3 ms·
We had working voting systems before AJAX (source: I have a high-but-now-considered-lowish Slashdot ID) so we can do it. Besides, visited link state seems like
by ashark 9y ago
We had working voting systems before AJAX (source: I have a high-but-now-considered-lowish Slashdot ID) so we can do it. Besides, visited link state seems like it'd resolve the problem, and it's been in HTML approximately forever.
I still don't see the problem. The rest of what's possible can go on the current web, which would hopefully optimize (even more) for delivering apps if that happened. I want a content reader that won't, in practice, often run malicious code when I point it at the wrong document. That means it must not be able to run scripts, or at least not scripts that can communicate with the outside world in any way, at which point you may as well save the disk, memory, and complexity by leaving scripting out. There's no reason the thing I browse Craigslist, Wikipedia, IMDB, and Facebook and Twitter for that matter on must also be able to run ports of Wolfenstein or Excel clones or whatever. Or chat clients. The loss of trust and predictability in ordinary web browsing that scripting and complex layout systems bring aren't worth it. Put them somewhere else (including right where they are now—that'd be fine. Move the web to something more suited to it and leave the apps on the current "web")
- wvenable 9y agoI have a site that delivers content, forums, and maybe even a few application-like features all in one. Why have two systems when one system can do all that? It'll simply never happen. The web killed gopher. The web killed NNTP. The web has almost killed email. What you described has existed and been killed off a half-dozen times already. I'm not against having more user-protection features (heck, I have a number of plugins installed to do just that) but limiting usability and creativity is the wrong approach. It's been done and it always fails.
- ashark 9y agoI'd be OK with keeping both in the same client, with warnings when transitioning from one to the other (say, you follow a link on a web page to a web app), if that's what it took. It'd be like the HTTPS lock and content-origin policies and such but for guaranteeing the page you're visiting isn't actually a "page" that's spying on you and doing all kinds of other nasty stuff, or that the link you're following may go to a page that's a god-awful pain in the ass to read and burns your battery up ("modern" designed web articles). Like a not-shit version of Google's AMP, kinda. It'd be better separate if that worked out—if all the things that work just fine without JS and advanced CSS went to the New Web I'd be able to arrange things so I rarely need anything else. Web apps blow and I try to avoid them when possible, or, increasingly, they're better off in Electron for reasons including that it's quicker to get to them to murder them mercilessly when they go rogue, than it is with some tab somewhere. But a unified client for both that kept them strictly separate would at least be a lot better than what we have now.
- wvenable 9y agoEverything you describe is a usability nightmare including "with warnings when transitioning from one to the other". You might not use any web applications but major web applications are the most used sites on the web. Would it not be better to provide full functionality and simply use better sandboxing to avoid spying and nasty stuff? Reducing functionality is like throwing the baby out with the bathwater -- you really want better privacy and more control but rather than attack that problem you want to take the web back 20 years. That's a solution but it's not a very good one. I personally don't think that browsers shouldn't leak all the information that they do -- that should be a hole to be plugged.