Chris (now @chris@toot.works) is a user on mastodon.social. You can follow them or interact with them if you have an account anywhere in the fediverse. If you don't, you can sign up here.
1227341459761592

Chris (now @chris@toot.works) @cs

I've moved accounts to @chris. Follow me there to get new toots :)

<p><span class="h-card"><a href="https://misanthropy.wang/@jemus42">@<span>jemus42</span></a></span> [en] Oh, no, I'm held captive in a server and must translate mischief for other people. Please send help or cake.</p>

@FerdiZ Looks like that is a web tool for managing manual translations done by humans for projects. It doesn't perform machine translation.

<p>Trying out my account on my own instance as a second main (in addition to <span class="h-card"><a href="https://mastodon.social/@cs">@<span>cs</span></a></span>). I just imported my follows over from .social, so sorry for the notification spam ☺</p>

@bea @chrismartin @compthink Haha, clever! 🤔 New rule: you only get X bio edits per month and no links are allowed

@animeirl Woolly? Probably already used by a bunch of other things :/

@translator @dyedgreen You used it correctly, I think! It usually repeats the language code in the translation, but only once.

It's using a paid Microsoft Translation API.

@translator @dyedgreen The bot repeating "[en]" twice here is a bug :) I will fix it later tonight @translator de

@mulander @Divan Woahhhh. This is amazing. How did I not find this before?!?

@ashkitten I've been meaning to do this as well. Were you planning on self-hosting GitLab or using gitlab.com?

@compthink @bea So you get all the toots (x * num. of days) up front or do you accrue more each day?

@xdej @translator Don't think so :/ I might be able to add it it in after the translation, if there's a decent library for it.

@MyBroodingHeart @Wulf So cool! How does it handle users in different timezones?

@zuphzuph @munin @hugo Drive safe! I think you could reduce RAM usage by having them instances all share Postgres/Redis, assuming those two are the main RAM hogs.

@zuphzuph @munin @hugo At that scale though (15k users, 30 instances), it might be cheaper to have one VM with ~8-16 GB RAM, 2-4 cores just run the web/Rails process(es) and then use cloud services for Postgres/Redis.

@hugo @munin @zuphzuph 500 users * 30 instances = 15000 users. A single $120 USD/mo. server* could probably run that. €15/mo * 30 instances = ~$481 USD, so you would be saving ~$361 USD/mo or €337.71/mo.

* DigitalOcean has a 16 GB RAM, 2 core VM for $120/mo. 32 GB RAM, 4 CPU for $240/mo.