Fwd: Re: Fwd: Re: Text reflow woes (or: I want bullets back!)y

[James Tomasino]

On September 2, 2019 12:20:24 PM UTC, Jason McBrayer <jmcbray at carcosa.net> wrote:
>But separating these issues: maybe we should just specify Gemini maps
>as
>max 70 columns of monospaced, unflowed text. If people want better
>typography, they can send text/markdown.

If we assume that mime-types other than text/gemini are rendered according to their own rules (eg, markdown) then the issue of wrapping is left to text/gemini and any types that don't have inherent rules (eg, text/plain). I think we have just a few choices:

1. Display as-is authored.
2. Wrap everything
3. Wrap only long lines (need to decide on arbitrary width)
4. Define markup to be explicit about wrapping vs. fixed.

#1 is the easiest to implement but leaves the same problem as gopher, and it's likely that some client authors will go rogue and implement #2 or #3 anyway. #4 adds complexity and opens a door for more and more markup.

What it we look at combination mime-types, though? What if we allowed something to be defined as a Gemini map & in markdown, or asciidoc, etc? Parsing mine types to indicate "this is a Gemini file" first gives you links, and if the secondary type is known to the client can give you formatting. It can fall back on "display as-is" for unknown types.

Now this doesn't necessarily solve text/plain if folks are determined to wrap it, but I think it makes Gemini map files more flexible over time without adding burden to the spec directly.