• 4 Posts
  • 356 Comments
Joined 3 years ago
cake
Cake day: June 22nd, 2023

help-circle








  • The webview isn’t supposed to show the bloaty page directly. Your client would retrieve the bloaty HTML, strip the junk out or maybe completely reformat the page, and then show the “purified” page in a browser. I did this for a while with local proxy cgi’s but they eventually stopped working due to AI bros over-scraping everything. Also, are you going to write custom de-bloating for different sites that your client reads? It helps quite a bit. I can send you some of mine if you want.

    Yes a normal web page can’t pull arbitrary feeds but I had thought there were permissions you could give to enable that. Like I think Manifest v2 lets you turn off the cross origin check, so you could write a Firefox extension. But, they shut down Manifest v2 in Chrome in order to prevent ad blockers from working.


  • We already have lite.cnn.com and there must be tons of RSS readers already? Emacs’s gnus.el has an RSS reader that I use on my laptop if that helps. It’s probably not usable on a phone.

    Anyway, I’m supportive of the idea but for the same reason, I don’t have a Play account. If the app eventually shows up on F-droid I might give it a try.

    The name “Compass” must already be in use for tons of other things, fwiw.

    Why a built in reader instead of a web view? Don’t we already have too many apps with their own special text UI’s?

    In fact I wonder if the whole thing can be done as a static web page with local storage and permissions to retrieve the remote RSS, if that’s a thing. Or it could be a very simple standalone httpd that a local browser could point at.

    There’s another issue, which is that news RSS feeds often only give you the headlines. Here’s an example:

    https://www.sfgate.com/bayarea/feed/bay-area-news-429.php

    if you want to see a complete story, you have to click a link that gives you the usual bloaty web page full of ads. It’s fronted by the now-common bot detection stuff so you would have to do some fancy scraping to get the underlying page into your special client. If you want to go that route, you probably have to orchestrate a real browser anyway.


  • Um obviously it will depend on the cost, and tape in general these days is seen as an enterprise thing with prices to match. Are you going to make a tape drive? What about the tape?

    1TB isn’t that interesting since you can get LTO-5 or LTO-6 (1.5TB or 2.5TB native) fairly affordably these days. The physical size of the cartridge isn’t that important for most of us.



  • This appears to be about medical coding, nothing to do with writing computer software. When someone gets medical treatment, someone has to translate the treatment description into insurance billing identifiers, called “codes”. I don’t know anything else about it except that it takes a lot of training, and some effort goes into gaming the system to choose the most expensive codes for a given treatment. CPC stands for Certified Professional Coder.



  • What does that even mean? Fastmail is mostly or entirely FOSS on the mail backend but the web front end is currently non-FOSS. You can of course use a FOSS client instead.

    Hosted services are more about hosting than about the backend software anyway. You can use something like mailcow to self-host email if you don’t mind headaches.

    mxroute.com is probably all FOSS but they don’t make an issue of it. Actually idk what web control panel they use now. It was formerly cpanel but they switched, probably for cost reasons. The webmail ui’s are roundcube and a few others.






  • I’d be interested to know if there’s a way to fix this issue with extensions. Do you mean browser extensions? I use a browser sometimes and newpipe sometimes.

    Using a browser and ublock origin, I find viewing youtube to work ok except for the occasional age gated video that I can’t view without a google account. What’s stopped working is search ordering and seeing timestamps. Like old comments might get labelled “2 years ago” with mouse-over no longer showing the actual date/time. It’s not shown in the unrendered HTML in any obvious way either.