If you ever want to reward me for a good job, then I have several ways you can help. https://kissu.moe/donations
This page has donations as well as some affiliate programs which provide me commissions on sale.
Donation page updated for more professionalism.
I'm working on a premium statistics feature.
1 year left on the primary-haiji servers lease.
2 years left on the other-luna.
In this now unlocked thread I'll discuss Kissu's financial plans and my justifications for them.
Kissu has two servers. One purchased for 4taba and Kissu utilities(luna). The other hosts kissu's software(haiji). Both share the file space equally on luna.kissu(small files) and haiji.kissu(large files and thumbnails).
Kissu faces two issues with our current setup. Periods where all server resources are used up on the haiji server leading to 5second+ post times(a multi-factor problem consisting of the operations performed by posting and background operations done to automatically update the site's information). Given that we don't get a post a minute yet it's not a terrible problem, but it arguably prevents people from wanting to use the site for chatroom like functions.
On both servers we constantly teeter on the edge of using up 40GB of data. This 40GB of data can be eternally worked with. It's not like we will run out of space any time soon. I can always backup then reset the server to remove pointless programs I added and lost track of. Most of it is likely being wasted from my experiences learning how to work with VPS and Linux servers.
For anyone wondering: 4taba = 6.4G, luna.kissu & services = 4.8G , haiji.kissu = 13G :: total = 24G . Libraries and other unknowns = 16GB . So there's a lot of "things" that are taking up disk space that I have not investigated
What I'm saying is In the current state of Kissu nothing needs to change except perhaps who provides the server, VPS vs Cloud etc.
However there is more to the decision than the current state. I have accumulated from the community 911.96USD(April 2020 -> January 2022). If I do a rough average, that's 1800USD per 4 years or 450USD per year.
To renew both servers and domains for 4 years with our current provider would be something like 419.20USD leaving 1300USD. This level of support warrants an upgrade. However, I believe I deserve some money for my effort put into the technical aspects of managing the site.
I want to propose a cut I'll give to myself from the sum of donation funds and a budget for server purchases. I'm still working out the details, but lets say that I decide that haiji is going to be a 2GB/40SSD/2Core (3rd party evaluates it as a 1.6x increase in performance for 2.2x the cost) and luna will stay the same, then that's 619USD allocated to the site, and 1181USD to me (34% site budget, 66% my cut).
There are likely some more affordable mid range VPS providers than what we have or maybe I want to switch to Digital Ocean and go for cloud hosting. Our main server expires December 1st. There's a lot of time to spend on this question.
I would like the community to give me options to explore. Concerns about how much is going to the site. Offer other suggestions on what a good use of the money would be and etc.
There are many good(and lets not ignore wasteful) ways to put funding to use and I can't say I know all of them.
Paying yourself is fair.
Are there any other imageboards that use cloud hosting? Have you checked how well it works out for them?
Sounds fine by me. All I would ever be concerned about would be the transparency of where funds are going, but frankly I've never cared so long as the site isn't in jeopardy of not meeting funding requirements to stay online. Regardless, for being combined admin and developer, the time put into working on the site is in my opinion perfect reason enough for you to get a cut anyways.
>Periods where all server resources are used up on the haiji server leading to 5second+ post times
Does this account for the type of post? As in, the difference between posting with an image vs just text. From a user's perspective at least, I've never noticed particularly long post times outside of uploading larger files at times, but that's to be expected.
No, I don't want that. He pays for servers and does the technical stuff. I just post as I would normally
The two others have their own form of payment for doing what they do and it might be more fair to say that they are the biggest donors and shareholders.
While I'm executive I'm an employee more to Kissu than a member of the site.
This sort of hierarchy is common in community clubs such as sailing where the manager(myself) handles technical operations and pays other people to perform skills while the community forms itself in the space created by management. Often club fees are required which are allocated based on various people, but it's assumed that the volunteers are in fact the people who have spent the most on the club (decade long memberships)
So while I have offered cool some money, he has declined and berun provides me money and is one of the main reasons I could weasle my way into an excuse not to jump into a 9-5
If my reinstalls don't free enough space then I'll start planning for a server upgrade.>>9026
Digital ocean isn't uncommon, lots of websites use Amazon Web Services to get the distributed processing power of Google for the small business. It's not about what other people use though. Cloud hosting offers you to adjust the harddrive or CPU space with a purchase(modularity) whereas on a VPS I need to get a new server with more resource allocations. Thing about Cloud is I've never used it so I'm not sure about what advanced issues I would encounter. The software that I've written has been for a VPS and tested on an OS that behaves like it. Server hosting websites sing praise about cloud and it's hard to find objective pro-vs con reviews that are not written by "influencers". They obviously will not tell you it's failings without you discovering that on your own. Probably the main issue is consistency of user experience. When one server is performing poorly other users will get shifted onto other servers automatically without any control over this. Cloudflare does this to some extent already and at times you begin to wish it didn't exist and you had more control.
If Kissu were to ever reach the levels of popularity of the major imageboards then we would have to either go from a small cloud provider into AWS which would be very trendy and make the site modern. If we stick with VPS then eventually up to a dedicated provider like 4chan does, discrete and stable.>>9028>Does this account for the type of post?
Yes, it's quite variable. Making a post on 4chan is generally always going to be faster than kissu. I'd say we're probably one of the slowest to post imageboards.
>>9069>I'd say we're probably one of the slowest to post imageboards
The bulk of the background problem >>8993
sounds like it could be solved with a beefier server, always, but have you considered changing the imageboard software?
I see, from the main page, kissu is using vichan https://github.com/ECHibiki/Kissu-Vi
And you've(?) invested considerable time into modding it for your purposes, which is great.
I'm a big proponent of 'if it works don't fix it' always, but with your concerns listed (and despite the time already invested into this vichan fork) have you thought of exploring different imageboard implementations altogether?
With this bottlenecking>multi-factor problem consisting of the operations performed by posting and background operations done to automatically update the site's information
I would be curious if some software rewriting could fix it
Yeah, I'm going to pick a VPS with double the resources, but probably a 2 year plan.>rewrite
yes, but I'd need someone to provide me for a year and I don't think it's likely going to happen. Perhaps I could begin some drafts on how to do it so that if I ever have the right situation it will be a faster process. IDK if anyone really gives a damn about what sort of language a server is running on or what it looks like, they just want to post and me nerding out about it is probably a selfish venture.
Vichan is slow, but the amount of things that are happening is what makes it slow. Switched to a split server idea (vichan handling the posts and Hazuki[Golang API and etc. server] ) to give myself a starting point for writing my own. But not too sure if I would actually use the code in an imageboard engine project.
It seems the bulk of your demographic wouldn't care
but I think it's still something that's interesting and worth
It's a worthy venture. The easier the code, the easier sysops, I think.
Things like JS or Rust - not too much of a fan of software with more than 10 dependencies.>yes, but I'd need someone to provide me for a year and I don't think it's likely going to happen
What are your current ventures for money?
It is quite unlikely anyone would pay for a rewrite...
Frankly I'm curious how you get people to donate in the first place.
yes, rust core running on top of golang would be my ideal situation. Rust-Golang bindings
are something worthwhile. I'd probably experiment the idea on 4taba first and then refine it for kissu.
You have to figure out the answers to the other questions yourself. I don't casually discuss that.
Rather, Rust as the core of an engine performing tasks such as thumbnail generation, connection pooling, data storage, template generation and having anything else do tasks such as API writing, routing and scripting (Python, Golang, Lua or perhaps multiple instance of Node) would make it one of the most efficient imageboard software if done right.
I only asked because the boldness in such a statement confused me – guess you just get donations through up-selling features.
Just the thought of having someone pay you to rewrite an imageboard is a foreign idea to me
Are you aware of this?https://github.com/bakape/meguca
Kissu borrowed its captcha code, so I assume yes.
Oh, I've yet to get a captcha. My mistake
>>9069>I'd say we're probably one of the slowest to post imageboards.
we take it easy here
>>9028>transparency of where funds are going
We're looking at anywhere from 20-40% reinvestment into the site. In USD our per month upgrade in server spending will be from 3.95 * 2 to 10.00 consolidating our two servers into one big one. Kissu doesn't require funds for services such as backups or CDN since I provide my own hardware and software or use free services.
This means that funding beyond this level(currently 6.00 - 17.00) will have to be put back into it in some meaningful way. This reinvestment plan should begin this month in some form as either stockpiling for later or using outright. Possibly as upgrades to our graphics removing our font-awesome effects for designs or gifting games to community members. Anything to make Kissu a more enjoyable place.
This remainder 60-80% (currently 50.00 - 67.00) going to me pays my internet bill and food. If I don't eat I will die. If I have no internet I can't post.
Banners.kissu.moe is back up and now has a feature for funders
There was a feature that I was working on, rich text, which would let you see the styles and colors as you type your message. A live preview basically. I discontinued it because it was completely absurd in difficulty to get to work on all platforms and be bug free.
I'm contemplating getting this feature working again and making it available for donating users.
A private userscript being given out which will override the existing QR box and append options to the options menu.
In turn for the user who doesn't want to pay, this will begin to create an API for the Kissu-UI that allows for other people to attach userscripts and override various features.