💾 Archived View for bbs.geminispace.org › u › decant › 17507 captured on 2024-07-09 at 06:07:08. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2024-06-20)
-=-=-=-=-=-=-
May 31 · 6 weeks ago · 👍 mozz, Addison, agh, Queen_City_Nerd
😎 decant [OP] · May 31 at 07:08:
ok.. so convert the image to webp, now it weights 48kb. But webp is nowhere near as universal as png. Another point: security risk of image viewer getting exploit by crafted image, I think the risk still is somewhere below pdf ?
🕹️ skyjake [...] · May 31 at 08:23:
— Image Over Wire Zine: A one off experiment (text extracted from image)
With modern OCR technology, publishing text documents as images isn't as silly as it might at first appear. Of course, there needs to be a compelling reason to use a format that's an order of magnitude larger than the text content alone (the text is about 2.3 KB as UTF-8).
The link above is the text extracted using Apple's built-in text-from-image feature, which is based on machine learning ("AI"). I added the cat emoji manually. 🙂
Publishing content in multiple alternative formats is a great solution if a particular one feels too restrictive. The common image formats like JPG and PNG are universally viewable practically everywhere, although I'd still opt for PDF personally as it supports text natively.
Beyond the additional size incurred by using an image format over a text is accessability. As @skyjake pointed out OCR is an option for text extraction which can obviously be used in combination with TTS of some description. People without vision issues may prefer not to consume this way though and the problems here are that during image creation the author chose a specific font and dimensions. One dimension issue is that you may need to pan around the content. One issue related to font that springs to mind is that there are specific fonts to aid those with dyslexia. A client for pure text can auto-wrap content based on the font and display and can be configured to use a font or fontset of the users choosing.
I have to ask why? Why take a format specifically designed not to do something, and come up with a very convoluted way to do it badly? If there was a need for it, we would have it by now. Conversely not having it implies the lack of need or desire for things beyond text or minimal markup.
I like to save webpage I like using the print to pdf function of my browser.
Is there a way to render the pdf as a continuous scroll as opposed to standard size of paper? pandoc?
😎 decant [OP] · Jun 01 at 08:28:
@stack, I would need to think about that. But one of the reason I could think of is math notations. I could parse simple latex in my head, but still, It's nicer to just see it. Of course, I could post latex source code, but the latex software is complex. Think of image over gemini not as pure text++ but pdf--. Each format will surely have their own best use cases.
@drh3xx I'm dyslexic myself, the post was edited in libreoffice in OpenDyslexcis font, I should have made sure the font is copied over to GIMP as well. I should write something on my struggle with the education system.
@skyjake I think if there is a way to turn off interpolation in image views, one could make small
😎 decant [OP] · Jun 01 at 08:33:
image file filled with text look sharp. then again, if the image is too small pictures will not look good. there is no point in using image for pure text.
I would like to thank everyone for there input. reading you replies, I feel like hearing myself talking but from different vantage points.
Math is a good use for that, as well as various illustrations. I am against abusing images to express 'rich text', the kind specifically not supported by Gemini. Some of us are here specifically to avoid that, and be in control of how our text is rendered for us.
