2024-11-17 external brain thinking

from Diary 2024-11-17

There's also a feeling that automatic English translation might have stopped. So it's not that the translation has stopped, but that the import is mocking it up.

Based on "Yasukazu Nishio's Scrapbox -> Yasukazu Nishio's External Brain", there is a growing sense of "Maybe I should quit importing to Scrapbox".

look back

. Then I remembered mem.nhiro.org.

• but it doesn't have the data, it just hits Scrapbox's API when accessing it and renders it.
• Good features:.
  • Custom Domain Delivery with Vercel
  • Free UI design
• Bad features:.
  • Data is acquired via Scrapbox's API without a DB, so processing using data that cannot be obtained from Scrapbox is not possible.
• Read. - What is mem? - > 2022-02-24 I've decided to call it mem because it's tedious to say every time "based on Scrapbox content, customized views, and delivered using Vercel on a custom domain".
  • 2022-03-08
    • Issues/needs of crossing multiple projects
  • Next-mem
    • Needs to span multiple services
• https://github.com/nishio/mem

I put etude-github-actions together exactly one year ago (I started using it around April).

• 2023-11-17 I didn't know anything about Github Actions, so I made it as a practice, but the experiment automatic translation of Scrapbox with Github Actions went better than I expected, so it went straight to the production phase.

• Automatic English translation.

• Good features

  • You're straddling the language barrier.

• Bad Characteristics

  • Difficult to indicate in the UI that the view is a translation since it is a Scrapbox

• https://github.com/nishio/etude-github-actions

• Also related to [Vector Search in Nishio

• The data is automatically taken and EMBEDDED.

• Experimental search across multiple projects.

• It should be possible to do extract dense masses from the EMBEDDING data as we did the other day

• https://github.com/nishio/vecsearch-service

  • Is this PRIVATE?

• embedding side

  • https://github.com/nishio/omni
  • I created omoikane-embed and then applied it to /nishio.
  • from omni 2023-08-31

  • I have installed the Omoikane Embed system, which automatically makes Omoikane Project into an embedded vector, in this Scrapbox project/nishio.

• Then nihia was born, right?

  • from pNihia 2024-05-05

    • ⿻Plurality Assistant was useful to make GPTs for this Scrapbox reading assistant.

    • Server side is in private repos: nishio/vecsearch-service.

• So Late May 2024, a turbulent time. started and now w

Perimeter Search

• Obsidian Vault at /villagepump/Quartz.

  • static rendering

• [/villagepump/JSON Canvas](https://scrapbox.io/villagepump/JSON Canvas)

  • Obsidian Canvas

• [/villagepump/Bluemo Obsidian vs Scrapbox consideration 202301](https://scrapbox.io/villagepump/Bluemo Obsidian vs Scrapbox consideration 202301)

  • Three points of contention are scalability, communication with others, and website publication.

    • Good organization, I was thinking to myself as I wrote this far, "You have multiple needs mixed up that need to be resolved.
    • My needs are not necessarily equal to this, so I have to think it through myself properly.

• Consideration Forked [/villagepump/Bluemo Obsidian vs Scrapbox study 202301](https://scrapbox.io/villagepump/Bluemo Obsidian vs Scrapbox study 202301) and created [/villagepump/view addition study 202411](https://scrapbox.io/villagepump/view addition study 202411). After some thought.

• Obsidian or Scrapbox is a false choice

• Since "flexible in expansion" is vague, make it a little clearer what kind of expansion you want to do.

• I want to extend the View.

  • Backend can remain Scrapbox.
    • Scrapbox can continue to be used for the Japanese real-time part.
  • The conversion delay is not a problem because it is a replacement for what was originally translated into English and published at a rate of once a day.
    • I've been importing into Scrapbox.
      • But then, for those for whom English is not their native language in the first place, it would simply be "a site that for some reason cannot be machine translated"?
      • Frustration with not being able to add to the UI that it is a translation.
      • I'm feeling like I don't want to rely on the "rewrite services without a write API by importing" hack anymore, since imports are now mocking me up.
  • In other words I want a place to host things after they are automatically translated into English.
    • I'm thinking that the same mechanism could be used to host the Japanese version, and that new features could be added along the way.
      • The higher the level of abstraction, the wider the range of application.
      • The problem that because there are so many things we want to do, we have to raise the level of abstraction to be able to do them all, and as a result, we get too far away from the specific requirements, which reduces the clues for decision making.
    • Should we limit ourselves to English hosts for once?
      • There is a lot to be learned from actually doing it.

• Unknown if the Obsidian path should be followed in replacing View

• We're talking about Quartz two years ago at this time.

  • V4 was released on 2023-08-20 and the configuration has changed significantly

  • Hugo to Preact.

  • I wasn't used to writing Hugo templates, so being able to customize the front end with JSX would be helpful.

    • [Obsidian published in Quartz v4 https://ikorihn.github.io/digitalgarden/blog/Quartz-v4%E3%81%A7Obsidian%E3%82%92%E5%85%AC%E9%96%8B%E3%81%97 %E3%81%9F]
  • Whoa, you can customize it with JSX, React-like.

next action

.

• Read back https://github.com/nishio/etude-github-actions
• Create a new one with external-brain or something.
  • Save JSON
• Create Japanese-English Obsidian format with [ScrapboxToObsidian
• Quartz v4 to Github Pages
  • (/nishio-en and mem.nhiro.org are operated in parallel without erasing once)

imagination

• /nishio-en will not be updated as it is, and when it works with GitHub Pages, I'll pin the "see here" page and be done with it.
• MEM is another story.
  • I'd like to see them join forces in the long run, but not at this point.
  • Interesting things to do with embedding data could be derived from omni.

This page is auto-translated from /nishio/2024-11-17外部脳考える using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.