The new user flow isn’t smooth. Granted, I may be feeling additional frustration because I had to work through a segfault bug with wget (see http://unix.stackexchange.com/questions/280968/wget-segmentation-fault-core-dumped).
So the first page at www.seafile.com: there’s no Getting Started link anywhere. Try It Now is a tease which doesn’t have a flow to actually get started. But it’s enough of a tease that I want to do it. The closest Getting Started seems like the Download.
From there, I see a bunch of Download links. Before I need a desktop client, I need a server, right? But the Server link comes AFTER the Desktop software downloads.
The Server download says: “No installation needed, just extract and run setup script. See https://manual.seafile.com/”. Yep, that’s right, it doesn’t point to anywhere in particular (or the setup script) in the manual spanning probably hundreds of pages.
So I started browsing through the server manual and eventually ended up at https://manual.seafile.com/deploy/using_sqlite.html but the instructions are not all that clear and are a bit odd. For example, “Click the tarball link and save it” appears on https://manual.seafile.com/deploy/using_sqlite.html. But most people installing a cloud server are probably not using a GUI.
I spun up a DigitalOcean droplet and followed instructions anyway and eventually got it running, but then I found https://github.com/haiwen/seafile-server-installer! Which I would’ve loved to see on the Downloads page to begin with, or something similarly quick.
You also shouldn’t have your quick script on a Github page. I’m a developer but when I see something on a Github page, I usually assume it’ll be painful to deal with as there are so many outdated open-source Github pages suffering from bitrot. Whereas if something is on your own site, I figure it’s probably better maintained.