switch (parts of) puredata.info to a modern site-generator

[umlaeute]

as mentioned in #15 it might be good idea to switch https://puredata.info away from Plone to something simpler and modern (and fast).

@Lucarda proposed to use github pages (which i opposed in order to not be to dependent on github and instead prefer something self-hosted).

it should be possible to setup something with GitLab pages (hosted at https://git.iem.at), and using some site-generator like hugo that takes markdown and turns it into a nice page

things to keep in mind:

• deken currently relies on webdav uploads to puredata.info (something that we cannot do with a static, generated page)
  • currently all uploads go into the /Members/ subfolder, it should be easy enough to keep this on the plone instance
• the permission model for editing the repository is different (we would somewhat lose the open "wiki"-like permission model; in practice this is probably not a very big deal)
• we cannot really migrate the userbase (so contributors have yet another user to keep track of; otoh git.iem.at allows you to authenticate with your github/gitlab/google users)
• search
• ...

[umlaeute]

@danomatika wrote in #15 (comment)

Also, the plone setup at puredata.info is not ideal but I am sure it can be updated, given an updated/modern theme, and reorganized.

unfortunately in practice the plone setup became immutable (it's about 20 years old; in the meantime plone has changed enough to make any migration unfeasible) there's also a couple of GB of data involved...

[umlaeute]

@Lucarda asked in #15 (comment)

can we host @ IEM?

yes

and make PRs to IEM gitlab?

yes

and avoid using githubpages.

yes

[Lucarda]

☝

I'm totally happy. :)

I'll install HUGO and see how to play with it.

[Lucarda]

I found HUGO a PITA to start something from scratch. May be someone else want to try. Anyway there's a lot of documentation to read and all that that will scare away someone who wants to contribute. (I might be wrong).

I'm proposing a simple way of getting the site updated.

see a preview on: https://pd.iem.sh

to start editing in your machine download:

https://git.iem.at/pd/pd.iem.sh/-/tree/master/

• install pandoc and add it to your path so that a terminal knows about it.

• edit the current .md files.

• run the script.

• open index.html in your browser.

• run the script each time you save or create an .md file (it converts all .md files recursively).

[umlaeute]

hmm, no.

the repository is really meant as a repository that builds the webpage automatically.
no HTML pages there, only markdown and a CI-configuration that "runs the script".

the process of working with hugo is something like:

1. install hugo

2. clone the repository and start the hugo live renderer:

   $ git clone --recurse-submodules https://git.iem.at/pd/pd.iem.sh
   $ cd pd.iem.sh
   $ hugo serve

3. open http://localhost:1313/ in your browser

4. edit/add markdown files

5. whenever you save, the live preview at http://localhost:1313/ will be automatically updated

once you are happy:

6. commit and push
7. done

[Lucarda]

clone the repository and start the hugo live renderer:

I did all this but never liked what I was getting.

I will give it another round trying to implement the custom.css and stuff.

[Lucarda]

@umlaeute get https://git.iem.at/pd/pd.iem.sh/-/tree/custom-css and start hugo server.

Is something like this what we need?

Can people fork on their gihub?

[Lucarda]

ping @umlaeute and @porres

check https://pd.iem.sh/ (still in a draft)

[porres]

check https://pd.iem.sh/ (still in a draft)

This looks great and much better than old puredata.info, which hurts my eyes and is really hard to edit...