `SQLException` or a `JSONParseException` but very probably I don't even care about that at all because I can't do anything specific in either case and will just throw it further (i.e. not handle it, other than potentially logging it).
Java actually tried the whole "specify all error situations with checked exceptions and otherwise the code won't even compile" and it failed miserably and you are hard pressed to see anything new derive from `Exception`. Everyone uses `RuntimeException`. It does come at a cost, because now I no longer have the hassle of explicitly knowing and deciding what to do with these exceptions and I may only figure out that a particular type of error can happen once I "see it in the wild" (e.g. in my logs, coz something failed) or I'm lucky enough to have actually read the documentation and handled all the exceptions I wanted to handle.
But that happened precisely because it was just too much to have all your code specify these exceptions when everyone figured out that 99% of all code just threw them further up the stack. You call one new library method that specifies an exception and you suddenly have to adjust 127 other files and the only thing you do is to declare all those methods will also just throw the exception further up the stack.
For example, in Authentication_Filter_Error union, you will have another union called SQL_Verify_Account_Error, that it will contain SQL_Error enum with the Closed_Conn value.
Imagine your stack trace like this Authentication_Filter_Error -> SQL_Verify_Account_Error -> SQL_Error.Closed_Conn
Now when you know that can happen (through CDD), you can create a switch statement to catch the specific stack trace, to call the system administrator in the middle of the night to check what happens.
This is how software should handle its errors and there is not even a need to log it.
In your scenario, you wake up, you go to work, everyone is screaming at the office, you check the logs, you see the problem, and then you call the system administrator for the problem.
However, that's still about the exceptions thrown from down thread, not from the call path part of the "stack trace".
I.e. your situation would never happen.
Authentication_Filter_Error -> SQL_Verify_Account_Error -> SQL_Error.Closed_Conn
But let's say there was a distinction made with proper exception types and instead of `SQLException("Connection closed")` and `SQLException("Statement timeout")`, I actually received `SQLException(ConnectionClosedException())` vs. `SQLStatementTimeoutException`. Now, without string parsing, I can know that either the connection just closed or that the statement was aborted due to timeout. If these are checked exceptions, I have to declare that I'm aware these can happen and what I want to do with them: Handle or rethrow.
However, a myriad of such exceptions can happen. I would probably have to declare 20-50 exceptions way up in a REST resource layer. Not only can these two happen, but many other situations on the network or database side and on the JSON parsing side for the payload I receive, some exceptions from my business logic etc.
And for most of these, what can I do? If the connection to the database closed, all I can do is to log the error and return a `500 Internal Server Error` to my caller. Guess what I can do when a statement timeout occurs? I log the error and return a `500 Internal Server Error` to my caller. For a statement timeout I can't even return a `400 Bad Request`, because it's not knowable if the statement timeout occurred because the database was simply overloaded in that moment or if the request itself was created with such parameters as to always cause a statement timeout. Until we see the logs and through investigation figure out that it wasn't a bad request after all anyway. We were missing an index and the table finally grew large enough for that to matter.
So yeah, I'm good with `RuntimeException` and handling only very few specific ones ever.
Also nobody will be screaming when the token is invalid and I definitely don't call any system administrator. That's something you as a developer look into. Same with the statement timeout.
You could make the software call the system administrator and return a message to the user "Try again in an hour, the system administrator is fixing it now"
Or if it is a timeout, the software will call amazon to buy a new machine to scale the database and send a message to the user "Try again in an hour until we scale the system".
Developer's job is to automate error handlers and not be the error handlers.
If you can see the stack trace tree, then you can plan far ahead, but to do that we need to destroy this "agile" mindset, that is always in a hurry and doesn't let you to think that far ahead.
you can make the software to handle these kind of errors.
You could make the software call the system administrator
"Try again in an hour, the system administrator is fixing it now"
Or if it is a timeout, the software will call amazon to buy a new machine to scale the database
Developer's job is to automate error handlers and not be the error handlers.
If you can see the stack trace tree
we need to destroy this "agile" mindset, that is always in a hurry and doesn't let you to think that far ahead.
Ask your boss, "do you want your users wait until the next morning for the system administrator to wake up and fix the issue? Or do you want the software inform the user how long it will take to fix the issue in seconds and start calling the system administrator to wake up and fix it?"
Because you answer based on your preferences as a worker who wants to avoid the extra work to make the system perfect and not your boss's preferences.
And yes, there's an on-call person that does get paged when something happens that likely needs immediate attention. A page for every single time there's any error? Not bloody likely mate.
To pick up your last point: My boss is not in the business of paying for you, who will spend countless extra days building useless error "handling" for stuff that has already been handled and who is trying to get out of the responsibility of writing resilient software by paging someone else to "pick up the tab".
A "boss" never wants to pay for a perfect system. That would take way too long and nobody has figured out how to actually build that anyway (no, Odin is not the answer). They want to pay for the "slightly less than good enough" system, because that's cheaper and still gets the job done. And especially when I hear you talk here, I'm with them: Perfect is the enemy of good enough. We just have to ensure that it really is good enough and not less (coz many a boss will happily take way less than good enough if it gets them to market faster.
Also in a similar regard your article mentions other things like microservices being the root of all evil so to speak and you need to know multiple services.
When in fact microservices are meant to do the opposite. You are supposed to be able to focus on just your service and what it needs to do and make it do it well. Instead of having to wade through a huge monolithic and probably very interwoven code base. We can of course talk about whether microservices achieved that goal.
They are not fundamentally different from having worked in large enterprises 25 years ago where you would've encountered many many different services making up your application landscape. It was probably using SOA and the services were deployed to a bunch of application servers like an IBM WebSphere and those services might be talking to other services incidentally deployed on some JBoss app servers and they'd talk via message passing via a queue (JMS comes to mind but other ways existed).
today fullstack developers have to pivot between 2-3 languages on the same day
All the SOA back then was JAVA, they had just different types of HTTP servers to run JAVA
I constantly write software in Typescript, Java and Kotlin. Some days it's all of them other days it's only a subset. I don't see an issue with that at all. If you don't have a standard set of languages I would agree with you but a stack of one frontend and one backend language should not be a problem for anyone that calls themselves full stack.
The SOA point was not about languages any longer. It was about having multiple services that make up an overall feature and app landscape. Also you seem to have missed the part where I said that I was not in fact HTTP calls is manyif not most cases ;)
I fail to see how one has anything to do with the other.
Knowing and applying both Java and Typescript is not an issue at all. What is a pain is CSS and why I tell everyone I'm a BE dev. But I happily build vertical slices of functionality where I have full control over FE and BE code and can adjust as I see fit instead of having to create or follow a rigid interface definition and where the BE isn't user testable when it gets written. What is a pain is the libraries on the FE but that's not an issue with knowing multiple programming languages but with the web dev FE ecosystem in general. A non full stack FE developer would have the same pain.
While we did learn a programming language at university to code our exercises in, we didn't learn that language specifically as "this is the language we teach and this is what you will use in your job until you retire". On the contrary we were told and taught about programming concepts, data structures and algorithms and it was expected that you can program in any language. The (mandatory) operating systems course for example would just assume that you learned enough C by yourself to do the programming exercises. We had to write a working memory compactification algorithm for Minix. In the scripting language course (non mandatory) you had to choose any scripting language you wanted and implement a project in it. In the mandatory hardware course we were expected to pick up enough VHDL to implement whatever we learned in the course.
You can't program in any language, don't fool yourself. Take, for example, SQL that is used in all companies. Can you write SQL functions with the same ease as Java? You may ask why do you need functions in SQL? Because it is faster to run the algorithm in SQL, rather than in Java from an ORM.
I'm not saying I can instantly program at expert level in any language I've never heard of. But yes, I can read most programming languages even without ever having seen them. Take LUA. I had never used LUA and when I started out I was reading code examples and writing LUA code without first "learning" it at all. I just went and did it. Did I probably do some "stupid things" that an expert would do differently and/or fast? Of course. But I got stuff done and learned as I went along. Same when we introduced Kotlin for example. I was instantly productive but someone using Kotlin for a longer time (such as today's me) would "cringe" at how I was writing Kotlin as if it was Java.
Notable exceptions would be languages that have such a weird syntax as Brainfuck ;)
SQL is a great example. I have not been using SQL directly for quite a long time now. But about 20 years ago my bread and butter was SQL and I was writing stored procedures every single working day and I was not writing any Java at that point in time but was using two other languages other than SQL on a daily basis as well (Perl and Bash).
But an unexpected change will come and you will need to go back studying on your free time. This is always happening in our industry, you were just lucky for not living it because you sticked with Java.
In professional capacity over the years I have used Perl, PHP, Bash, SQL (various dialects), Java, Groovy, Kotlin, Scala, Javascript and Typescript. I did not move from Java and OOP knowledge to other programming languages. I started with with languages that don't (have to) use OOP. I still work with Typescript (and until recently Javascript) while also working with Java and Kotlin. And in Java and Kotlin there is very little traditional OOP going on.
I guess the point there is to always keep learning and evolving. I learned Kotlin at work, while doing it. I learned Typescript at work, while doing it. I learned about various libraries at work, while using them. Until a few years ago, I had never used React and I have never used React except at work.
Talk about it. Prolifically. Say that you have no problem working on "that thing" but that it's going to take more time, because you're unfamiliar with that language / library / service or whatever it is that would slow you down.
Then adjust your estimates. If you are working on something completely new that you know nothing about, put in "Spike" tickets to prototype things. Put larger estimates for the tasks themselves after for the "unknown unknowns". When you discover an unknown unknown that can be extracted as a new ticket from what you are currently working on, create a ticket for it to make that fact visible and adjust your current ticket to say it got extracted so nobody still expects your ticket to do it.
Tickets are not evil, tickets are protection for developers - create them and comment on them, explaining what is going on and what needs to be done. Most companies aren't really "evil" in that way. They just don't know better. If you do happen to be in one of the 'actually evil' ones, go find another company to work for. E.g. if developer are not allowed by process or actually prevented by the ticketing software from just creating tickets and moving things around as necessary: run. Don't walk. Run from that company.
Tickets are useless. Real tickets are unit tests that need to pass for the software to be ready for production.
Tickets cannot protect you, only Odin's switch statement on stack trace can protect developers from bad changes.
I will give you an example, you worked for a company and on a ticket, you wrote “this code works this specific way and should never be changed”. After 6 months, the company fires you and another senior developer takes your place. The new developer never read previous tickets because no one read old completed tickets, so no one will read the message you left, and they will change the code accidentally.
However, in Odin with unit tests and locking stack traces with switch statements, the code is really protected from these kinds of accidents.
The “while doing it” does not work if you don't have experience with similar language. If you don't believe me, try learning Rust while writing a server for your work.
But I think I'm barking up the wrong tree here anyway, because you seem to be completely set in your thinking that nobody should need to know more than exactly one programming language and that should probably be Odin and that anyone thinking or saying anything else is just wrong. Well good luck to you getting paid and having fun. I definitely know that someone with your attitude towards learning and expanding ones horizon would not last long at my place.
only Odin's switch statement on stack trace can protect developers from bad changes.
on a ticket, you wrote “this code works this specific way and should never be changed”
What I was saying about tickets being protection is that they're protection against "the company" and process and being rushed all the time "because agile". Embrace them. Use them to your advantage. Someone gives you one ticket to do X and do it quickly but you don't know the library or service involved? Convert it into an Epic and extract 10 tickets including some prototyping in spikes.
Odin with unit tests
Making simple tickets to epic shows that the company does not have a software architecturer. It means that you create a monster application with full of technical debts left by junior. If CDD used in your company, nobody would ever be able to replace a ticket with an epic, because the architecturer would show you from where to start and what steps to take.
You thinking in DDD because in CDD there is no domain logic. Everything is part of stack trace tree
For some “can't” statements, you don't need to write any error.
The “can't” statements in CDD are for actions and input that you could take, but the software deliberately prevents you from doing them.
There are no actions or input, requested from the exercise that make it possible to do the list you gave me, so it does not make sense to create an error for them.
People who think that specifically TDD is a programming technique didn't get the entire memo. It can drive you to design and re-design your app's inputs. But that part of TDD is tacit - hard to teach.
where in the article say that CDD does not care about inputs?
Also in the article says "The main idea of TDD is to design the software through tests." do you disagree with that?
Not being able to quickly think of more test cases for whatever you're trying to test at any given moment does not prove anything. You also seem to be very hung up on something about stack traces that I'm really not able to figure out what it really is, even though you keep mentioning it.
It architects the code from the errors by following specific rules
or, #ifndef GONE is more powerful/flexible than #ifdef PRESENT
Of course, sometimes one wants exactly that specific things that positive assertions do. But IMO that's rarely the case, it's just going positive seems simpler/easier to think of.
one may say, in some system of rules, "Dont's" are more usable than "Do's". Think cultures..
From what I understood it seemed like you listed a collection of requirements for TDD failure cases, named them "can't" statements and then did TDD on them. But I believe there is more to it, but I couldn't see it.
Could you list the types test cases you would have created if you had done TDD on this same problem? And how they'd be different?
In CDD, I created first the software design by translating "can't" statements to errors and then iterate over the errors following specific rules to create a tree of stack traces that created the software design and then I TDD over that design to complete the application.