#dev 2025-01-31

2025-01-31 UTC
#
[Joe_Crawford] I like `cursor: not-allowed` as a hint, at least to desktop browsers on mouseover.
jbrr joined the channel
#
[tantek] right, the question is what to do with dead links. Internet Archive is one possibility, and if none, then maybe a local tombstone?
geoffo and grufwub joined the channel
#
doesnm IA can delete some copy of sites if someone want.
#
claudinec capjamesg[d]: I add links to the Wayback Machine or archive.is where possible, with specual markup - https://www.claudinec.net/posts/2024-06-22-resisting-linkrot/
#
claudinec Oh, or did you mean pages that have gone from your own website?
#
doesnm try to enter any link from habr.ru
#
doesnm https://doesnm.cc/Screenshot_2025-01-31-10-01-26.png
doesnm and jak2k joined the channel
#
[tantek] that's usually a robots.txt directive causing that
#
[tantek] IA still has the archived the content, they are just not returning it in queries
GuestZero joined the channel
#
capjamesg[d] claudinec Yeah, I was thinking about web apps I have made where I don't run the back-end anymore.
#
capjamesg[d] An archive link is less relevant because it would show an archived page of something that no longer accepts user input.
#
[tantek] Ironically, an Archive link to a page with forms that run completely with in-page JS should work fine
#
[tantek] for example http://asin.cc is a "web app" that accepts user input and does not require any back-end when JS is enabled
#
Loqi [preview] Tantek Çelik
#
[tantek] IMO this is a good goal for web apps. Both no-JS-needed back-end support, then progressively enhanced to work fully (or as much as possible) clientside without requiring any back-end requests.
nemonical and barnaby joined the channel
#
capjamesg[d] Does anyone have a recommendation on a spell checking software to use on a folder of markdown files?
ttybitnik, nemonical and [Murray] joined the channel
#
[Murray] [capjamesg] how are you viewing the markdown? or do you want to point it at the whole folder and run it through the lot? Asking because there are some good spellchecker plugins for various IDEs ‒ I use one on VSCode which is solid
#
capjamesg[d] [Murray] Which one do you use on VSCode?
#
capjamesg[d] I would ideally like to run it on a full folder.
#
capjamesg[d] I have a lot of posts to check.
#
capjamesg[d] I haven't really had a good spellcheck up until now.
#
[Murray] I use "Code Spell Checker" on VSCode
#
[Murray] never really considered running a bulk spell check; what would the output of that even be? or would you automate the changes?
#
[Murray] I guess in an IDE you can at least open a bunch of files at once 😄
#
capjamesg[d] That's what I am doing haha.
#
capjamesg[d] It doesn't help that I have technical blog posts with deliberate "typos" (i.e. prefixes).
#
[Murray] I wonder if something like Grammarly has a CLI? This also feels like a genuinely useful use case for an LLM: scan through, suggest changes, let you accept them in sequence
doesnm and jak2k joined the channel
#
GuestZero are there any non-intrusive, privacy-preserving analytics tools out there (i'd guess not really)?
#
aaronpk depends on what you mean by privacy preserving, but I use https://usefathom.com/
#
GuestZero thx aaronpk
#
GuestZero does anybody know of free alternatives?
#
[Joe_Crawford] So you're looking for "analytics," likely open source, "non-intrusive", "privacy-preserving" - probably self-hosted. I don't know of any, though I have to say that if you have access to your webserver logs you have *some* data that could be collated into something structured using tools like... webalizer and similar.
#
[Joe_Crawford] I can't think of anything but then I''ve not explored this space beyond GA in a while.
#
GuestZero thank you Joe, webalizer looks promising
#
[Joe_Crawford] AWStats is another in that category. I suppose old school web pages means consideing old-school analytics techniques and tools.
#
sebbu matomo.org or plausible.io ? (i haven't tested them)
#
sebbu I remember MRTG too to have nice stats/diagrams for a website
#
sebbu (multi router traffic grapher)
#
jeremycherfas Homework done https://www.jeremycherfas.net/blog/more-accessibility-fun
#
jeremycherfas Another black mark for Tailwind--
#
Loqi Tailwind has -1 karma over the last year
jak2k joined the channel
#
[Joe_Crawford] Nice job jeremycherfas++
#
Loqi jeremycherfas has 3 karma in this channel over the last year (11 in all channels)
#
jeremycherfas Well, the menu is still a bit messed up, and there are other things that irk me, so I'm hoping this gives me more incentive for the complete redesign.
#
jeremycherfas I like Safari as my main browser, but I do think Firefox might be easier to use when it comes to developing and troubleshooting a site.
#
sebbu wrote 2 bugreports today
GuestZero_ and corlaez joined the channel
#
corlaez Old article but thought it would be interesting for the no JS folk: gemini://xavi.privatedns.org/small-web-browser.gmi
#
[Joe_Crawford] Is there a web version of that link? Also, search results for "gemini browser" are now wall to wall "Google AI bot usage links." Not loving the name "Gemini" right now. Super duper overloaded name.
#
corlaez sorry i meant to share this link instead: https://portal.mozz.us/gemini/xavi.privatedns.org/small-web-browser.gmi
#
corlaez Yep the gemini ai takeover is tragic for that tiny protocol. the exclusionism as well (not allowing regular browsers to open such links). That is part of what the article addresses and links to other interesting articles as well.
#
corlaez Lagrange is the leader of the gemini clients out there if you do want to install it.
#
corlaez The css in portal.mozz.us is really bad, but whatever.
#
[Joe_Crawford] Yeah, just found that. Weirdly last I tried some Gemini clients all of them were just... unusable command line. Somehow didn't encounter Lagrange, which was straightforward to install.
#
corlaez I am having the same issue with Scorpion protocol, I have no idea how to find an off the shelf browser for it.
#
[Joe_Crawford] It's a fun idea, but my own biases cause me to think setting the tech of an interactive document network to 1991 doesn't really help very much.
#
[Joe_Crawford] If it gains adoption, all the incentives to make it more powerful will turn it into the web in a few years.
#
corlaez Yeah, I am not sure it would be a replacement of the interactive modern internet.
#
[Joe_Crawford] I could of course be wrong.
#
corlaez It would probably be just like gemini, an alternative space of a more document focused web
#
[Joe_Crawford] What we need is a planet with different incentives and checks on bad intrusive behavior.
#
corlaez but with the benefit that you can read it in your regular browser
#
[Joe_Crawford] And that is politics.
#
corlaez Partly partly, but I do think that offering the technical and practical, easy to use alternatives is part of the equation as well
#
corlaez Because if everyone is locked into Facebook, Twitter, and other big corp services they will just Lobby the hell out of the goverment with their sweet surveillance bucks
#
[Joe_Crawford] It is definitely a good sign to see enthusiasm for it. It's a kind of exemplar for good aspects of the web. We've seen pushback in the browser space against things that are abused. 3rd party cookies are the biggest. In the past: onunload abuse, new window spawning, Flash's poor performance on mobile. Web tech can die if it doesn't play well with the world as it browsers.
#
[Joe_Crawford] Right, push on your governments to put a check on such services.
gRegor, doesnm, MyNetAz, nemonical and ttybitnik joined the channel
#
[Joe_Crawford] ↩️ 😞 yeah, I feel like I didn't notice it last time. I think I went directly to https://en.wikipedia.org/wiki/Gemini_(protocol) before and avoided them. But yeah, wow. All AI.
#
doesnm gemini is not only AI. It also crypto exchange :D
#
gRegor what is gemini
#
Loqi Gemini may refer to Google Gemini, an LLM tool like ChatGPT, or for developers, the Gemini protocol, an alternative to HTTP https://indieweb.org/Gemini
#
[Joe_Crawford] ...or a skeevy crypto exchange. Wow.
[Jo] joined the channel
#
[Jo] people are still doing crypto??
#
[snarfed] also https://www.bridgy.com/ 😆
#
gRegor nooo 😭
#
sebbu I got a nostr account, with a lightning address, does that counts as crypto ?
#
[snarfed] yup
marmadilemanteat joined the channel
#
arekenaten I think it's really interesting that there feels to me like there's always a bit of a choice between accessibility and anonymity. If you make a system flexible enough to accommodate the widest variety of users, is it also true that you necessarily build a system capable of narrowing down options to the point of identifying those users?
#
arekenaten Repo that got me thinking about it: https://github.com/OliverBrotchie/CSS-Fingerprint
#
IWDiscord <a​rekenaten>
#
arekenaten Did you know you could fingerprint without JS being turned on?
#
Loqi [preview] [OliverBrotchie] CSS-Fingerprint: Pure CSS device fingerprinting.
#
arekenaten [edit] Repo that got me thinking about it: https://github.com/OliverBrotchie/CSS-Fingerprint
#
arekenaten Did you know you could fingerprint without JS being turned on?
#
IWDiscord <a​rekenaten>
btrem joined the channel
#
[snarfed] arekenaten yeah that's a pretty well-understood academic field, https://www.google.com/search?q=de-anonymization
#
[snarfed] Cynthia Dwork, Rakesh Agrawal, etc
#
[snarfed] check out hippocratic databases and differential privacy
#
[snarfed] (...but we're also headed toward #indieweb-random)
sebbu, sp1ff and corlaez joined the channel
#
corlaez this gravy hack seems huge: https://www.removepaywall.com/search?url=https://www.wired.com/story/gravy-location-data-app-leak-rtb/
[tw2113] joined the channel