<p>Using a image matching algorithm (no idea which one or how).
Or, frankly, even a scaled down image with binary closeness algorithm.</p>
<p>Either way, take the best matching picture and use that as the basis to describe the picture.
Unsure how I would go about describing unlimited amount of new memes, but that is to be seen.</p>
<p>Best idea is to get in contact with the guys who make <ahref="https://imgflip.com/memegenerator">this meme generator</a> as most memes there are at least partially described.</p>
<p>These commands are nice, but why is there not a <codeclass="highlighter-rouge">openvpn-client@random</code> so that I can randomize out of the possible files?</p>
<p>I am sure there would be a way to have a systemd service envoked like so:</p>
<h4id="31-add-zhuyin-support-for-epub-with-pinyin">3.1 Add Zhuyin Support for <ahref="https://github.com/TTWNO/epub-with-pinyin">epub-with-pinyin</a></h4>
<p>It currently only supports Pinyin.
At the time, HTML <codeclass="highlighter-rouge"><ruby></code> support was low,
and so even when implemented properly the characters showed at the top.
They should be shown on the left hand side.</p>
<h4id="312-make-epub-with-pinyin-easier-to-use">3.1.2 Make <ahref="https://github.com/TTWNO/epub-with-pinyin">epub-with-pinyin</a> Easier To Use</h4>
<p>Currently it requires setting up Python and some libraries.</p>
<p>Ideally it could be a web application, executing in Javascript.
Storing .epub files on a webserver even temporarily (especially if they are copywrited) is
a legal risk I am not willing to take.</p>
<h4id="32-look-into-linq">3.2 Look Into <ahref="https://www.lingq.com/en/">LinQ</a></h4>
<p>I have heard this system implements something very similar to an idea I had called “Minks”.</p>
<p>The general idea was that you could save words you knew,
thus stopping their pronunciation to appear above them (Pinyin/Zhuyin).</p>
<p>Given my interest in similar things, possibly apply to work for them,
or encourage them to BSD some of their libraries.</p>