Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I'm going to be that guy. I skimmed a few sections and wanted to come back here and had to click the back button a dozen or so times to achieve that. Why does the URL update when you're scrolling down the page? It seems like a pointless "look what I can do!" kind of thing in JavaScript.


Firefox on Linux, I had to disable uBlockOrigin to see anything at all. It was a blank page, although ViewSource showed a large body of html and content.

This doesn't appear to be a commercial site at all. Weird that an ad blocker would be in conflict with it.

Also: I was able to read it with no problem with lynx, a text browser, after scrolling wwaaayyyy down to get to the article. Why don't designers (or whoever's responsible for this decision) put the content physically first, and the cruft last, in html? You can obviously arrange what gets displayed first in CSS/javascript. It would make life easier for text viewing people and their tools.


Websites should fail like escalators, not elevators. If your JS isn't working properly, the site should default to just dumping the text. I'll be OK dealing with the consequences of bad formatting.

/rant


>Websites should fail like escalators, not elevators.

I like this analogy and will be stealing it for future use. It's for this reason alone my site is plain text with few images and few styles.

Time to see how my site does in a CLI browser like Lynx or eLinks. :)


> I like this analogy and will be stealing it for future use.

Heh, I stole it from someone on here. I purposely built my site to work with minimal unnecessary components (JS, CSS, etc), and that makes it "just work" with things like Lynx. As it turns out, making sites simple also make them more compatible!


> making sites simple also make them more compatible!

For some reason (probably that I recently ran across it), this reminds me of:

"designing the system so that the manual will be as short as possible minimizes learning effort."

- Mike Lesk, as quoted in Expert C Programming, Van der Linden. https://en.wikipedia.org/wiki/Mike_Lesk


It sort of ignores the fact that escalators also fail by ripping people apart, maiming or killing them. I'll take the occasional broken back button and crappy contrast.


Operation failure - not structure failure.

If an elevator fails to operate - you can't get anywhere. If an escalator fails to operate - you can use them as stairs.


An escalator shredding someone is not a structural failure. Anyway, it wasn't meant all that seriously - just that a quip about failure modes has failed to consider the failure modes.


Just speculating, but there might be some UX justification. If you're reading a length bulleted list, and one of the items strikes you as shareworthy, I think it actually kind of makes sense to make it easy to link to that exact bullet.

The alternative is making users share the root URL, or scroll back to a menu and click a link to get the url directly to that #anchor.


You can use anchors without pushState.


I imagine it updates to make it easier for a reader to grab a link directly to the section they're currently looking at.

They could use `replaceState`[1] instead of `pushState` though so as not to mess with your browser history.

[1]https://developer.mozilla.org/en-US/docs/Web/API/History_API...


it's strange that the post we're replying to is 37 minutes old, but we both posted replies at nearly the same instant.


I'm curious why it seems to be connecting to maps.googleapis.com.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: