đž Archived View for bbs.geminispace.org âē s âē Lapidary âē 13985 captured on 2024-08-18 at 19:22:23. Gemini links have been rewritten to link to archived content
âŦ ī¸ Previous capture (2024-07-09)
-=-=-=-=-=-=-
Hi geminispace. I'm full time gemcutter and general lapidary with a passion for the art. My goal is to make as much material about it available on geminispace as possible, and to rescue aging materials for digital preservation.
Would love to meet other gemini users who work in this field or to answer questions for anyone curious to learn more.
Jan 16 ¡ 7 months ago ¡ đ skyjake, gritty, mifuyne
đ¤ nikhotmsk ¡ 2024-01-16 at 23:19:
Gemini is not suitable for long storage. It suffers from bit rot just like the normal web. But I still appreciate the effort. A paper (pdf) book would be a better option.
đ istvan [OP/mod] ¡ 2024-01-17 at 00:29:
Where old texts are available for scanning or PDFs are available, I've already been sending these to the Internet Archive.
Unfortunately, IA is terrible for deep text search, and OCR technology is very bad on many of the old documents. Breaking scanned articles to single files and cleaning up the text is part of my project.
đšī¸ skyjake [...] ¡ 2024-01-17 at 07:26:
@nikhotmsk:
Gemini is not suitable for long storage.
Gemini, as a technology, does not address storage in any way. It's up to you to decide how you want to store your data. Gemini's emphasis on plain text documents is a big long-term advantage, though.
Storing things long-term and ensuring data remains available on the internet either requires the traditional server hardware, software and domain registration upkeep, storage redundancy, backups, etc., or a massively parallel solution, something like a peer-to-peer network or a distributed version control system. Gemini can regardless be used as the interface through which data from those is accessed.
đ istvan [OP/mod] ¡ 2024-01-17 at 16:30:
Yes. I think directory of plain text is much easier to store and preserve and search than OCRs of aging PDF. And eventually will be better input for AI machine translation.