Breaking up with WordPress after two decades | Hacker News Reader

Breaking up with WordPress after two decades

yusufaytas.com

73 pointsowenbuilds39 comments

The Achilles heel of WordPress is its backward and brittle storage format: every post is stored in the database, in "html kinda" format, plus random shortcodes initially, and now blocks encoded as JSON inside HTML comments. Plugins then add their own random formats on top of that, and store their metadata in another table, usually in serialized PHP object form.

It's a wobbly Jenga tower of technical debt, and WP's underlying architecture makes this mess fundamentally unfixable. There is zero political will to address these deep-seated design issues, or at least Matt Mullenweg has no will to fix them, which translates in reality to never getting fixed.

They seem to be interested mostly in adding more and more visual editing features that keep making the whole thing more unstable and messy. There's no hope. We already have clients who were with WP for many years switching to Directus and a custom front-end. The only reason others are still in it is the huge ecosystem and the custom integrations they have built around WP.

Directus seems an unorthodox idea for a website. Is this a marketing post?

Nope, I just found about it recently. It seems familiar for a WP user but decoupled from the front-end. We decided to try it - time will tell if it's a good decision or not.

The design issues are also in other parts of the system. Major silliness of WP is, that basically every tutorial and documentation encourages a pattern of concattenation, instead of composition. This is in PHP tradition, which just runs scripts from top to bottom, which means you can change code while it is running and next request it will run your changed code, if it passes by that part of the code to handle the request. In the WP ecosystem countless plugins engage in making their parts unreusable, by following this concattenation idea. They will start some HTML element in one part, but not close it. That job they leave to another part, which you just have to implicitly know must come later, to close all the tags. The different parts are described at [1]. This also implies treating HTML as a string, also in old PHP fashion, with all the issues coming along with that.

Worst is, when you have people, who even manage to break composition on the procedure level by doing the following: (1) open php "tag", (2) write function header, (3) close php "tag", (4) write plain HTML inside the PHP function body, which results in outputting that HTML as a side-effect, instead of returning it as result value, (5) open php "tag", (6) insert closing brace of function.

Literally, makes the code only usable in very specific places, where that side-effect is just in time for the document creation, otherwise cannot be used. You cannot use this and then wrap it with some other HTML, because it is already output as a side-effect. Of course since it is stupid tradition to treat HTML as string, you also cannot pattern match on that stuff and cannot have intelligent operations work on the result.

But not only this major F up. No. The use of global state is very pervasive. One doesn't pass in information to templates to then have information locally and render templates independently from the rest of the system. Instead rendering templates happens while frequently retrieving information from the global state.

Another bad design is how one provides metadata for plugins inside comments inside the code of the plugin, instead of having a clearly separate metadata file. Of course this doesn't lend itself well to being edited automatically, as one would have to parse PHP code and then work on an AST to change it. But I guess no one responsible for this design has ever thought that far.

The whole thing feels more like it is made for people, who live edit code on a server and then check 20 times whether it now works and the moment it looks right, they drop everything and call it a day. This is also the impression that the quality of plugins leaves. You want to see bad code, you just grab a random WP plugin.

WP is in essence a guide showing by example how not to develop a website.

[1]: https://developer.wordpress.org/themes/classic-themes/basics...

A lot of WP's architecture could be fixed internally, with shims added for backward compatibility. I took a stab at doing just that in a fork, but what I really don't know how to fix without major BC breakage is WP's reliance on global state. And these days there seems little point in incremental fixes, owing to the AI elephant in the room that doesn't need further mention. I suppose if WP were my livelihood, I might think otherwise, but even though I have several good friends in the WP community, I've pretty much written off the platform as a lost cause.

I worked with and advocated for WP for projects for years. It’s not pretty, but it is well-supported and well-understood; it gets the job done. With Matt’s treatment of WPEngine (with whom I’ve worked with on and off over many years, and have had nothing but good experiences) I’ve written the project off. It’s not just his response to WPE, but the threats, the dictatorial reactions to anyone speaking up, the blatant takeover of plugins etc. I was sad at first, but I’ve come to peace with never touching WP again if I can help it, and this is as a top contributor to the WP StackExchange - I _loved_ the project and the community, and helping others made my heart genuinely happy.

I know it's not the main topic of the post, but I think it may be worth pointing out that Bluehost is owned by a company infamous for poor quality hosting solutions and trying to monopolise the web hosting field:

https://en.wikipedia.org/wiki/Endurance_International_Group

So regardless of whether you go with WordPress or a different solution, you may want to be cautious about that company. They tend to oversell their services a lot, and care mostly about cost cutting.

Unfortunately that industry has been in a race to the bottom for a very long time. Most customers just care about the dollar amount they're paying. In a race to the bottom there is no corner that can't be cut.

I've met so many people over the years who've worked in the industry at some stage (including myself), and not one of us has had a positive experience.

All that to say, it's almost certainly true that unless you've specifically heard otherwise about a hosting company, you're probably supporting assholes whoever you use.

Where is that race to the bottom visible? Surely not in the pricing -- bluehost's intro offers are already expensive (10gb space shared hosting for $3.99 a month -- in 2026?). After a year, it jumps to $11.99 with 12 months terms. That's more than $1 per gb storage. In 2026.

There is no race to the bottom.

Could you do it cheaper?

The $10/month gets you storage, but also bandwidth and hosting and a bunch of tooling. Worth it? Probably so if you want something that mostly just works.

Me, individually, providing it to one customer? no.

At scale, providing it to tens of thousands? yes.

It's a perfectly fine price for a customer to pay and not worry about it, but it's not squeezed to extract every fraction of a cent because competition is so fierce. In a race to the bottom you'd expect the bottom to be approached, but it isn't.

Bluehost, Kinsta, WPEngine, GoDaddy etc, are marketing companies that sell webhosting, and they have very healthy margins. They compete on ads, not on price.

After a terrible, possibly criminal experience with a hoster, I ended up with: https://www.interserver.net/

I am very happy. The speed is insane. I always thought and was told that WP is the reason for a slow website. No, it was my host. I pay around 10 USD per month but I think the smaller plans starts lower.

For what it's worth, I am very happy with them. But I only host a few WP and FreshRSS. I think they support python too but for Django I use: https://www.pythonanywhere.com/ I pay 5 USD a month there but I think this plan is not sold anymore.

> trying to monopolise the web hosting field

What does this even mean?

They own at least 74 different web hosting companies and are buying more brands all the time. They want to own a large percentage of this field at least.

Seems like they went through a bunch of vibe coding only to come up with a half-baked copy of Astro. And why not just use Astro? It's got the exact features that they've built, many more, and a large and bustling community.

Might want to check out Kirby cms - sounds like you’ve vibe coded your way to something similar but Kirby will have more niceities.

+1 on Kirby. It brought back fun in working on projects with CMS as a requirement. The cost for a license is easily justified when I calculate working hours and nuisances that would come with Wordpress.

For now I didn't find the comparably tiny plugin ecosystem to be a problem. It's well documented and lets me write missing functionality quickly.

My first encounter with it was because a wordpress project annoyed me so much I started to look around for alternatives. I took the project as a real-world example to take a look at Kirby and rebuild the site in an afternoon.

Yeah likewise. Most of the things you’d need a plugin for in WordPress are already part of the core system, things you need are easy to extend.

I’m doing some really powerful stuff with the custom CLI commands, highly recommend the YouTube video on it. Makes bulk import / api enrichment and other things a breeze.

    Indemnification  
     
    By accepting the Agreement, you agree to indemnify and otherwise hold harmless Us as well as Our officers, employees, agents, subsidiaries, affiliates and other partners from any direct, indirect, incidental, special, consequential or exemplary damages arising out of, relating to, or resulting from your use of Kirby or any other matter relating to Kirby. This paragraph also applies to you if you are not the licensee (e.g. if you use Kirby while someone else is the licensee).

Doesn't sound too nice

The MIT license comes with pretty much the same sort of thing at the end, and in all-caps at that.

Disclaiming warranty and indemnification are different things.

Kirby rocks. It's so much better designed than WordPress.

Not open source though.

Yes it is. You can view it on GitHub?

It’s licensed, and I’m happy to pay for a professional sustainable solution.

There must be something in the air, because I just migrated away from WordPress & Bluehost too: https://www.codeofhonor.com/blog/migrating-away-from-wordpre.... I'm so much happier with the site.

The solution I'm using is certainly not as mature and featureful as WordPress, but I don't have to patch it for security holes so frequently either.

I had been using WordPress since, what I remember, back in 2004 or 05. I think that was the first time I used WordPress and I stopped using it last year when Vibe Coding helped me build something with NextJS and Sanity and other CMSs as well. I've actually quite enjoyed the Vibe Coding part but I do miss WordPress from time to time.

Chat -> static site

I've gotten my most tech-averse friend on this flow.

WordPress had its time, and I'm thankful for that time. But the overhead is no longer worth it for the broad audience it use to serve

I need to do this also. Thanks for the inspo.

a decade late... kidding, i totally get it

Allright everyone, back to Drupal and Joomla.

Wordpress is done, SaaS is done. Static Software is done. Now people need dynamic, fully personalised and compounding software.

Usage stats - still on the up. ~37 millions sites up from ~32 million a year before https://trends.builtwith.com/cms/WordPress

I had a friend who used it for promoting businesses and one the main advantages was it wasn't "dynamic, fully personalised" which meant you could use standard plugins, swap in different programmers who knew how it worked and so on.

WordPress is dynamic and incredibly customizable. You could vibe code a huge variety of web applications onto WordPress. Not saying that’s a good idea. But WordPress is successful exactly because it’s dynamic and highly personalizable. So your comment makes no sense.

What is "compounding" software?

Maybe compounding interest on all the technical debt generated by the LLMs?

Perhaps like a pharmacy that mixes their own medications from more basic ingredients and is known as a “compounding pharmacy”.

Probably just means everyone has to vibe code their own software.

Sure buddy. Wake me up when it actually happens. What good is being able to vibe code all your software if you have to maintain it all yourself after? On Anthropic's token quota no less?

LinkedIn post nonsense

Reply on news.ycombinator.com