Sutter publishes to both Gemini and the Web.
I am able to link internally to pages on my website inside a Gemini file, and Sutter will interpret that internal link as an https:// link from the Web version, and a gemini:// link from the Gemini capsule. For example, in a .gmi file I can link to the archive like so:
=> archive
This results in a link to archive.html on the Web and a link to archive.gmi on Gemini.
The alternative would be to link to the same page twice, if you did not have this feature, like so:
This blog is available both on the World Wide Web and on Gemini:
=> https://blog.ungovernable.love Web site
=> gemini://blog.ungovernable.love Gemini capsule
This is rather cumbersome both on the web and in Gemini.
Proposed solution
I propose writing external links with the protocol stripped off the front, and having Sutter interpret them with the correct protocol:
=> //geminiprotocol.net The official Project Gemini website
I'm not the first person to make this proposal: https://admin.flounder.online/gemini_text_guide.gmi
For links to content that is hosted at the same URL on Gemini and HTTP(S), such as flounder.online links, I recommend using schemaless URLs such as //alex.flounder.online as they will properly show up on both gemini and http pages.
If this is in fact the intended behavior, then this should be a bug report instead. Right now external schemaless links also add the .html and .gmi to every link, which is not appropriate for many websites.
=> //geminiprotocol.net without a trailing slash results in https://geminiprotocol.net.html/ on WWW, and gemini://geminiprotocol.net.gmi/ on Gemini.=> //geminiprotocol.net/ with a trailing slash results in https://geminiprotocol.net/.html on WWW and gemini://geminiprotocol.net/.gmi on Gemini.
None of those links work.
I'm open to alternate implementations, but I think something along the lines of schemaless URLs would be warranted.
(削除) In case this wasn't clear, you can in fact check out my Sutter page displaying the issues discussed above: (削除ここまで)
EDIT: You can no longer see my pages displaying these issues because my web host went out of business and I'm not interested in trying to run Sutter myself while this issue is unsolved.
Sutter publishes to both Gemini and the Web.
I am able to link internally to pages on my website inside a Gemini file, and Sutter will interpret that internal link as an `https://` link from the Web version, and a `gemini://` link from the Gemini capsule. For example, in a `.gmi` file I can link to the archive like so:
```gemtext
=> archive
```
This results in a link to `archive.html` on the Web and a link to `archive.gmi` on Gemini.
The alternative would be to link to the same page twice, if you did not have this feature, like so:
```gemtext
This blog is available both on the World Wide Web and on Gemini:
=> https://blog.ungovernable.love Web site
=> gemini://blog.ungovernable.love Gemini capsule
```
This is rather cumbersome both on the web and in Gemini.
## Proposed solution
I propose writing external links with the protocol stripped off the front, and having Sutter interpret them with the correct protocol:
```
=> //geminiprotocol.net The official Project Gemini website
```
I'm not the first person to make this proposal: https://admin.flounder.online/gemini_text_guide.gmi
> For links to content that is hosted at the same URL on Gemini and HTTP(S), such as flounder.online links, I recommend using schemaless URLs such as //alex.flounder.online as they will properly show up on both gemini and http pages.
If this is in fact the intended behavior, then this should be a bug report instead. Right now external schemaless links also add the .html and .gmi to every link, which is not appropriate for many websites.
- `=> //geminiprotocol.net` *without* a trailing slash results in `https://geminiprotocol.net.html/` on WWW, and `gemini://geminiprotocol.net.gmi/` on Gemini.
- `=> //geminiprotocol.net/` *with* a trailing slash results in `https://geminiprotocol.net/.html` on WWW and `gemini://geminiprotocol.net/.gmi` on Gemini.
None of those links work.
I'm open to alternate implementations, but I think something along the lines of schemaless URLs would be warranted.
~~In case this wasn't clear, you can in fact check out my Sutter page displaying the issues discussed above:~~
EDIT: You can no longer see my pages displaying these issues because my web host went out of business and I'm not interested in trying to run Sutter myself while this issue is unsolved.