The Future of Software Development?
Вставка
- Опубліковано 12 тра 2024
- What will software development look like going forward? Taking a look at how the past may have some insights into the future...
Table of Contents:
01:29 - The Internet Takes Off
02:36 - Drag and drop tools
05:26 - Legacy modernization tools
06:18 - Modern visual tools
08:33 - Backend and other tools
10:29 - Why it feels strange
13:17 - The new developer split
FYI, none of these companies paid me to do this video, no affiliate codes etc.
For the companies that offer self-host, I've included links to the relevant GitHub repos as well.
www.scullinsteel.com/apple2/
infinitemac.org/
emotionalyst.blogspot.com/201...
netbeans.apache.org/
tsri.com/solution
docs.openrewrite.org/
aws.amazon.com/windows/produc...
bubble.io/
airtable.com/
appsmith.com/
github.com/appsmithorg/appsmith
budibase.com/
github.com/Budibase/budibase
retool.com/
github.com/tryretool/retool-o...
tooljet.com/
github.com/ToolJet/ToolJet
supabase.com/
github.com/supabase
Software Engineering/Development imo is never gonna "die" it's just gonna shed it's skin and become something different. I definitely feel like everyone is in reserve mode until something big happens. The best thing you can do right now is prepare yourself for when the tide turns and the economy roars back to life again.
Love hearing thoughtful takes from peers that have been in the industry with a lot of experiences. Appreciate you taking the time to share your thoughts.
Thank you for sharing this through your wisdom and experience, that was very valuable! Much love and all the best from Bulgaria
These low-code app builders have been common in the enterprise for a long time. I've been working with one in particular for close to a decade now. While you can do a lot without writing code, knowing how to code makes you a MUCH better developer, as almost all of them have ways they can be extended with code, as you mentioned. Most developers working on these tools in large companies don't have much formal programming knowledge or experience, so there's a lot of opportunity out there for people with programming skills who want to pick up one of these platforms and work with it. Also, since they're all specific tech stacks, the labor pools tend to be smaller individually, and the salaries are actually pretty decent. One thing to remember is that using visual tools to code is still programming. You're still a knowledge worker!
A hot take people can try out for themselves however. I was actually just thinking as a current CS student that I should just go ahead and take one of my project ideas and literally just see how fast I can do it with AI and more abstract tools (like drag and drop). I think it will be good to know the limits of what can be done and also see how it personally feels to do that. Interesting that this video pops up right after thinking that.
Man this brings back memories. My first rig was a Vic 20 and coding BASIC was my after school thing too. When I was a teen, my parents would say I need to learn a real skill to be able to support myself later. By 25, I was making more than both of them combined. I don't think our parents saw that coming!
Thank you for the English subs that help non-native speakers better understand your videos.
Great video. Thank you. The bifurcation that you describe between the core developer and drag-and-drop developer seems like a reasonable direction. Clients, at least in that small to medium business space, have become increasingly intolerant of the overhead of traditional development and that discomfort is hastening in step with the acceleration of new technologies. Add to that, user experience seems to be getting more homogenous and this might become more so as different modes of interacting with applications take hold (more chat, voice, gestures, vision etc). A lot of the typical ui elements we navigate around with today might possibly become just friction in users getting what they want from an application.
Visual editors have been around for a long time and they always work fine when the complexity is low. If the complexity increases, you will get weird bugs that need you to look to the actual code.
At the moment designers live in Figma (and it's a wonder why UIs can't be exported directly from there). But it's quite reasonable to think that frontend will split in two (front of frontend, e.g. Figma + UI. Versus the back of frontend, e.g. state management, libraries, etc) in a similar way that backend has split into two (front of backend, e.g. API building. Versus the back of backend, e.g. DevOps)
Omg i remember that logo on your shirt from one of my childhood lego space sets it was on the astronaut minifigs 😭
Yeah, makes sense
I think the visual tools, i.e. drag and drop can be useful for describing visual aspects, i.e. screen layouts. Graphical flow tools may have some use in describing what I'd call higher level 'orchestrations' i.e. if this then go do that....however when it comes to even moderately complex business rules (as opposed to standard CRUD logic) I still think you need code. I do accept that the number of jobs to maintain the programs that the world needs may well drop drastically and may not be as well paid. Personally I loved the way the old WinForms apps worked, with a visual builder for the UI and then the ability to write C# against any event - I'd like to see a web based version of that.
Great observations. I enjoy craft and take pride in what I create... as opposed to prompting an LLM to do it. Interesting times indeed.
I used various drag and drop over the years, meaning home grown or open source. These nocode lowcode variant as a hosted service, without the ability to install a module in react (or your fav framework) and compile and run without the cloud - I mean if I can't run it without the internet, never mind about the subscription price - then its worthless. At least for my work, complex applications going in embedded systems or internet based enterprise applications interacting with embedded (IoT without all the cloud - as in a simple LAN). I think drag and drop to design UIs is great, but only if i can later refine and calibrate the code with my own fingers and a "must" is that it is stand alone - not running on someone else's cloud.
I love it when tools make my job easier, and I will definitely be looking at things like Bubble and Buildship to see what's possible. My only skepticism is about the tradeoffs. Will I still be able to drop down to write custom code whenever it's necessary? What new frustrations will come from trying to interact with an LLM to get a usable result? I
I feel like it's just a natural maturity of the Software industry like most other high skilled profession. Maybe I'm just a snob but I never considered building something like a web-store as engineering. So yeah, we might lose the need for a lot of "coding" jobs, but I don't see Software Engineering being in danger any time soon, but it might be over for those who were able to find a 100k job by learning some html/css/js in a 3 month bootcamp.
I love your videos
From an education standpoint, is there merit in learning how to write code by hand before learning these drag and drop tools? Or should it be the other way around?
I'm old enough to remember when computers were going to replace us all. Instead, they ended up creating more jobs. I suspect AI will alter the software development industry, but not in the way everyone thinks it will. It will empower those in the know (younger-cheaper developers who grew up using it) and bury those who fight it (old and expensive developers.) It will do the same with artists.