2025-09-02 Feeds

I am once again thinking of the feed as a generic single-file export format (RSS or Atom, it doesn't matter to me). Does that make sense? Or would it make more sense to just zip HTML and images? The only benefit of using a feed is that the format is standardised, I guess.

Feeds as an export format only if the feed items contain images as attachments or as data URIs 😨 I guess? I mean, feeds are good enough to feed the text of pages into a search engine like Xobaque, but if the feed is supposed to act as real export format, the other assets are also important.

Xobaque

And what happens if an image is used by two pages? The feed format would mean that both pages need to declare it as an attachment, no? So I'm guessing, the feed as a universal import and export mechanism is just not good enough.

Nevertheless, Oddμ has two feed exports on the command-line:

oddmu-export(1)
oddmu-feed(1)
Emacs
online feed

And yet, I think the *archive* action (not enabled on this site) is the more useful one, in general. Except the archive contains the Markdown files, not the HTML files. So you still need to run oddmu-static(1) to generate the static site.

oddmu-static(1)

Lately, I've been writing a shell script to take an HTML export from Oddmuse sites and convert them to Oddμ sites. This requires a conversion from HTML to Markdown (getting rid of the quirky wiki markup). I'm using html2markdown even though it's "sponsored by 🔥 Firecrawl, where you can scrape any website and turn it into AI friendly markdown with one API call." 🤮 There are two alternatives written in Python: html2text translates HTML to text that happens to be Markdown; html2markdown does that and preserves untranslated HTML. Something to keep in mind!

Oddmuse
Oddμ
html2markdown
html2text
html2markdown

#Oddμ ​#Xobaque

Your export file doesn’t actually contain your media library (like photos, videos, audio, etc.). -- Export your website’s content
Export your website’s content