Ivan

@notegone

9 Following

28 Followers

Matthew pfp

First screenshot essay ever: "Where Meetup went wrong" Please let me know what you think 😃

2 replies

1 recast

7 reactions

Ivan pfp

I would like to use someone’s unmetered 5 or 10 gigabit fiber for a startup I’m working on (the economics of buying my own storage devices are far better) and I’m willing to do your software design/database/sysadmin labor or take on debt.

0 reply

0 recast

0 reaction

Ivan pfp

I have a 2TB PostgreSQL database with 3 billion rows used to track content; I use periods so that I can UPDATE without losing old data, and I use ephemeralpg for unit tests and ad-hoc experimentation to avoid risking a production server. I have no real alternatives to PostgreSQL so I am glad that it is pretty adequate.

1 reply

0 recast

0 reaction

Ivan pfp

the point is not to hold myself to everything in the initial plan, but rather to have an adequate plan at all. also, goodreads “want to read” is not good enough: it accumulates too much stuff and you can’t assign yourself part of a book.

0 reply

0 recast

0 reaction

Ivan pfp

https://gist.github.com/ivan/5095670735ba941a6090a69fce4183df my 2023 reading list. I have found this time-intensive planning process to be quite worthwhile, causing me to drop many texts which I don’t really need. If you are an autodidact you basically have to spend time figuring out what to assign yourself.

1 reply

0 recast

1 reaction

Ivan pfp

I spent almost a week triaging my backlog of 1000+ things to read (mostly ignored, many sampled) and thought about whether I would benefit from reading each thing this year, to arrive at a list of about 50 texts, some short; some made shorter by assigning myself only part of a book. I have found the planning worthwhile

0 reply

0 recast

0 reaction

Ivan pfp

unbook was really only designed to work with ebook formats like mobi/epub that contain text and good markup. not sure what your best bet for djvu would be; it depends on what you want? perhaps djvu2pdf + ocrmypdf + pdf.js

1 reply

0 recast

0 reaction

Ivan pfp

hierarchy of trying to fix your web browsing experience: a per-site setting a uBlock Origin filter a userscript a Manifest V2/V3 extension a Chromium patchset a new browser but existing engine a whole new browser and engine

0 reply

0 recast

1 reaction

Ivan pfp

zsh-autosuggestions btdu periods (for PostgreSQL) ephemeral_pg

1 reply

0 recast

1 reaction

Ivan pfp

2) the MutationObserver is reinstalled after an SPA navigation (because there might be new elements to click on). This currently relies on the 'navigate' event which is still Chromium-only (https://caniuse.com/mdn-api_navigation), but I am open to contributions to make this work elsewhere.

0 reply

0 recast

0 reaction

Ivan pfp

This is implemented using MutationObserver and querySelectorAll to watch for certain elements that should be clicked on, but crucially: 1) the MutationObserver is disconnected after e.g. 100 mutations to avoid slowing things down, and

1 reply

0 recast

0 reaction

Ivan pfp

I have released a userscript called Expand Everything (https://github.com/ludios/expand-everything) which clicks on all the "show more" links on a page, so that you have a lot more of the page text available for searching, copying, or archiving. It supports these sites right now: https://i.imgur.com/6CwCCvu.jpg

1 reply

0 recast

0 reaction

Ivan pfp

unbook 0.7.2 is now fixed for macOS, and all platforms have improved error reporting that is less confusing when e.g. calibre ebook-convert is not found. There are improved install instructions on https://unbook.ludios.org

1 reply

0 recast

1 reaction

Ivan pfp

The inlining of images was inspired by SingleFile, which convinced me that long data URIs and large HTML files are generally no problem. The rewriting of image URLs to data: URIs is handled by Cloudflare's very helpful lol-html library: https://github.com/cloudflare/lol-html

1 reply

0 recast

0 reaction

Ivan pfp

But HTMLZ output was created for editing, not for reading (in 2011: https://mobileread.com/forums/showthread.php?t=128076…). Browsers can't read zipped HTML; I kept extracting .htmlz files and fixing styling issues like the awful `text-align: justify`, all of which unbook now handles. https://i.imgur.com/eeigZJx.jpg

1 reply

0 recast

0 reaction

Ivan pfp

unbook was created after I repeatedly noticed that I had a better reading experience with Calibre's HTMLZ output in Chrome or iOS Safari than with any ebook reader software. Such programs often have peculiar UX issues which are not always outweighed by whatever benefits they offer.

1 reply

0 recast

0 reaction

Ivan pfp

I am pleased to announce unbook, which solves the DRM-free .{epub, mobi, azw3, ...} ebook reading problem by converting the book to a self-contained HTML file with all the images inlined using data: URIs. You can now just read everything in your browser. https://unbook.ludios.org https://i.imgur.com/uftndd3.jpg

1 reply

0 recast

0 reaction

Ivan pfp

Stanisław Lem’s _Solaris_ David McCullough’s _The Wright Brothers_ William Langewiesche’s _The Outlaw Sea_ Elif Batuman’s _The Idiot_ Hamilton Helmer’s _7 Powers_

0 reply

0 recast

0 reaction

Ivan pfp

gist.github.com/ivan/caea21522a563d1413a32b5a95f54638 a combined list of filters that I put into uBlock Origin's 'My filters' to fix website typography problems and other annoyances

0 reply

0 recast

0 reaction

Ivan pfp

https://github.com/ludios/useful-bookmarklets a collection of bookmarklets I use frequently: kill sticky, replace all font-family, override all colors, inject width=device-width, invert all images, set a background-color for images with a transparent background (bookmarklets can indeed be used on mobile too.)

0 reply

0 recast

0 reaction