SMF and the old code

[Dilip]

I was an SMF user and waited a lifetime before moving to other forum software. Mainly because they started the dev in 2009 /2010 and released the software in 2020s.

I wonder how much code is from the 2010 era in the latest SMF? Does it have an old base framework over which the current software runs?

 

[Cedric]

@Arantor has wrote many codes being used currently. I’m sure he’ll be able to write a very informative reply.

 

[Dilip]

Indeed. He was a very helpful figure in the official community when I started forums. A lot of the old gang left and I heard it caused the development to take this long.

 

[Arantor]

Mainly because they started the dev in 2009 /2010

Nonsense.

SMF 1.0 was started in 2003, released in 2004.
SMF 1.1 was released in 2006.
SMF 2.0 was released in 2011.
SMF 2.1 was released in 2022.

I wonder how much code is from the 2010 era in the latest SMF?

It's hard to have accurate stats, but approximately 120,000 lines of code existed as 2.0.0.

The GitHub stats - allowing for the fact that what was dumped onto GitHub (because SMF's code repos predate GitHub... they predated the entire existence of Git) was slightly ahead of 2.0 final, and allowing for the very immediate changes thereafter, so from early 2012 through to today...

442781 additions, 311694 deletions - where 1 line changed counts as 1 deletion (the old line) and 1 addition (the new line). That to me suggests quite a lot of reworked code in that time, and a significant amount of new code added.

2.1's dev cycle is mostly the product of 'we don't actually know when we're done because we can't decide what should be in it' but that includes a replacement editor, draft post saving, a new theme, alerts, native support for international domains (including autolinking based on links and domains even including things like accented letters), and a lot of works to ensure that mods play better going forwards (assuming mod authors update and use the features provided) plus good compatibility with PHP 8 / MySQL 8 / recent Postgres / Maria 10, in ways that SMF 2.0 just couldn't really do as well.

The thing about old code, it's a known quantity. It *works*. You've battle tested it, you've had it running for *years* and many things just don't suddenly stop working in that time. One of the biggest lessons out there is to never do a complete rewrite - because it'll inevitably take you longer and more money to do than to overhaul what you have.

Examples? Well, Borland has several - they lost out in multiple of the wars to Microsoft, over Quattro Pro vs Excel, not to mention several of their lesser known products, because they started over, had to re-learn all the wisdom.

But the best example I have is Netscape. Netscape 4.x was considered by a very serious group of people to be a better browser than IE, and that had a point, to be sure. But they started on a full-bore rewrite, which never saw the light of day - this was Netscape 5. By the time they realised the hole they'd gotten into, they were so far behind it was unwieldy. They *tried* to fix that, with the joy and mess that was Netscape 6, but to no avail. What remains was eventually cannibalised into Firefox years later. But the landscape could have looked so very, very different if that hadn't happened.

I give you this from someone who was there at the time - https://www.joelonsoftware.com/2000/11/20/netscape-goes-bonkers/ - and in case anyone's wondering who Joel is, he's one of the co-founders of Stack Overflow (and its entire network, Stack Exchange). Dude has been running software businesses for years. Trello was another of his originally. Dude knows things.

SMF has many faults, not least of which a community that has no idea what it wants the software to be except to remain exactly what it is forever as far as I can tell. But the people at its heart understand only too well that a complete rewrite isn't that useful.

I should know. I actually started to try. I'm now 6 years in from that journey and backing out of it because trying to 'modernise' it doesn't get me anything at all, and starting from scratch just means redoing what's already there!

 

[Arantor]

I also think, in hindsight, there's actually even more to this story than meets the eye.

So, there was some major drama internally at the start of 2010 which derailed 2.0's progress and caused more than a few people to leave - including the #1 beta tester (not me) - and a schism began in the form of Wedge, which would remain through to 2013.

Then there was another round of drama in 2012 that led to the end-of-2.0/those-who-started-2.1 dev team all quitting. 2.1 has had at least 4 complete changes of developer that I can think of, partially because no-one wants to pick up a 'it's nearly finished, just get it over the line' project when they want to build new interesting things and partially because no-one could agree on what 2.1 should actually be either.

From my perspective, most of what changed between the last time I did any real work on 2.1 in 2014 and the release in 2022 was mostly internal stuff that not-so-many people would notice. Not small stuff, either, but stuff that took a lot of time and effort and sadly few will ever realise it did its thing.

What I will say though, is that you look at phpBB and MyBB from 2011 and compare them to now, I think you can make an argument that SMF has - for its faults - tried to keep up on some level.

As to its future... they and I are on different paths and I sadly think our paths may not cross again.

 

[Dilip]

SMF 2.0 was released in 2011.
SMF 2.1 was released in 2022.

Thanks for the stats

That is exactly what I was trying to say.

Nice to see that your avatar has not changed in years

How are you? I remember that you were working on an SMF fork. Did you complete it?

 

[Arantor]

I've only really had 4 avatars (plus one for a meta-joke) over the last 12 years or so

I was working on StoryBB as a roleplay specific SMF fork, but I spent most of the last 6 years trying to modernise it and honestly... it didn't get me anything, so now I'm back to stripping out as much as possible to be lean and reimplement all the roleplay stuff in a useful way.