Understand, Design, Build: A Framework for Problem-Solving | Hacker News Reader

Understand, Design, Build: A Framework for Problem-Solving

lob.com

209 pointsharryzhang25 comments

Reminds me of this incredible chart / graph which you can use to apply to a lot of situations. I actively refer to this and show it to teams constantly. The gist of it is that you require many skills and capabilities to create alignment. Without that, you have different outcomes which we may initially experience as frustration / failure.

https://intenseminimalism.com/2015/a-framework-for-thinking-...

thanks for sharing, pretty sure that is a post I will come back to often

There's a methodology of teaching programming focused on understanding and designing programs.

It's in the book (available online for free) called How to Design Programs: https://www.htdp.org

It introduces the notion of a "design recipe" for structuring thinking and problem-solving for programming.

This framework is good. And like the other frameworks in other comments here, it's a good enough start in a small company, start up, or small department within a larger organisation.

A better framework might be cobbled together from various disciplines. For instance, "Understand". This part is huge. Well, bigger than those three questions the author lists.

"Understanding" begins by stepping back to look at the environment as a whole. Bound the situation. Assess organisational culture, the environment (immutable aspects of a transformation/process), the organisation's technology landscape (Active Directory?), regulatory requirements, competition, suppliers and customers.

There's also the social system (roles, norms, values), and the political landscape.

Then you define your objective. But before you can do that you need to know who your client is. Sure, in SV that's often easy. Outside that bubble not so much. Identify the client by asking some questions - whom do you need to satisfy? Who will judge the success of the project? Who will fail if the project fails? Who has authority for the project? Who pays for the project? Whom do you report to?

After all that you're ready to set an objective. Vision/scope stuff. Specific, measurable, attainable, realistic and time-boxed. Cliche'd, but when a project goes wrong the specification is the only version of the truth a team has to fall back on.

Setting project objectives is much bigger than I can do justice to here. However once you have that you can start with the fun stuff. Design, build, deploy, repeat until done.

However while you're doing the fun stuff you still have to do the unfun stuff too - risk management, and stakeholder management. And it's stupefyingly incredible how many people think they know how to do these things but can't readily explain the difference between mitigation and contingency, exposure and impact, or why you have to be nice to the CIO for your project to succeed, even though the CTO and CEO are on your side.

I've written about all this stuff. https://www.wittenburg.co.uk/Entry.aspx?id=8ec91ced-b3a4-4b0...

IMHO the "understanding" part belongs to the "skills".

Social skills, communication skills is just as important as solving practical problems. This skill can be trained and often underrated.

I think you've just nicely summarised why architects & project managers exist! There's a whole level of treacle to get through before coding can start, and that stays with the project throughout its lifespan.

The challenge I see most is getting the right level of understanding and design freedom to the developer level, where they know why they're being asked to do some work but without burdening them with extraneous detail or processes.

A lot of developers get annoyed at project managers and architects but they dislike project politics even more.

"A lot of developers get annoyed at project managers and architects..."

If a PM or architect is annoying developers then there's a very big problem with the PM or architect. Like any profession you get good and bad apples. It's why anonymous feedback from people who manage you and whom you manage is so useful. Of course ego is prevalent the higher up the chain you go so this is something many organizations won't do.

Yeah, usually the hardest problem is to define the problem, or even know that there exists a problem that could be solved. Once you have discovered, understood and defined your problem, solving it becomes easy.

Problem identification > problem solving

That's why "vision" is important.

I've seen too many companies tries to be the "next Apple" or "Next Google". I mean come on that's not a vision. Google's vision was to organize human knowledge. That's a god damn good one.

A more thorough framework to solve problems might come from being familiar with all of the available problem solving techniques in your domain. For instance, if your job is software development, you may want to know all of the problem solving techniques in computer science. Now is the point that I will be a shill for my own book: https://www.manning.com/books/classic-computer-science-probl...

Your book in swift with the same title is awesome as well.

Thanks for the shoutout!

> common misconception among new software professionals that the best engineers are the best at writing code

> Our job is to find and solve problems that move the business forward.

Couldn't agree more. Would also add that this "best dev==best at coding" ain't only restricted to the young ones. Never stop asking theses "why" questions that sometimes show us that crappy and cheap systems may work wonders ;)

These are very important steps to creating something, but there's not really anything new here.

This is basically a simplified version of the design thinking framework with the implied "build" step at the end made explicit.

For example, the process that has been taught at Stanford for years is: empathize, define, ideate, prototype, test. The first two steps could map to "understand" here with the last three falling among "design" and "build" depending on the fidelity of your prototype.

PDF link to an overview from Stanford: https://dschool-old.stanford.edu/sandbox/groups/designresour...

This is an adaptation of Design Thinking which designers have used for years. It's found in Human centered Design, Design Sprints, Jobs To Be Done... and others.

I've seen lots of projects fail, cost way over budget, take years longer than projected (personal recorded: I paid a developer to write me - his estimate - a 6 week project. 8 years later he returned my money), and/or completely miss the target because Dev jumps right in and starts coding - not being concerned with the "soft" stuff like business, usability, experience design, etc...

If this helps dev teams join the holistic effort to success, I'm all for it. I entirely agree that coders need to be problem solvers but also solvers of relevant problems.

Good talk on this subject:

Hammock Driven Development https://youtu.be/f84n5oFoZBc

"But our job isn’t to write code. Our job is to find and solve problems that move the business forward. Writing great code is a necessary but insufficient skill for doing that job."

If only more / most people understood this. Coding has a finite use. Problem solving (and critical thinking) are universal and possible uses approach infinity. They are the equivalent of the proverbial "learning to fish."

Well written content. However, I am not sure if thoroughly complete. The whole article covers areas of shortcomings related to understanding the problem, designing the structure and building it. But fails to address "upgrading current solution" or "optimizing the current solutions" that suits the current market needs.

> However, I am not sure if thoroughly complete.

Of course a 1,000 word web article is not going to be complete. There are dozens of books you can check out for more info, including from Tim Brown CEO of IDEO [1] and from Columbia Business School

[1] https://www.amazon.com/Change-Design-Transforms-Organization... [2] https://www.amazon.com/Solving-Problems-Design-Thinking-Publ...

Is... is this waterfall?

Essentially, yes. I find it implausible the PMs have a deep enough understanding of a product to be able to elicit technical requirements in the beginning without proof from a prototype or UX research. I personally prefer a Lean UX or Just in Time process where understanding, design, and build develop in smaller increments.

I don't know about this author's intention, but in general with design thinking frameworks, the idea is not that each step must be followed rigidly in order and never re-visited. It's more about calling attention to the importance of each part of the framework. Once you know all the inputs needed for creating a great product, you can seamlessly move back and forth among the phases, iterating until you're arrived at the solution.

I agree. In waterfall, you close the 'design door' behind you when you go into implementation. There is nothing wrong with taking a phased approach as long as you keep that door open for the inevitable changes you discover while implementing.

In fact, that's the way it should be. There is no human endeavor that does not benefit from an understanding of how something should be done, before it is attempted. I'm not sure why some seem to feel that software engineering is somehow special in this regard. I can't count the number of times we have had people shout down attempts at serious engineering design, often by using the term "waterfall" in a derogatory manner.

Not if you call the understanding and design processes a "design sprint"!

As someone else said, the individual phases can be huge, which is why in large projects the first step is often to break down large business needs into more manageable chunks, each of which can be worked on seperately, either linelarly or by multiple teams. In a healthy process as the work progresses it is used to reevaluate previous decisons & designs. So basically agile.

Reply on news.ycombinator.com