/tech/ - Technology

101010

Posting mode: Reply

Check to confirm you're not a robot
Name
Email
Subject
Comment
Password
Drawing x size canvas
File(s)

Remember to follow the rules

Max file size: 350.00 MB

Max files: 5

Max message length: 4096

Manage Board | Moderate Thread

Return | Catalog | Bottom

Expand All Images


Anonymous 12/28/2018 (Fri) 00:28:45 [Preview] No. 13050
step aside lynxchan fans there's a new "imageboard engine" in town that doesn't suck shit like your abomination does.

http://futatsu.org/

https://github.com/DangerOnTheRanger/maniwani


Anonymous 12/28/2018 (Fri) 00:35:01 [Preview] No.13051 del
>>13050
>Per-thread auto-generated IDs, with avatars!

Interesting gimmick but it'll get old and only slow down the website. Esp if you browse with Tor or on a mobile.


Anonymous 12/28/2018 (Fri) 01:56:28 [Preview] No.13052 del
(12.14 KB 380x380 yotsuba question.png)
What the fuck is this reddit horseshit?
the 2chan/yotsuba style has persisted for a reason nigger, and don't say it's better for mobile because zero-effort phoneposters are the cancer that's killing all forums including imageboards

If you want altchans to take your software seriously you need a UI that isn't designed for cerebral palsy and you need to actually talk about why your engine is better. I mean people are still using vichan for fucks sake, unless you can show some real performance/feature wins lynxchan and blazechan are going to eat your lunch and you'll just have yet another imageboard abortion like gochan


Anonymous 12/29/2018 (Sat) 15:41:59 [Preview] No.13056 del
>>13051
Not sure it'll get old, but you make a good point about speed. I'll make a note to keep it configurable.
>>13052
>the UI
I don't think the 2ch/Futallaby style was necessarily a stroke of UX genius, and I think imageboards have been hurt by people not experimenting with the formula more often. The frontend is subject to change and I've been taking suggestions and integrating them, but I intentionally want something a bit out there to separate Maniwani from the current offerings and to see if we can't come up with something better.
>features
Couple things come to mind that I don't see in other engines:

1) Super easy deployment thanks to Docker; if you got docker-compose installed, spinning up a fresh instance of Maniwani backed by Postgres/Minio/nginx from scratch is 3 commands. It's a lot easier to try out than other imageboard engines. I've tried to put in a lot of effort to make rollouts easily automated; you can define your boards, board settings, rules, and general configuration with two files and keep them where you like, mount them in Docker, and bam, Maniwani is configured to your liking. I'm able to easily roll out new updates from Github to Futatsu for this reason.

2) Great media support; Maniwani is the first imageboard engine (or if not, definitely one of the first) to support WebP, and one of the few to support uploading text directly as an attachment with the thumbnail generated from the file contents. I also plan to add the ability to render Latex attachments and view attached 3D models in the browser, which will be a really great help to people uploading art and literature. I also have some stuff planned for the translation board that should make it trivial to collaborate on subtitling anime and other video without needing any sort of software installed, with the end result being an SSA file that you can plug into any video player.
The thumbnailer in general accepts arbitrary MIME types (though it can only properly thumbnail text, images, and video for now), and you can configure which MIME types are valid on a per-board basis. If you wanted to allow mp4 uploads, you could do that. I'm fairly certain AV1 will just work too (ffmpeg supports it and I use it for video thumbnailing), but I haven't bothered to try.

3) Media storage is handled with S3 and not via flat files or a document store the way a lot of imageboard engines work. So proper CDN support and whatnot is extremely turn-key, and the application server doesn't have to strain and serve both general requests and media. Maniwani actually does include a flat file storage backend, but I use it for testing only.
4) Markdown support in posts, I think that's pretty neat.
5) A REST API that supports both reading and writing, i.e, you can post and moderate via REST. The goal is to be able to do everything you that you could normally do on the web interface via REST, which seems a reasonably rare feature among imageboards.
6) Fast development pace; everything you see was done in the only past 7-8 months starting from scratch. I also pride myself on taking criticism and suggestions well and being a generally accessible head developer.

There's other stuff related to Futatsu itself (the narrow board topics, elected moderators, some other stuff), but those are the feature points that Maniwani in general can claim to have.


Anonymous 12/29/2018 (Sat) 15:49:55 [Preview] No.13058 del
>>13056

> Maniwani is the first imageboard engine (or if not, definitely one of the first) to support WebP

nntpchan supported that ages ago


Anonymous 12/29/2018 (Sat) 16:06:53 [Preview] No.13059 del
>>13058
Good to know, I hedged that statement a bit since I didn't look too hard outside of the bigger HTTP-backed engines. I need to give nntpchan a shot one of these days.


Anonymous 12/29/2018 (Sat) 16:08:26 [Preview] No.13060 del
>>13059
technically it supports it by not supporting any specific filetype, it supports anything your local imagemagick and ffmpeg supports. on failure it uses a placeholder.


Anonymous 12/29/2018 (Sat) 16:24:11 [Preview] No.13061 del
>>13060
Yeah, I think delegating responsibilities to ffmpeg for video is the way to go to begin with, the hard part is supporting formats that don't have turn-key thumbnailers. Difficult as it may be, I think there's a lot of potential for properly hosting other kinds of files, though that blurs the line between imageboards proper and BBSes, I guess.


Anonymous 12/29/2018 (Sat) 16:25:20 [Preview] No.13062 del
>>13059
nntpchan is just a usenet http bridge with an imageboard flavored web ui, the css is godawful as it should be https://2hu-ch.org/


Anonymous 12/29/2018 (Sat) 16:47:02 [Preview] No.13063 del
>>13062
Sometimes I wonder why even if nobody has messed with the 2ch-style UI, by and large everyone keeps the same CSS too. I think arisu does a good job of doing something familiar yet new, but most boards just use the same Yotsuba look and I don't get it.


Anonymous 12/29/2018 (Sat) 17:05:21 [Preview] No.13064 del
(2.89 KB 171x204 1519100873347s.jpg)
>>13056
>uploading text directly as an attachment with the thumbnail generated from the file contents
literally W H Y


Anonymous 12/29/2018 (Sat) 17:14:26 [Preview] No.13067 del
>>13064
Super useful for sharing/uploading code. Especially longer stuff that might be unwieldy pasted in directly, and there's no need to link to a pastebin either.


Anonymous 12/29/2018 (Sat) 18:53:50 [Preview] No.13068 del
>>13062
>nntpchan is just a usenet http bridge with an imageboard flavored web ui
vimscript was designed


Anonymous 12/30/2018 (Sun) 00:52:51 [Preview] No.13070 del
>>13056
>Super easy deployment thanks to Docker
how is docker easier for an anon than "install a list of packages, clone a git repository and run a bash script"

>Great media support
I'm pretty sure lynxchan can be very easily modified to thumbnail arbitrary mimetypes. The reason it doesn't is because thumbnailing arbitrary mimetypes is a great way to get rooted or DoS'd by an ImageMagick exploit. Nobody uses WebP, vichan already has TeX support, and all that other shit is retarded.

>Media storage is handled with S3
<semi-hard dependency on the AWS jew
>markdown support
<when vichan markup is just the subset of markdown that imageboards care about
<when markdown is the offical markup language of redditniggers

>A REST API that supports both reading and writing
What can you do with a REST API that you can't already do with lynxchan's API?


Anonymous 12/30/2018 (Sun) 01:20:13 [Preview] No.13071 del
Not a fan of the catalog as default view.


Anonymous 12/30/2018 (Sun) 01:56:33 [Preview] No.13072 del
>>13070
>how is docker easier for an anon than "install a list of packages, clone a git repository and run a bash script"
Docker has a couple advantages over that more manual approach. The first obvious one is that you're guaranteed a clean environment every time, which is a huge deal for testing. You can do the same thing with lynxchan and other software suites with tools like Vagrant, but you're spinning up an entire VM versus a simple container, with everything that entails.
The other big advantage is a repeatable environment regardless of the host OS. You don't have to worry about which versions of what are installed on the host machine, what the package manager is, or in most cases with Docker what the host OS even is. Maintenance becomes a lot easier too; when I commit new changes to Github pushing them to Futatsu is as easy as building a new Docker container and swapping it for the currently-running one.
The last big advantage that ties in to your security concerns is that damage from a hacked Docker application is limited to the Docker container itself. So no system-wide filesystem access or anything like that.

>The reason it doesn't is because thumbnailing arbitrary mimetypes is a great way to get rooted or DoS'd by an ImageMagick exploit.
Why (and how?) would you use ImageMagick to thumbnail arbitrary MIME types? Do what Maniwani does and let the board admin set what types are acceptable; you don't need to have a large attack surface. In practice I don't think there's much danger of a crafted attachment bringing down your board if you don't do things like call out to a shell to thumbnail. Anything beyond that is something you already have to deal with already - libpng/libjpeg vulnerabilities, etc...
>and all that other shit is retarded.
I definitely think there are a lot of artists who would like others to be able to view their work without needing a 3D modeler installed. We'll see if I'm proven right though, it's just a hunch.

<semi-hard dependency on the AWS jew
S3's API is implemented by several different open-source object stores, like Minio and Ceph. I use Minio to test out S3-dependent stuff with Maniwani, actually - way easier than calling out to AWS. It's not an AWS dependency at all, which is why I chose to have the primary storage backend target S3 in the first place.
<when vichan markup is just the subset of markdown that imageboards care about
I disagree on that, I've seen a decent amount of people explicitly give Maniwani props for the full Markdown support.
>What can you do with a REST API that you can't already do with lynxchan's API?
Nothing really, I like lynxchan's API, though it's definitely an exception to the rule among imageboard engines in that regard. I'd like to point out you've compared Maniwani's capabilities to the union of both vichan and lynxchan's feature set, though.

>>13071
What in particular do you like about the classic 2ch view? I'm assuming that's what you prefer, at least.


Anonymous 12/30/2018 (Sun) 02:23:41 [Preview] No.13073 del
>>13072
Not having to open the full threads to see if I want to reply anyone in them or its just shitpost, or merely to absorb the content in general.

I only ever open threads to see older posts from threads I haven't read before, I even reply from the index (using dollchan and similar tools).


Anonymous 12/30/2018 (Sun) 05:53:43 [Preview] No.13077 del
Probably not the right place to ask, but which engine is the best to create an imageboard and why?


Anonymous 12/30/2018 (Sun) 06:22:46 [Preview] No.13078 del
>>13077
none their all shit or not really an imageboard and still shit
https://tanami.org/overscript/


Anonymous 12/30/2018 (Sun) 16:34:31 [Preview] No.13081 del
>>13073
I'm going to be cutting back pretty hard on shitposting overall on Futatsu, though admittedly that's only a solution for a specific Maniwani installation.
>I only ever open threads to see older posts from threads I haven't read before, I even reply from the index (using dollchan and similar tools).
I'm not sure I've ever heard of anyone else doing this, anon. I'll think about adding a 2ch-style thread preview, but I think overall the problems it solves are mostly better left to moderation.


Anonymous 01/02/2019 (Wed) 15:07:26 [Preview] No.13082 del
I have nothing against trying to adopt a new style for imageboards, but I don't think default bootstrap crap is it. Actually, I wouldn't even mind bootstrap if it was implemented properly, but your site makes my eyes bleed. I thought you would fix it eventually but it's been like a week and it's still the same.

>>13072
>I disagree on that, I've seen a decent amount of people explicitly give Maniwani props for the full Markdown support.
If you think the opinion of people on 4chan /g/ is worth anything, sure.
Also, you making a thread on 4/g/ if the biggest red flag for me. But from your choice of captcha I'm sure that's exactly your target audience.
t. someone barred by recaptcha v3


Anonymous 01/02/2019 (Wed) 15:45:01 [Preview] No.13083 del
>>13082
>but I don't think default bootstrap crap is it.
Fair enough, I don't think stock Bootstrap is that great either.
>I thought you would fix it eventually but it's been like a week and it's still the same.
I'm just one man juggling running the site itself with all the different feature requests different anons give plus obviously other non-imageboard-related responsibilities - I'm no superhero. I'll ideally get to theming by the end of this week. If you look at the commit log I've been actively adding other features people have requested for a while now.
>If you think the opinion of people on 4chan /g/ is worth anything, sure.
Any other boards you think I should pull advice from? One of the really nice things about /g/ is the number of people there means I can get feedback very quickly compared to other places. We can go back and forth about the quality of the feedback, but the turnaround speed remains. I'm quite open to pulling feedback from other places, though (like here).
>But from your choice of captcha I'm sure that's exactly your target audience.
I've noted before that reCAPTCHA was a quick interim solution meant to ward off spamming. After I finish implementing signups, Futatsu won't have any CAPTCHAs to speak of, though I'll keep the reCAPTCHA functionality in Maniwani itself. Sorry to hear reCAPTCHA v3 doesn't like you, though. Try to hang tight for a bit longer and it should be gone.


Anonymous 01/03/2019 (Thu) 15:23:42 [Preview] No.13086 del
>>13078
why are they all dogshit?



Top | Return | Catalog | Post a reply