from Hacker News

Ask HN: Is no-code a future of web development?

by kandruszkow on 1/5/22, 1:04 PM with 59 comments

• by DanielHB on 1/5/22, 1:59 PM

  Relational Databases are essentially no-code replacements for b-trees. I have worked in several companies where the managers could not do any coding, but could wipe out an SQL query to get some data point they needed. Heck building an UI that generates SQL queries for you based on some DB schema isn't even that hard. Relational databases removed the need for most b-tree and binary-streams manipulation coders in the 90s yet the demand for all types of coders (even b-tree and binary-streams coding!) only increased

  As tools evolve programmers get exponentially more productive, but they will still be coding. They might be coding different things though because what we do today can now be done by a tool

  People point UI work will be replaced, I actually highly doubt that, unless there is an standardisation of UI-toolkit like we used to have in the windows 98 days (which seems incredibly unlikely). Remember Delphi and Visual Basic? Sure there was code in there, but very rarely did you need actually need to write UI-code like animating a text box around, most of it was business logic code

  If anything the UI landscape is changing to be more custom for each app instead of consolidating. I expect UI coding will look very different in the future though, more focused on unique user interactions and animations instead of API calls and Form glue code

• by usrbinbash on 1/5/22, 2:15 PM

  This is a cycle that repeats every 10-12 years imo.

  And then, after 2 or 3 years, the hype is gone, because it turns out that

  a) Complexity can be hidden and displaced but not eliminated

  b) Vendor Lock-In issues

  c) Interoperability issues with existing tools, platforms and systems (ever tried to version-control a nolo-code project?)

  and last but certainly not least

  d) As projects get bigger and requirements pile up, at some point there is a requirement not satisfied by the platform; then the dance to somehow interface it with "some code" starts, then the next requirement comes along, and the next, and at some point somehow aska the question "Wouldn't it be simpler to rewrite this in XYZ" and at that point the cycle usually ends.

  That doesn't mean that these things do not add value. The hype is usually gone after each cycle, but useful things that help a lot of people making things they would't have otherwise made remain.

  But it doesn't replace coding.

• by Ma8ee on 1/5/22, 2:56 PM

  Low code is what we are doing every day with modern programming languages and developer tools. We write a fraction of the code someone had to write 30 years ago for the same functionality. In that regard, the future is already here.

  And developers haven't been replaced, and it is unlikely that they will be. But the role has already changed. A bigger part of being a developer is to understand business requirement and talking to customers (internal and external), while previously you could rely on your precious technical skills to be irreplaceable.

  Finally, for many people No Code seems to mean "graphical code". That is certainly nothing new and is great for some things, like developing interfaces and I have seen some reasonably successful ones for developing database schemas. But for expressing flow and logic, text still rules supreme (for a number of reasons).

• by codingdave on 1/5/22, 1:22 PM

  We can look to Salesforce for how this will likely play out - they aren't the only no-code platform out there, but they are big and have been around a while, so their dev culture has had a long time to evolve.

  In that world, most devs work primarily in the no-code portion of that platform, maybe with some basic coding skills. That group does the majority of app dev work. But there is a smaller percentage of people who do write code, and who do the work that goes beyond what the no-code features can do.

  So most other platforms will likely match - a wide swath of people, functions, and apps are no-code, and you absolutely can succeed both doing no-code work for your job, with most businesses not needing more than that. But there will always be another layer underneath the non-code to let businesses who need more to have coders who add more features.

• by tjpnz on 1/5/22, 2:02 PM

  For relatively simple projects it's already happening, I recently booked a doctors appointment online using a site built with a no-code tool. Turns out the physician built it himself and admitted to having very little in the way of technical knowledge. Not so long ago he would've had to have hired a fullstack developer for the job.

  Would a hospital booking thousands of appointments in a given day be able to get by with the same no-code solution? No, but there's only going to be so many of them.

• by ksec on 1/5/22, 1:58 PM

  I still think no one has fully crack the low code / no code tools on the Web. You cant have a No Code solution that doesn't offer Low Code function. And you cant have a Low Code solution that you cant expand to Full Stack Code.

  Think of it like Game Engine, where you could make games with editors, and then you add some code, all the way to some AA Games ( May be not AAA ).

  May be a better example for the Web would be something like Yahoo Pipes.

• by spcebar on 1/5/22, 2:20 PM

  I think there will always be a boundary with these tools where if you are building something complex you will reach a point in the project where you need something more advanced than the tool you're using can provide. For me at least, that was the spark to learn to code in the first place. You could call it the law of conservation of code, but I think there is a limitation on how complicated the output a no-code tool can create before it eventually becomes a code tool.

  That said, for the chiropractor who needs a website to list his phone number, address, and services, I think a future where he doesn't need to hire a web designer and potentially get scammed (I've seen this a number of times) is a bright one, and I do think we're headed towards a future where the tools are good enough to make very basic front end dev work for situations like this obsolete.

• by DanielHB on 1/5/22, 2:32 PM

  These kind of no-code or low-code tools are VERY common in enterprise, like SAP, Salesforce, CMS solutions, IBM Tivoli products, and on and on. But companies are starting to realise that those tools often have some hard limits on what you can do and the problems they are solving get so complex that hacking the tool around to solve the problem becomes unsustainable. If anything even enterprise companies are pivoting away from this kind of tool towards custom-built software.

  I call it the "army of consultants problem", sure you can use a low/no-code tool, but if your problem is complex or big enough you will eventually need an army of consultants specialised in said tool to maintain your solution and you will be limited by what the tool can do. Eventually you get to that point it becomes more expensive than just building custom, but by then your company is too invested in the tool and only a band-aid ripping can fix it. Most often than not ripping it out would kill the company/product if it happens too late. Thing is, the tools give the impression they are cheaper because you get a lot "out of the box", while a custom-built tool takes a long time to do anything useful

  Reminds me an article I read once, it was about why My Space failed, it propositioned that the main reason was that they were using Dreamweaver and their product was so big and complex that the devs could no longer handle the complexity to add new features or fix performance problems. And the company as an organisation could not find enough dreamweaver devs because they needed an army of them

  One of the main triggers for this is cloud infrastructure is so much better and cheaper today that a single developer can orchestrate a really complex system of services and servers in a few hours if he knows what he is doing. So the industry as a whole is shifting from "army of low skill specialised devs" to "low amount of highly skilled generalist devs". You can see it in the job openings, nobody wants to hire low-level anymore, or if they do it is to support legacy systems

• by rchaud on 1/5/22, 2:14 PM

  This questions keeps coming up, and without knowing the specifics of the tools being discussed, the answer will always be "it depends".

  I recently started using a no-code tool called Glide[0], because it emphasized that the starting point could be as simple as a Google Sheet. That data is pulled into the app's visual builder interface, so you can start playing around with filters and conditional logic to show/hide information.

  Then you can set up user logins, Stripe integration and basic CRUD functionality with read/writes happening on the Gsheet. You can string together actions with Zapier and create calculated columns like you would on an Excel pivot table.

  Ultimately, I can see myself creating a small community app using this tool, but not something that has hundreds of thousands of users, as then we get into tricky issues of vendor lock-in.

  So IMO, the use case for something like this is for rapid prototyping and and an MVP. You wouldn't want your entire business hosted on someone else's platform, so as you get user feedback on the MVP, you should be talking to a dev about how to move off-platform: hosting, deployment, using a proper DB, security reviews ---- all the ugly stuff the no-code tool abstracted away.

  [0] https://www.glideapps.com/

• by pkrotich on 1/5/22, 1:47 PM

  I think it will eventually provide most of the scaffolding especially for frontend. Main resistance is with OG developers - new devs will embrace it going forward.

  I remember when automatic cars came about - people who were used to manual shift resisted it, but now it’s almost silly to push back on automatic.

  Another good example is house printing / prefabs - it’s just a matter of adoption but builders provide the resistance because of money they’ll loose.

• by gostsamo on 1/5/22, 1:18 PM

  No-code tools fail when it comes to leaky abstractions and complex logic. They might solve some trivial cases, but they are not the end of programing.

• by oli5679 on 1/5/22, 1:57 PM

  Huge amounts of analysis and data-recording happen with people juggling spreadsheets, because they are easier for 'business/finance' people to learn how to use them vs. coding and databases.

  I think there is huge potential for improved spreadsheets, either with generic things such as greater scalability/ability to schedule reports or with inbuilt capability for common tasks for specific industries.

• by ericsilly on 1/5/22, 3:02 PM

  One yet-to-be-mentioned element here is the dynamic of developer preference and resistance. As systems bifurcate into relatively cookie-cutter "lego"-like technical components (the Wixification of the entire stack via cloud, frameworks and no-code), vs the heavy-lift coding to make the framework "legos" themselves, a lot of the fun development is going to go away (development with a direct end-user feedback loop and the reward of solving some real-world problem). This dynamic also eliminates a good entry point for developers (basic apps and web sites), which allow them to build the knowledge to create and love those more sophisticated re-useable legos, because they understand the true pain of solving those original problems without them. I predict a big, negative impact on the future population of developers arriving to software from "creative" entry points, like design, marketing, or finance, because the bridge will be gone. What will be left are only formally educated engineers.

• by llimos on 1/5/22, 1:56 PM

  It's definitely 'a' future, in that there's no good reason why it would go away. It'll get better and better, but there'll always be something it can't do well.

  For me, the key to a no-code solution is that it provides an escape hatch into code when necessary. If I have that, the no-code stuff is actually really nice.

• by tikkabhuna on 1/5/22, 1:56 PM

  We don't use Low or No Code solutions at work, but I wish we did. There's tons of automations we could do, however, because everything has to be a proper app with a proper deployment, etc, its a big job. I hope that Low/No Code solutions make it easier for a wide range of people to automate their jobs. I expect this will be accelerated with the next generation being more tech savvy.

  We're already seeing similar parallel in Finance with End User Computing[0] where non-developers (traders, etc) are using tools like Python, Jupyter, etc as part of their jobs.

  I don't ever see it replacing developers though. Maybe if your entire job is workflow automation, sure, but developers will focus on solutions that are difficult to do with Low/No Code tools.

  [0] https://en.wikipedia.org/wiki/End-user_computing

• by nonameiguess on 1/5/22, 1:52 PM

  Even if they take off, they're implemented in code, so they can't obsolete developers. Similar to the existence of cloud platforms not meaning there is no longer a need for platform dev and system administrators or even basic lab techs to physically install computing equipment. It just means for a sufficiently small business that doesn't want to roll its own infra, those roles are now employed by the cloud provider rather than the small business directly. Similarly, if no-code ever truly takes off, it just means devs who specialize in small business one-off websites won't be employed directly by small businesses and will work for no-code toolchain companies instead. The result will be consolidation and probably some amount of sector downsizing or else such products can't offer any real value, but it won't result in complete elimination of human labor.

• by haspok on 1/5/22, 2:17 PM

  >Will they replace developers...?

  What an odd question to ask! Who do you think is going to code / maintain those "no-code" tools then? If for nothing else, this is one reason why you'll never obsolete developers.

  But actually, if you think about it, the best you can do is that you raise the level of abstraction - that doesn't mean the new level will not need developers.

  Of course, for certain common "boring" software will be no-code, which I'm actually quite happy about. Who would want to code the N+1st CRUD app over and over again?...

• by PragmaticPulp on 1/5/22, 1:28 PM

  No-code is fantastic for basic websites and apps. Prototyping basic businesses has never been easier than with no-code tools that let people string together common functions.

  The key is to know when it’s time to outgrow the no-code version and start a real custom implementation. Not all businesses will get there, but the developers need to know when it’s time to switch over.

  Fortunately, replicating a no-code site as a custom solution is actually quite easy in most cases because the no-code building blocks have a lot of easy to use code library equivalents.

• by soueuls on 1/5/22, 1:21 PM

  I think it will reach a point where it's similar to page builder. It works, sometimes that's all you need. But it's bloated, not always optimal, you will likely reach a point when it's clearly not enough.

  In my opinion there is a market for it but they should not fall into the trap of trying to be generic (like bubble for example), when they do, they are quite hard to use that you might as well learn coding.

  But for specific use cases building a simple and custom CRM, showing structured data, prototyping etc it's good.

• by TYPE_FASTER on 1/5/22, 3:29 PM

  I've been seeing ads for No Code products for decades. I've watched the various categories of development tools (IDEs, configuration management and version control systems, etc.) mature over the years.

  Good development tools can be used to accelerate development when they help you leverage your knowledge of the system and technology. What I haven't seen them do is replace a deep understanding of how things work that's necessary to build scalable and performant software.

• by dogleash on 1/5/22, 2:16 PM

  The tide of software is always turning from "requires custom development" towards "turnkey solutions exists." "No-code" is a buzzword that sits on one of the contours between the two.

  Of course the web is/will be changing. The only interesting aspect of the web is that it expends so much effort changing it's fashion rather than it's form that this question even has to be asked.

• by ermir on 1/5/22, 2:03 PM

  Development will still be hard intellectually, but much easier when you don’t have to deal with symbolic manipulation. I don’t think no code is a solution for the web, but in specialized domains such as video games it could have a major impact. Think of games like DotA that first came out as mods made by specialized editors instead of programming languages.

• by Kathy8888 on 1/5/22, 1:15 PM

  For me is a great opportunity. I'm "no code" person so I'm really happy that finally I can create website alone. Of course sometimes I have some troubles but its not something that I can't learn by reading some article or watching tutorials. So for me no code tools are big "Yes".

• by shaolinspirit on 1/5/22, 2:02 PM

  I think it depends on the product itself. If it is a salesy one pager then TailwindCSS has already a lot of prebuild components, if it is something really creative, like with 3d graphics or crazy font animations then you need a dev.

  in an age where everything will look almost the same there will be demand to be authentic.

• by BackBlast on 1/5/22, 3:29 PM

  Depends on requirements. Most no code platforms I've seen are VC funded with all the baggage that contains.

  If I'm building a system with an intended life span of >=5 years I would avoid a no code platform.

  If I'm doing a one off proof of concept test app, it's on the table.

• by mule1 on 1/5/22, 2:21 PM

  Looks like software developers will become luddites too if their job is threatened. Big portions of development not just UI development will need little to no human input after artificial general intelligence becomes common place.

• by dmingod666 on 1/5/22, 1:54 PM

  'No code' is comparable to the 'serverless' trend.

• by jakubkopacz on 1/5/22, 1:36 PM

  Recently, I came across some stats about no-code/low-code. For example:

  - Forrester, said that low-code can make software development 10-times faster than the method of writing complex code.

  - On the onther hand, According to Gartner’s report, by 2024, low-code application development will be responsible for more than 65% of application development activity.

  Of course, these are only predictions, we'll see if they come true...

  However, in my opinion, No-Code / Low-Code tools can become invaluable support for developers, especially by eliminating simple, repetitive activities.

  Thanks to them, those who know how to code will be even more valuable. Developers will focus on developing tools that require more complex skills, not provided by no-code tools

• by caporaltito on 1/5/22, 1:56 PM

  And remember, if an article has a question in its title, the answer is always: "No."

• by est on 1/5/22, 2:12 PM

  no-code is more like the past of web development. Macromedia Dreamweaver, Shockwave Flash, Frontpage, etc. And you know what, a good authoring tool unleashes more creativity than today's boilerplate SPA bullshit.