>>107296
What's in >>>/v/8055339 is what HW said in the stream, I'll copy it here for the lazy and those with limited bandwidths (thread there is very long).
>> Can Next be fixed
>Yes. There are two main problems with next. The first is that Josh didn't bother to target 2ch's server hardware; he wrote it to be hardware-agnostic and then optimized for the test servers that have twice the number of cores as 2ch's servers. So currently the 2ch servers get overloaded by Next's constant I/O since they're built to cache stuff, not keep rewriting and responding to requests on-the-fly.
>The solution for that is to add a cache, which they're doing, called Varnish. It may or may not have been completed a couple weeks ago, if I'm reading the logs right.
>The second problem is that Next is missing some stuff. From a development perspective, it's minor things. From a user perspective, it's major. The UI doesn't quite work right, and it doesn't quite look right. There's a lot of minor display bugs that make the site act funky. Webms don't repeat, and the images are big. The footer is missing the disclaimer. There's a bunch of stuff like this that just is missing, but nothing major. But it's contributed to this problem where people hate Next because it looks and feels wrong and Josh has gone out of his way to ignore requests/complaints about look & feel because he thought he knew best.
>Getting new hardware or fixing Next to work well enough on 2ch's current servers is doable. It's a project, but it's doable. Fixing minor issues will come with time and having an active developer that can listen to the community. Also doable.
>> Should next be fixed?
>Yes. From a technical perspective, Next is incredibly easy to fix. Both vichan and lynxchan need major overhauls just to get working. Vichan will probably die eventually, and Lynxchan is immature and needs a lot of work, but it shows promise. Still, it could be awhile until it's ready, and it will need modifications, and it's written in a language that's mostly used by hipsters.
>> Can the software running the board now be fixed.
>Vichan is what the board runs now. Hotwheels is probably going to end up trying to fix it, but the main problem is the way it's structured. It rewrites the main file constantly, and it's just too intensive and too hard to keep up with. That's the simplest way to put it. There's no easy solution to that, just mitigating how often that needs to happen and improving performance, but these are band-aid fixes. I'm not confident the new front end controller will solve all the problems, and it might introduce a lot of bugs in the mean time. Additionally, tinkering with vichan while it's live could result in a lot of problems being pushed to production. At least Next and a theoretical lynxchan test would be contained to a beta
>> If Hotwheeels adopts Lynxchan, how would that be different from the board we're using now?
>You can go over to endchan and see for yourself. The main difference is that we need to port all the infinity functionality over, and the developer doesn't work for us or associate with us directly. I doubt users will see a big difference, since the UI is almost identical, but we also can't migrate our posts over or do much of anything else. Switching to lynxchan would basically mean starting over from scratch using someone else's platform.
>In short, after spending the last month shitposting at Josh about his retarded project management skills, and even shilling for Lynxchan once or twice, I think Next is our best best going forward. It's the most effective use of our time, and has the highest chance to pay off.