Huh? at least in web, the "big ones" (angular/react/vue/svelte) have been around for YEARS at this point and IMO mostly stablized (though i still don't understand why angular needs to releast a breaking change version every 6-12 months)
the major 'issue' is often near 0 understanding in the fundamentals i'm talking like, super basic, what is an onclick function, how do we get our website to talk to our backend. if you can have clean domain and abstraction cuts, the rest really is a 'technical detail' - i.e. language / framework / technologies truly are all subjective. there are probably combined over 1000 valid tech stacks for a standard "show me table entries in a web dashboard" - not one is 'more correct' or 'more wrong' than the other, but rather in a given _organization_ or with a _given set of devs_ etc. that makes it 'more wrong'. the tech doesn't care, its the team of humans and HOW that team of humans interacts with the stack that is where things can go off the rails
there are so many ways to skin a cat and there ARE tradeoffs (positive AND negative) to each of these tech decisions... i'm not sure what the author is getting at - he seems to hint there are a select few sets of known best practices and tech choices, but fails to list them explicitly... this is dubious at best, and in some way counteracts his claim that it is "incorrect" that tech choices are subjective.
if you want to be non-subjective, be objective. name exactly the tech decisions and best practices you are talking about!
That's very simple. The balanced path depends directly on how much of the requirements and assumptions are going to change during the life time of the thing you are building.
Engineering is helpful only to the extent you can forsee the future changes. Anything beyond that requires evolution.
You are able to comment on the complexity of that large company only because you are standing in the future into 50 years from when those things started take shape. If you were designing it 50 years back, you would end up with same complexity.
The nature's answer to it is, consolidate and compact. Everything that falls onto earth gets compacted into a solid rock over time, by a huge pressure of weight. All complexity and features are flattened out. Companies undergo similar dynamics driven by pressures over time, not by big-bang engineering design upfront.
>> If you ignore a dependency and try to fix it later, it will be more expensive. More time, more effort, more thinking. And it will require the same level of coordination that you tried to avoid initially.
The 1980s and 90s were full of DOD-497 multi-kilogram documents being analyzed atomically to determine the specification, and they rarely came in close any of the 3 main dimensions of success: time, quality, or cost.
On the other hand, neither has Agile with a capital A, with the ceremony of documents replaced with the ceremony of JIRA tickets and t-shirts.
2. I would be more receptive to this argument if they had listed some famous examples of successful, large systems that were built like this. On the other hand, I can easily list many failures: FAA Advanced Automation System (1980s), IRS Tax Systems Modernization (1990s), UK NHS National Programme for IT (2000s).
3. Waterfall vs. agile is a continuum. Nobody plans everything, down to each if-statement, and nobody wings it without some kind of planned architecture (even if just inside one person's head). Where you are on the continuum depends on the nature of the problem (are all requirements known?), the nature of the team (have they done this before?), and the criteria for success (are there lives depending on this?).
4. The analogy to building a building is flawed. At large enough scale, software is like a city, and all successful cities have gradually evolved in complexity. Come back to me when someone builds a 1-million person arcology on some island in the Pacific.
5. Just as some PhDs are sensitive about being called "Doctor", some software engineers are sensitive about being "real engineers". Stop thinking about that. What we do as software engineers is immensely valuable and literally changing the world (usually, but not always, for the better). Let's stop worrying about whether or not what we do is "engineering" and focus on what we do best: building complex systems that have never before existed on earth.
> you lay out a huge specification that would fully work through all of the complexity in advance, then build it.
Because requirements always change. Humans always change. That's it. No need to elaborate.
Incidentally this highlights a problem when using chatbots to build large software projects that are intended to be used for a long period of time.
The key is not how much code you can add but how little you can get away with.
Chatbots only solution ever is to ADD code. They're not good at NOT writing code or even deleting it because after all the training set for the lines of code that do not exist is an empty set. Therefore it's impossible to train a robot to not write code.
What's better than generating 10kloc really fast? Not having it in the first place.
In fact, this is how you build an aerospace program, satelite, and more.
It is even possible to develop software with agile processes in such a framework, even though strictly speaking it’s not fully agile.
Nah I’m good. I’ve watched system architecture framework views be developed. Years of prep and planning. System is released and half the employees that had requirements no longer work there and the business already pivoted to a new industry focus.
There’s a reason we went this way in software development a quarter century ago.
Software is not a skyscraper.
In short: the tension described in "systems thinking" is the same one as the one between "spec driven" vs. "iterative prompting"
In a more typical modern sense systems thinking is more about relationships and wholes, rather than isolating parts, which is the traditional engineering approach.
While much of the base material on systems thinking will be based around cybernetics, it is really a complement to traditional engineering, used in parallel to identity more natural complexity boundaries and to help avoid confusion and accidental complexity.
Gregor Hohpe’s Architect Elevator is probably a good place to start on why this change in perspective is important and why investing in flexibility is crucial when there is uncertainty.
While you may have to accept this article’s definition in some groups, accepting the more modern definition will help you get jobs in places that are nicer to work.
This type of false dichotomy that is presented in the article is a warning that there is soft work to be done.
People mentioning mechanical engineering in this thread are possibly the people who may benefit most from examining the material. I encourage you to see if this is a path forward for your needs.
> The most prevalent one, these days, is that you gradually evolve the complexity over time. You start small and keep adding to it.
> The other school is that you lay out a huge specification that would fully work through all of the complexity in advance, then build it.
I think AI will drive an interesting shift in how people build software. We'll see a move toward creating and iterating on specifications rather than implementations themselves.
In a sense, a specification is the most compact definition of your software possible. The knowledge density per "line" is much higher than in any programming language. This makes specifications easier to read, reason about, and iterate on—whether with AI or with peers.
I can imagine open source projects that will revolve entirely around specifications, not implementations. These specs could be discussed, with people contributing thoughts instead of pull requests. The more articulated the idea, the higher its chance of being "merged" into the working specification. For maintainers, reviewing "idea merge requests" and discussing them with AI assistants before updating the spec would be easier than reviewing code.
Specifications could be versioned just like software implementations, with running versions and stable releases. They could include addendums listing platform-specific caveats or library recommendations. With a good spec, developers could build their own tools in any language. One would be able to get a new version of the spec, diff it with the current one and ask AI to implement the difference or discuss what is needed for you personally and what is not. Similarly, It would be easier to "patch" the specification with your own requirements than to modify ready-made software.
Interesting times.
The core point they're trying to make is that agile (or similar) practices are the incorrect way to approach consolidation of smaller systems into bigger ones when the overall system already works and is very large.
I agree with their assertion that being forced to address difficult problems earlier on in the process results in ultimately better outcomes, but I think it ignores the reality that properly planning a re-write of monumentally sized and already in use system is practically impossible.
It takes a long time (years?) to understand and plan all the essential details, but in the interim the systems you're wanting to rewrite are evolving and some parts of the plan you thought you had completed are no longer correct. In essence, the goal posts keep shifting.
In this light, strangler fig pattern is probably the pragmatic approach for many of these re-writes. It's impossible to understand everything up front, so understand what you reasonably can for now, act on that, deliver something that works and adds value, then rinse and repeat. The problem is that for sufficiently large system, this will take decades and few software architects stick around at a single company long enough to see it through.
A final remark I want to make is that, after only a few years of being a full-time software developer, "writing code" is one of the easiest parts of the job. The hard part is knowing what code needs to be written, this requires skills in effective communication with various people, including other software developers and (probably more importantly) non-technical people who understand how the business processes actually need to work. If you want to be a great software developer, learn how to be good at this.
> you lay out a huge specification that would fully work through all of the complexity in advance, then build it.
But I do still think there's a lot of value into coming up with a good plan before jumping in. A lot of software people like to jump in and I see them portray the planning people as trying to figure everything out first. (I wonder if we reinforce the jumping in head first mentality because people figure out you can't plan everything) A good plan helps you prevent changing specs and prepares you for hiccups. It helps by having others but basically all you do is try to think of all the things that could go wrong. Write them down. Triage. If needed, elevate questions to the decision makers. Try a few small scale tests. Then build out. But building out you're always going to find things you didn't see. You can't plan forever because you'll never solve the unknown unknowns until you build, but also good prep makes for smoother processes. It's the reason engineers do the math before they build a bridge. Not because the math is a perfect representation and things won't change (despite common belief, it's not static) but because the plan is cheaper than the build and having a plan allows you to better track changes and helps you determine how off the rails you've gone.
It is also perplexing to me that people think they can just plan everything out and give it to a LLMs. Do you really believe your manager knows everything that needs to be done when they assign jobs to you? Of course not, they couldn't. Half the job is figuring out what the actual requirements are.
> the size of the iterations matters, a whole lot. If they are tiny, it is because you are blindly stumbling forward. If you are not blindly stumbling forward, they should be longer, as it is more effective.
You are not blindly stumbling forward, you're moving from (working software + tiny change) to (working software including change). And repeat. If there's a problem, you learn about it immediately. To me that's the opposite of moving blindly.
> you really should stop and take stock after each iteration.
Who is not taking stock after every iteration? This is one of the fundamental principles of agile/lean/devops/XP/scrum. This one sentence drastically lowers my impression of the author's ability to comment on the subject.
> The faster people code, the more cleanup that is required. The longer you avoid cleaning it up, the worse it gets, on basically an exponential scale.
Unsafe tempo is as likely to happen in big-spec design projects as in small iterations. In fact, working in careful small iterations helps us manage a realistic tempo because we know we can't move faster than we can get things into production and evaluate.
The terrible outcomes listed in the same paragraph are linked to unwise practice and have nothing to do with small iteration size.
There are core differences in software engineering that, unlike in construction work:
- making changes is often cheaper
- we might not know beforehand everything that is needed to be built, especially unknown unknowns
I would still agree that the truth is somewhere in between, but I would argue that, for software, it's closer to the evolutionary approach.
The 1970 Royce paper was about how waterfall didn't work, and most "Agile" is a subset of DSDM, each flavor missing a necessary thing or two whether working in large systems or growing them greenfield from nothing. But DSDM wasn't "little a" agile (and SAFE just isn't). There is a middle way.
If you like applying this stuff (e.g. you've chatted with Gene Kim, follow Will Larsen, whatever, sure, but you've deliberately iterated your approach based on culture and outcome observability), feel free to drop me a note to user at Google's thing.
- you create large number of working small apps .
- you create a spec from these apps .
- create a huge app .
- make a Dsl to make extensible .
- extend the Dsl to fit what you need in the future .
- optimize the Dsl remove obvious N+1 stuff.
The hard part is throwing away the code in each step . Both managment and devs cant stomach the reality that the code is useless at each stage prior to dsl. They cant molt and discard the shell and hence the project dies.
That feels like a straw man to me. This is not a binary question. For each small design decision you have a choice about how much uncertainty you accept.
There are no "two schools". There is at least a spectrum between two extremes and no real project was ever at either of the very ends of it. Actually, I don't think spectrum is a proper word even because this is not just a single dimension. For example, speed and risk often correlate but they are also somewhat independent and sometimes they anti-correlate.
It didn't work, even for the "large" systems of that time: and Booch had worked on more than a few. The kind of "system" the OP is describing is vastly larger, and vastly more complex. Even if you could successfully apply the waterfall model to a system built over two or three years, you certainly can't for a system of systems built over 50 years: the needs of the enterprise are evolving, the software environment is evolving, the hardware platform is evolving.
What you can do, if you're willing to pay for it, is ruthlessly attack technical debt across your system of systems as a disciplined, on-going activity. Good luck with that.
Comments on actual blog post: 0
Why are people so afraid to leave replies on the author’s OC
Well, there isn't a formal version of it, because the answer is not formal, it is cultural.
In enterprise software, you have an inherent tension between good software engineering culture, where you follow the Boy Scouts' principle of leaving a codebase cleaner than you found it, and SOC2 compliance requirements that expect every software change to be tracked and approved.
If every kind of clean up requires a ticket, that has to be exhaustively filled out, then wait for a prioritization meeting, then wait for the managers and the bean counters to hem and haw while they contemplate whether or not it's worth it to spend man-hours on non-functional work, then if you're lucky then decide you can spend some time on it three weeks from now and if you're unlucky they decide nope, you gotta learn to work within an imperfect system. After once or twice or trying to work By The Book, most engineers with an ounce of self-respect will decide "fuck it, clearly The System doesn't care," and those with two ounces of self-respect will look for work elsewhere.
Or, you get together with the members of your team and decide, you know what, the program managers and the bean counters, they're not reading any of the code, not doing any of the reviews, and they have no idea how any of this works anyway. So you collectively decide to treat technical debt as the internal concern that it anyway was in the first place - you take an extra half hour, an extra day, however long it takes to put in the extra cleaning or polish, and just tack it on to an existing ticket. You give a little wink and you get a little nod and you help the gears turn a little more smoothly, which is all the stakeholders actually care about anyway.
You cannot replace culture with process. All attempts to replace culture with process will fail. People are not interchangeable cogs in the machine. If you try to treat them as such, they will gum up and get stuck. Ownership and autonomy are the grease that allows the human flywheel to spin freely. That means allowing people to say, "I'm going to do this because I think that it is Right And Good For My System Which I Own", and allowing them to be responsible for the consequences. To pass SOC2, that means treating people like adults and allowing them to sometimes say, instead of "can I get this reviewed because I legit need another set of eyes to take a serious look?", to say "can I get a quick rubber-stamp on this please?"
> The most prevalent one, these days, is that you gradually evolve the complexity over time. You start small and keep adding to it.
> The other school is that you lay out a huge specification that would fully work through all of the complexity in advance, then build it.
I doubt many people (if they stop to think about it) actually fit into either of these two schools of thought, they're both extremist positions. It's like claiming that the US population is filled with people believing in either big government autocracy and communism or small government libertarianism and free markets. That's an absurd position to take, just as these opening paragraphs are absurd.
The most generous interpretation is that the author is setting up a strawman.
These are extremist positions. No one except a fool would think that they can design a truly large system from scratch before ever writing a piece of code, and no one but a fool thinks they can write the code for a large system without ever thinking about the design.
The reality is that it sits in between, the author figures this out by the end, fortunately, but wants something they can't have:
> There should be some balanced path in the middle somewhere, but I haven’t stumbled across a formal version of it after all these decades.
First, they at least admit that they're stumbling. That's good, groping around in the dark is not an effective way to find an answer, turn on the lights. You aren't the only one thinking about this subject.
Second, for small projects and simple projects, or rehashes of projects you've done before, the approach often doesn't matter. This only matters for large, complex, and/or novel projects.
Software development is a design process.
Again, in case this was missed: Software development is a design process.
The idea of separating design from development is foolish (BDUF). The idea of separating development from design is equally foolish (extreme take on Agile, see Ron Jeffries failing at sudoku because he tries to use a development technique, TDD, without thinking about the design).
Take the "over 3000 active systems" from paragraph 4. There is no way anyone could have designed all 3000 active systems (either as the 3k systems or in a compressed form) from scratch in a reasonable amount of time. The only reason the author can think of a better design is because they have a design, even if it's not formally documented. The existing software is the design [0] that they can draw from to come up with the better design.
But wait, the foolish BDUF people would not try to refine the system, they'd try and build a new system from scratch. Don't be a fool.
The foolish extreme Agile people would not look at the whole (or a large enough section) and think about refining it, they'd just add to it or change the existing systems.
The sensible person says, "Wait, I acn'
IMO the problem isn't discussing the spec per se. It's that the spec doesn't talk back the way actual working code does. On a "big upfront design" project, there is a high chance you're spending a lot of time on moot issues and irrelevant features.
Making a good spec is much harder than making working software, because the spec may not be right AND the spec may not describe the right thing.
When they turn out to be based on false assumptions of simplicity the fallout is that the whole thing can't go forward because of one of the details.
Evolutionary systems at least always work to some degree even if you can look after the fact and decide that there's a lot of redundancy. Ideally you would then refactor the most troublesome pieces.
Ai assistance would seem to favor the engineering approach as the friction of teams and personalities is reduced in favor of quick feasibility testing and complete planning.
There is a related phenomenon in some types of software where the cost of building an operational prototype asymptotically converges on the cost of just writing the production code. (This is always a fun one to explain to management that think building a prototype massively reduces delivery risk.)
It was discussed here just 2 days ago intensively.
Software has 0 construction cost but that it does have is extremely complicated behavior.
Take a bridge for example: the use case is being able to walk or drive or ride a train across it. It essentially proves a surface to travel on. The complications of providing this depend on the terrain, length etc etc and are not to be dismissed but there's relatively little doubt about what a bridge is expected to do. We don't iterate bridge design because we don't need to know much from the users of the bridge: does it fulfill their needs, is it "easy to use" etc AND because construction of a bridge is extremely expensive so iteration is also incredibly costly. We do, however, not build all bridges the same and people develop styles over time which they repeat for successive bridges and we iterate that way.
In essence, cycling is about discovering more accurately what is wanted because it is so often the case that we don't know precisely at the start. It allows one to be far more efficient because one changes the requirements as one learns.
Those are called data/structures and design patterns (not only the ones in the GoF book). If you have a good understanding of those and you know your data and the operation you will apply to it, it’s easy to model your data using those structures. Software are state machines specifications. Knowing how to model states and thus figuring the transitions is helpful. And there’s a lot of samples out there.
Everything W3C does. Go is evolving through specs first. Probably every other programming language these days.
People already do that for humankind-scale projects where there have to be multiple implementations that can talk to each other. Iteration is inevitable for anything that gains traction, but it still can be iteration on specs first rather than on code.
Designers need malleability, that is why they all want digital design systems.
But if anything, all development is the search for the search for the requirements. Some just value writing them down.
If you are doing a CRUD web app for a local small business - there are thousands of examples. If you are writing control software for a space station - you may not have access to code from NASA/Russia/China but you can at least look at generic software that does the things you need and learn some lessons.
In this sense, web applications haven't changed so much in the last twenty years: client, server, database...
Some projects have been forced so far, by diverting resources (either public-funded or not-yet-profitable VC money), but these efforts have not proven to be self-sustaining. Humans will be perpetually stuck where we are as a species if we cannot integrate the currently opposing ideas of up-front planning vs. move fast and break things.
Society is slowly realizing the step-change in difficulty between projects in controlled conditions that can have simplified models to these irreducibly complex systems. Western doctors are facing an interesting parallel, now becoming more aware to treat human beings in the same way--that we emerge as a result of parts which can be simplified and understood, but could never describe the overall system behavior. We are good examples of the intrinsic fault-tolerance required for such systems to remain stable.
See also: "there is nothing more permanent than a temporary solution"
The law is maybe a little too simplistic in its formulation, but it's fundamentally true.
The first is too ambitious and ends in an unmaintainable pile around a good core idea.
The second tries to "get everything right" and suffers second system syndrome.
The third gets it right but now for a bunch of central business needs. You learned after all. It is good exactly because it does not try to get _everything_ right like the second did.
The fourth patches up some more features to scoop up B and C prios and calls it a day.
Sometimes, often in BigCorp: Creators move on and it will slowly deteriorate from being maintenaned...
Care to exemplify?
This is a really good observation and I predict you will be correct.
There is a consequence of this for SaaS. You can imagine an example SaaS that one might need to vibecode to save money. The reason its not possible now is not because Claude can't do it, its because getting the right specs (like you suggested) is hard work. A well written spec will not only contain the best practices for that domain of software but also all the legal compliance BS that comes along with it.
With a proper specification that is also modular, I imagine we will be able to see more vibecoded SaaS.
Overall I think your prediction is really strong.
> The WHATWG was based on several core principles, (..) and that specifications need to be detailed enough that implementations can achieve complete interoperability without reverse-engineering each other.
But in my experience you need more than a spec, because an implementation is not just something that implements a spec, it is also the result of making many architectural choices in how the spec is implemented.
Also even with detailed specs AI still needs additional guidance. For example couple of weeks ago Cursor unleashed thousands of agents with access to web standards and the shared WPT test suite: the result was total nonsense.
So the future might rather be like a Russian doll of specs: start with a high-level system description, and then support it with finer-grained specs of parts of the system. This could go down all the way to the code itself: existing architectural patterns provide a spec for how to code a feature that is just a variation of such a pattern. Then whenever your system needs to do something new, you have to provide the code patterns for it. The AI is then relegated to its strength: applying existing patterns.
TLA+ has a concept of refinement, which is kind of what I described above as Russian dolls but only applied to TLA+ specs.
Here is a quote that describes the idea:
There is no fundamental distinction between specifications and implementations. We simply have specifications, some of which implement other specifications. A Java program can be viewed as a specification of a JVM (Java Virtual Machine) program, which can be viewed as a specification of an assembly language program, which can be viewed as a specification of an execution of the computer's machine instructions, which can be viewed as a specification of an execution of its register-transfer level design, and so on.
Source: https://cseweb.ucsd.edu/classes/sp05/cse128/ (chapter 1, last page)
One issue is that a spec without a working reference implementation is essentially the same as a pull request that's never been successfully compiled. Generalization is good but you can't get away from actually doing the thing at the end of the day.
I've run into this issue with C++ templates before. Throw a type at a template that it hasn't previously been tested with and it can fall apart in new and exciting ways.
I highly applaud this idea. IMO this is why big upfront design is so risky.
I discussed some of this in https://www.ebiester.com/agile/2023/04/22/what-agile-alterna... and it gives a little bit of history of the methods.
We are nearly 70 years into this discussion at this point. I'm sure Grace Hopper and John Mauchly were having discussions about this around UNIVAC programs.
> But I do still think there's a lot of value into coming up with a good plan before jumping in.
Definitely, with emphasis on a _good_ plan. Most "plans" are bad and don't deserve that name.
> be specified up-front, planned on JIRA
Making a plan up-front is a good approach. A specification should be part of that plan. One should be ready to adapt it when needed during execution, but one should also strive to make the spec good enough to avoid changing.
HOWEVER, the "up-front specification" you mentioned was likely written _before_ making a plan, which is a bad approach. It was probably written as part of something that was called "planning" and has nothing to do with actual planning. In that case, the spec is pure fiction.
> estimates provided
Unless this project is exceptional, the estimates are probably fiction too.
> and Gantt charts setup
Gantt charts are a model, not a plan. Modeling is good; it gives you insight into the project. But a model should not be confused with a plan. It is just one tiny fragment you need to build a plan, and Gantt charts are just one of many many many types of models needed to build a plan.
> before they even sign the contract for the next milestone
That's a good thing. Signing a contract is an irreversible decision. The only contract that should be signed before planning is done is the contract that employs the planners.
> Anyone who claims upfront specs are the solution
See bove. A rigid upfront spec is usually not a plan, but pure fiction.
> My approach, especially for a project with a lot of unknowns, is usually to jump in right away and try to build a prototype.
Whether this is called planning or "jumping in" is a difference in terminology, not in the approach. The relevant clue is that you are experimenting with the problem to understand it, but you are NOT making irreversible decisions. By the terminology used in that book, you are _planning_, not _executing_.
> after the 2000 pages specification document was written, and passed down from the architects to the devs
If the 2000 page spec has never been passed to the devs while writing it, it's not part of a plan, it's pure fiction. Trying to develop software against that spec is part of planning.
My best project to date was a largely waterfall one - there was somewhere around 50-60 pages of A4 specs, a lot of which I helped the clients engineer. As with all plans, a lot of it changed during implementation, actually I figured out a way of implementing the same functionality, but automating it to a degree where about 15 of those could be cut out.
Furthermore, it was immensely useful because by the time I actually started writing code, most of the questions that needed answers and would alter how it should be developed had already come up and could be resolved, in addition to me already knowing about some edge cases (at least when it came to how the domain translates into technology) and how the overall thing should work and look.
Contrast that to some cases where you're just asked to join a project and help out and you jump into the middle of ongoing development, not going that much about any given system or the various things that the team has been focusing on in the past few weeks or months.
> It’s not hard to see that if they had a few really big systems, then a great number of their problems would disappear. The inconsistencies between data, security, operations, quality, and access were huge across all of those disconnected projects. Some systems were up-to-date, some were ancient. Some worked well, some were barely functional. With way fewer systems, a lot of these self-inflicted problems would just go away.
Also this reminds me of https://calpaterson.com/bank-python.html
In particular, this bit:
> Barbara has multiple "rings", or namespaces, but the default ring is more or less a single, global, object database for the entire bank. From the default ring you can pull out trade data, instrument data (as above), market data and so on. A huge fraction, the majority, of data used day-to-day comes out of Barbara.
> Applications also commonly store their internal state in Barbara - writing dataclasses straight in and out with only very simple locking and transactions (if any). There is no filesystem available to Minerva scripts and the little bits of data that scripts pick up has to be put into Barbara.
I know that we might normally think that fewer systems might mean something along the lines of fewer microservices and more monoliths, but it was so very interesting to read about a case of it being taken to the max - "Oh yeah, this system is our distributed database, file storage, source code manager, CI/CD environment, as well as web server. Oh, and there's also a proprietary IDE."
But no matter the project or system, I think being able to fit all of it in your head (at least on a conceptual level) is immensely helpful, the same way how having a more complete plan ahead of time can be helpful with a wide variety of assumptions vs "we'll decide in the next sprint".
And by doing this sort of exercise, you can avoid wasting time on dead ends, bad design, and directionless implementation. It's okay if requirements change or you discover something later on that requires rethinking. The point is to make your thinking more robust. You can always amend a design document and fill in relevant details later.
Furthermore, a mature design begins with the assumption that requirements (whether actual requirements or knowledge of them) may change. That will inform a design where you don't paint yourself into a corner, that is flexible enough to be adapted (naturally, if requirements change too dramatically, then we're not really talking about adaptation of a product, but a whole new product).
How much upfront design work you should do will depend on the project, of course. So there's a middle way between the caricature of waterfall and the caricature of agile.
A system of services that interact, where many of them are depending on each other in informal ways may be a complex system. Especially if humans are also involved.
Such a system is not something you design. You just happen to find yourself in it. Like the road to hell, the road to a complex system is paved with good intentions.
We have yet to see a largely llm driven language implementation, but it is surely possible. I imagine it would be easier to tell the llm to instead translate the Java implementation to whatever language you need. A vibe-coded language could do major damage to a companies data.
[0] https://iceberg.apache.org/spec/ [1] https://lists.apache.org/thread/whbgoc325o99vm4b599f0g1owhgw...
My approach, especially for a project with a lot of unknowns, is usually to jump in right away and try to build a prototype. Then iterate a few times. If it's a small enough thing, a few iterations is enough to have a good result.
If it's something bigger, this is the point where it's worth doing some planning, as many of the problems have already been surfaced, and the problem is much better understood.
> This has never happened and never will. You simply are not omniscient. Even if you're smart enough to figure everything out the requirements will change underneath you.
I am one of those "battle-scarred twenty-year+ vets" mentioned in the article, currently working on a large project for a multinational company that requires everything to be specified up-front, planned on JIRA, estimates provided and Gantt charts setup before they even sign the contract for the next milestone.
I've worked on this project for 18 months, and I can count on zero hands the times a milestone hasn't gone off the rails due to unforeseen problems, last-minute changes and incomplete specifications. It has been an growing headache for the engineers that have to deliver within these rigid structures, and it's now got to the point that management itself has noticed and is trying to convince the big bosses we need a more agile and iterative approach.
Anyone who claims upfront specs are the solution to all the complexity of software either has no real world experience, or is so far removed from actual engineering they just don't know what they're talking about.
indeed, i would argue 'big iterations' are the ones where all the problems which the author mentions crop up in the first place!
I suppose it's primarily a matter of experience. And as the article alludes, it's very important to deeply understand the subject matter. I highly value some of my non-programmer colleagues responsible for documentation. But can't put my finger on what exactly they brought to table that made their prose exceptionally good (clear, concise, spot on)...
Splitting coding and design is a bad idea. It’s like asking engineers not to draw and measure.
How rapidly has business software changed since COVID? Yet how many skyscrapers remain partially unoccupied in big cities like London, because of the recent arrival of widespread hybrid working?
The buildings are structurally unchanged and haven't been demolished to make way for buildings that better support hybrid working. Sure office fit outs are more oriented towards smaller simultaneous attendance with more hot desking. Also a new industry boom around team building socials has arrived. Virtual skeet shooting or golf, for example.
On the whole, engineered cities are unchanged, their ancient and rigid specifications lacking the foresight to include the requirements that accommodate hybrid working. Software meanwhile has adapted and as the OP says, evolved.
Point is though eventually some system runs out of ability. It works different in programming from physical construction, but the concept is the same, eventually you can't make a bad early design work anymore.
It's sort of the old General Eisenhower quote: "In preparing for battle I have always found that plans are useless, but planning is indispensable."
Of course, it requires some discipline to not just yolo the prototype into production when that’s not appropriate.
Nothing will get you to hit every milestone. However you can make progress if you have years of experience in that project and the company is willing to invest in the needed time to make things better (they rarely are)
You need smaller documents - this is the core technology we are using. This is how one subsystem is designed - often this should be on a whiteboard because once you get into the implementation details you need to change the plan, but the planning was useful. This is how to use core parts of the system so new comers can start working quick.
You need disciple to accept that sometimes libfoo is the best way to solve a problem in isolation, but since libbar is used elsewhere and can solve the problem your local problem will use libbar despite making your local problem uglier. Have a small set of core technologies that everyone knows and uses is sometimes more valuable than using the best tool for the job - but only sometimes.
If the definition of "complex" is instead something more like "a system of services that interact", "prone to multiple, coincidental failures", then I don't think it's impossible to design them. It's just very hard. Manufacturing lines would be examples, they are certainly designed.
Not sure why you're trying to bring AI development into this.
And things like "race conditions"/lack of scalability due to improper threading architecture aren't especially easy to fix(!)..
That’s certainly my experience
The design of the manufacturing lines and the resulting supply chain are not independent of each other -- you can trace features from one to the other -- but you cannot take apart the supply chain and analyze the designs of its constituent manufacturing lines and actually predict the behavior of the larger system.
AFAIK there's not a great definition of a complex system, just a set of traits that tend to indicate you're looking at one. Non-linearity, feedbacks, lack of predictability, resistance to analysis (the "you can't take it apart to reason about the whole" characteristic mentioned above"). All of these traits are also kind of the same things... they tend to come bundled with one another.
(And no, this is not "my" definition, it's how it's defined in the systems-related disciplines.)
Also, there's a certain point where you can't avoid management sabotaging things.
The set of system designs that exhibit naturally stable behavior doesn't overlap much with the set of system designs that deliver maximum performance and efficiency. The capability gap between the two can be large but most people choose easy/simple.
There is an enormous amount of low-hanging opportunity here but most people, including engineers, struggle with systems thinking.
There are two main schools of thought in software development about how to build really big, complicated stuff.
The most prevalent one, these days, is that you gradually evolve the complexity over time. You start small and keep adding to it.
The other school is that you lay out a huge specification that would fully work through all of the complexity in advance, then build it.
In a sense, it is the difference between the way an entrepreneur might approach doing a startup versus how we build modern skyscrapers. Evolution versus Engineering.
I was working in a large company a while ago, and I stumbled on the fact that they had well over 3000 active systems that were covering dozens of lines of business and all of the internal departments. It had evolved this way over fifty years, and included lots of different tech stacks, as well as countless vendors. Viewed as ‘one’ thing it was a pretty shaky house of cards.
It’s not hard to see that if they had a few really big systems, then a great number of their problems would disappear. The inconsistencies between data, security, operations, quality, and access were huge across all of those disconnected projects. Some systems were up-to-date, some were ancient. Some worked well, some were barely functional. With way fewer systems, a lot of these self-inflicted problems would just go away.
It’s not that you could cut the combined complexity in half, but more likely that you could bring it down to at least one-tenth of what it is today, if not even better. It would function better, be more reliable, and would be far more resilient to change. It would likely cost far less and require fewer employees as well. All sorts of ugly problems that they have now would just not exist.
The core difference between the different schools really centers around how to deal with dependencies.
If you had thousands of little blobs of complexity that were all entirely independent, then getting finished is just a matter of banging out each one by itself until they are all completed. That’s the dream.
But in practice, very few things in a big ecosystem are actually independent. That’s the problem.
If you are going to evolve a system, then you ignore these dependencies. Sort them out afterwards, as the complexity grows. It’s faster, and you can get started right away.
If you were going to design a big system, then these dependencies dictate that design. You have to go through each one and understand them all right away. They change everything from the architecture all the way down to the idioms and style in the code.
But that means that all of the people working to build up this big system have to interact with each other. Coordinate and communicate. That is a lot of friction that management and the programmers don’t want. They tend to feel like it would all get done faster if they could just go off on their own. And it will, in the short-term.
If you ignore a dependency and try to fix it later, it will be more expensive. More time, more effort, more thinking. And it will require the same level of coordination that you tried to avoid initially. Slightly worse, in that the time pressures of doing it correctly generally give way to just getting it done quickly, which pumps up the overall artificial complexity. The more hacks you throw at it, the more hacks you will need to hold it together. It spirals out of control. You lose big in the long-term.
One of the big speed bumps preventing big up-front designs is a general lack of knowledge. Since the foundations like tech stacks, frameworks, and libraries are always changing rapidly these days, there are few accepted best practices, and most issues are incorrectly believed to be subjective. They’re not, of course, but it takes a lot of repeated experience to see that.
The career path of most application programmers is fairly short. In most enterprises, the majority have five years or less of real in-depth experience, and battle-scared twenty-year+ vets are rare. Mostly, these novices are struggling through early career experiences, not ready yet to deal with the unbounded, massive complexity present in a big design.
Also, the other side of it is that evolutionary projects are just more fun. I’ve preferred them. You’re not loaded down with all those messy dependencies. Way fewer meetings, so you can just get into the work and see how it goes. Endlessly arguing about fiddly details in a giant spec is draining, made worse if the experience around you is weak.
Evolutionary projects go very badly sometimes. The larger they grow, the more likely they will derail. And the fun gives way to really bad stress. That severe last-minute panic that comes from knowing that the code doesn't really work as it should, and probably never will. And the longer-term dissatisfaction of having done all that work to ultimately just contribute to the problem, not actually fix it.
Big up-front designs are often better from a stress perspective. A little slow to start and sometimes slow in the middle, they mostly smooth out the overall development process. You’ve got a lot of work to do, but you’ve also got enough time to do it correctly. So you grind through it, piece by piece, being as attentive to the details as possible. Along the way, you actively look for smarter approaches to compress the work. Reuse, for instance, can shave a ton of code off the table, cut down on testing, and provide stronger certainty that the code will do the right thing in production.
The fear that big projects will end up producing the wrong thing is often overstated. It’s true for a startup, but entirely untrue for some large business application for a market that’s been around forever. You don’t need to burn a lot of extra time, breaking the work up into tiny fragments, unless you really don’t have a clue what you are building. If you're replacing some other existing system, not only do you have a clue, you usually have a really solid long-term roadmap. Replace the original work and fix its deficiencies.
There should be some balanced path in the middle somewhere, but I haven’t stumbled across a formal version of it after all these decades.
We could go first to the dependencies, then come up with reasons why they can be temporarily ignored. You can evolve the next release, but still have a vague big design as a long-term plan. You can refactor the design as you come across new, unexpected dependencies. Change your mind, over and over again, to try to get the evolved works to converge on a solid grand design. Start fast, slow right down, speed up, slow down again, and so forth. The goal is one big giant system to rule them all, but it may just take a while to get there.
The other point is that the size of the iterations matters, a whole lot. If they are tiny, it is because you are blindly stumbling forward. If you are not blindly stumbling forward, they should be longer, as it is more effective. They don’t have to all be the same size. And you really should stop and take stock after each iteration. The faster people code, the more cleanup that is required. The longer you avoid cleaning it up, the worse it gets, on basically an exponential scale. If you run forward like crazy and never stop, the working environment will be such a swamp that it will all grind to an abrupt stop. This is true in building anything, or even cooking in a restaurant. Speed is a tradeoff.
Evolution is the way to avoid getting bogged down in engineering, but engineering is the way to ensure that the thing you build really does what it is supposed to do. Engineering is slow, but spinning way out of control is a heck of a lot slower. Evolution is obviously more dynamic, but it is also more chaotic, and you have to continually accept that you’ve gone down a bad path and need to backtrack. That is hard to admit sometimes. For most systems, there are parts that really need to be engineered, and parts that can just be allowed to evolve. The more random the evolutionary path, the more stuff you need to throw away and redo. Wobbling is always expensive. Nature gets away with this by having millions of species, but we really only have one development project, so it isn’t particularly convenient.