My question is: is this just a matter of something going out of fashion (perhaps driven by monetization interests), or are there technical/UX/safety reasons why you'd recommend a different solution for a new community project in 2022?
My question is: is this just a matter of something going out of fashion (perhaps driven by monetization interests), or are there technical/UX/safety reasons why you'd recommend a different solution for a new community project in 2022?
18 comments
source: I ran & hosted zx6e.net, a niche but active motorcycle forum, for over a decade. Sunsetted the site for lack of user growth combined with becoming sick of operating php trash.
It's been many years since I shut down the forum so I can't speak to the current state of things.
At the time there were two major problems:
1. the php runtime was constantly in need of updating because of yet another vulnerability either due to bugs in a forum-required feature like gd (graphics package iirc) or another stupid php default resulting in exploitable bugs (remember register_globals anyone?)
2. the forum software was poorly written by the kind of people who wrote php web software back then, not well maintained by upstream, and I didn't have any time or interest in becoming a php forum software maintainer myself.
There's only so many obvious semi-successful attempts to hack my shared server via any combination of the above before I lose interest in hosting a small web forum for free.
If I were to host a web forum today I'd want it to be mostly static content and not a constant invocation of some php or php-like interpreter involved in every operation.
It's ranking algorithm was better than "newest activity", the collapsible trees made branching discussion tolerable, and the aging of posts kept those 2000 page long top posts with hundreds of branching discussions displayed linearly from happening.
In particular in a hobby like PC building, using a message board for build logs makes sense given the format. And probably other "log/project tracking" sorts of coverage. Like Reddit, you get a free captive audience that requires less effort than blogging about it, but unlike Reddit you can post build updates to bring your topic back to the top without making multiple threads.
I do know a few sites that run custom bulletin boards for my hobbies, but that requires a developer to be available much of the time to ensure uptime. 2 examples are:
https://www.ukriversguidebook.co.uk/ - which is originally based off phpbb I think
https://www.ukclimbing.com/forums/ - which I think is completely custom
I still go on certain forums (e.g. Pistonheads.com) but the niche groups are all on Facebook these days.
That being said, any time forums come up in discussion there is so much nostalgia there. Messing around with forums software like vBulletin and stuff as a kid taught me a lot about the internet, ftp, php, hosting things, graphics design, etc.
I wonder if older people feel the same way about BBS?
I know what you mean, but I'd describe that as more "visually stimulating" rather than mentally stimulating.
Forums were great to run and get many people at once on the same forum at the same felt exciting.
Getting people to visit your site is still exciting
The main problem is the network effect (which Reddit did right) and what at 2010-2012 there was a massive change in the demographics of the new users on the Net: more and more people started to interact with it through the phones only. phpBB-like forums never had usable UI for the phones (though they are still a magnitude lighter and usable than modern Reddit, lol) and had one main problem: discoverability.
To Average Joe (who thinks Internet is that 'F' picture on the phone) the process of discovery (Internet search) and registration (e-mail based, captcha ridden) is way too much.
Add all other shortcoming listed in the other comments (notably the need to spend money, time and brain power on just maintaining the forum itself) and it's pretty evident why most of the forums died.
NB I still frequent one forum, site OP moved it to XenForo from phpBB years ago and did numerous modifications and integrations both for the old and new versions, which isn't available on Reddit at all.
Another forum is still running ikonboard, but it's userbase is... I'd say "not Discord/Reddit-compatible", which is also partly answers to your question: it is viable for some small (on Internet scale of things) community to thrive on a 'classic' forum software... but only if Discord/Reddit doesn't provide way less friction and/or discoverability.
1) high admin costs of forums due to security issues and spam,
2) usability issues from clunky interfaces and separate accounts for each forum.
On Reddit/FB/etc., the admins don't have to take care of security, and users only need one login to access any community on the site. So yes, the demise of the forums is in some ways quite regrettable, but nonetheless understandable.
I still have more than fifty forum accounts in my keepass but I can't even remember when I've last been to budgetlightforum.com, for example, even though it's still superior to /r/flashlights.
In a way it's sad, because I've seen some forums suddenly die, and the knowledge of more than a decade just vanished overnight. Vast repositories of DIY and repair information on old petroleum lanterns just gone, while the people who have experience with them are slowly dieing out. But I guess I'm part of the problem, since all I do on these forums is lurking nowadays.
Anyhow I've been writing a modern, real-time multi forum system for the last 4 years on and off, but idk if it's ever going to be ready. Forum software may not look it, but it actually quite complicated.
Forums went away and got replaced by the horrible reddit, which Google pushes in their results. Every damn exotic search has a reddit result in the top 3 results.
Facebook added to that, forums were the main info banks, then Facebook came and not only mostly destroyed forums but also classic, colorful websites. I hate to talk like that but what once was a wild, arguably pretty and free internet, Facebook turned into a white-blue dystopia. A walled garden where censorship was and is the norm, where you're identified and tracked, where you habe zero choice about the look of your profile/wall. Is it even called a wall anymore?
And what's worse Instagram is replacing Facebook. Communication by images only. How dumb is that?
Yes we need text culture back. But nowadays if I would be super smart and have the resources, I'd probably build a speech recognizing feature, where you can just record paragraphs because people don't like typing. With the option to also post that recorded passages of speech and read that text aloud. Which brings me to Adsense has this new experimental feature, reading content out aloud and monetizing that. 1+1=2?
That would actually make forums have a comeback. Because now you actually CAN monetize single posts.
P.s. those ~2500 instances were hosted on a single core amd64 with 2GB of RAM and RAID1 only.
On transitional forums, you'd have to use Google. And your ability to find a community would be directly impacted by the owners skill at SEO unless you're constructing rather advanced queries.
There was also a poor UX compared to social media. Site owner doesn't pay the monthly hosting fee from their web host? Site is down. Can't really admin a MySQL DB? Site is frequently down when it grows past some point. Shitty web host? Site is slow (or down). Reddit solves this by being centrally managed.
There was not a lot of effort to make a middle ground -- but it is not clear what UX would be right for that. The UX is largely dictated by the scale of the community, e.g. FB treats all comments as dispensable, reddit keeps topics open for 30 days, stackoverflow cares about ranking etc. Also the predominance of mobile and app stores means that those forums became undiscoverable. Everything in the past 15 years have been tailored to make the big fish bigger.
eevblog, probably biggest english speaking electronics community runs on SMF
It also doesn’t scale well for the size of modern web communities.
People also changed the way that they want to interact, from text based BB’s to more media rich, action based social networks.
Regarding scaling, node.js is probably faster because of the JIT compiler, but this doesn't mean that scaling is not possible. Wikipedia is PHP, Wordpress is PHP, and Facebook is also PHP (with HHVM).
I guess this would be more of an architecture question, rather than one of the runtime itself.
There are many sites out there that scale decently that are running on slow or sometimes slower runtimes than PHP, like Python (the Django framework comes to mind) or Ruby (the Ruby on Rails framework comes to mind). For example, consider the following: https://www.techempower.com/benchmarks/#section=data-r21&l=z... (disclaimer: sometimes benchmarks contain pretty optimized code, but here I'm selecting for the more idiomatic and boring ones; I don't want to show the benchmarks that show PHP ahead of everything, because native plugins are used there)
Someone mentioned Wikipedia and WordPress being PHP, I'll also mention Instagram and Spotify using Python and GitHub and GitLab using Ruby. If your app is architected well, then performance of any of these runtimes (or others) can be good enough, though of course one can also talk about some of the benefits of needing less resources to do more (e.g. with Go, Java or .NET, or even Rust).
Though admittedly, PHP is also popular in shared hosting, which might give it a bad reputation due to the architectures typically employed there (single server with single instance) and security has indeed been a topic in the public eye, especially because PHP was easy to get started with, which lead to a lot of projects by devs that weren't skilled.
What do you mean? Old forum codebases can generally easily support tens of millions of users, what “community” is bigger?