<-- back to the mailing list

[SPEC] Backwards-compatible metadata in Gemini

Baschdel baschdel at disroot.org

Tue Feb 23 16:16:17 GMT 2021

- - - - - - - - - - - - - - - - - - - 

On 22.02.21 21:16, Bradley D. Thornton wrote:

On 2/21/2021 4:46 PM, Oliver Simmons wrote:
On Sun, 21 Feb 2021 at 20:21, easrng <easrng at gmail.com> wrote:
[snip]
I'm not going to weigh in here on the virtues of this, but wouldn't the
following be more applicable?
# My Super cool gemsite
```data:text/gemini-metadata+json;charset=ut.......
Welcome!
```

Ok I think have to throw in my two ct here …

Please don't do fancy data link black magic in gemtext.

The reason we (at lest I) like gemtext is that it's easy to read and write, and that for both humans and machines. Putting metadata behind two or three layers of encoding and/or hiding it in elements that were not intended to store page metadata won't make it easier to parse it (for both)!

I personally like Oliver's approach, that makes the metadata both machine and human readable and forces it to the end of the document, so that it stays out of the way on clients that don't have support for hiding it.

Meta about metadata

Of course we have to standardize some keys and value types to make the metadata actually useful for machines, but in a way that is human readable/writable and easy to parse.No "anything that <command> can parse" is not a good standard at all,if that approach had been used for gemtext it would have turned into a mess pretty quickly and nobody would care.

Greetings- Baschdel

^^^version: 1date-written: 21-02-23 11:09 +0100author: Baschdel=

mailto:baschdel at disroot.org author contact E-Mail=
mailto:gemini at lists.orbitalfox.eu?subject=… reply contactstyle: E-Mail message=
https://lists.orbitalfox.eu/archives/gemini/2021/005420.html answerto=
https://lists.orbitalfox.eu/archives/gemini/2021/005522.html mirror