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

The "end game" from the user's point of view would if someone indeed came up with "content selectors". It would be quite an achievement to build one that was actually useful.

Maybe easy version could work exactly the same way as blockers do, except they'd invert the way the rules are evaluted, and of course the rules would need to be custom as well. And then if the rules would fail to find some actual content on the page, then I guess you might not even know about it :).



That closely describes uMatrix in default deny mode. For each individual site and subdomain, the user gets to choose if they want to allow cookies, images, CSS, JS, XHR, frames. Too bad it's unmaintained, with most of its practical and commonly intended use cases fitting the simpler uBlock Origin.


But the underlying principle is still filtering out the undesired content instead of selecting the desired content.


Surely the principle the parent suggested was to initially filter out all of the content, and then selecting the desired content to load.

But isn't "filtering out all of the content" just another way of saying, "not loading any content" - until explicitly requested?


Perhaps I misunderstood. But let's say I do run uMatrix in deny more, I get the site HTML as the Tim Berners-Lee intended. Yet that already may contain some kind of advertising material, so to view only the actual content I'd need to have e.g. an XPath do it. So I would have a database of sites and XPath (or other ways) describing how to find the content, instead of describing how to get rid of non-content.

Can I tell uMatrix to only show a certain XPath from the page? If so, I agree it is a content selector, albeit pretty impractical :).


uMatrix is the absolute best, I can't imagine browsing without it. So the unmaintaned part is very sad and troubling.

I do install uBlock Origin for all non-technical family members, but uMatrix is so much more capable.


The end game would be some AI running on a gpu in your monitor, so the browser nor the originating website need to know about any blocking.


Except you'd still have client-side execution in that case. Obviously not everyone cares, but there are some straightforward objective reasons to care if you're blocking it being rendered anyway, like why waste cycles on it.


True, but it's hard to block execution without the website not knowing about it.


That's basically what the "reader mode" in the various browsers do


Anyone remember Proxomitron?


The end game for the web to be scraped into entries in a database to be displayed by real software instead of a web browser.


Sounds miserable and authoritarian




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

Search: