Github has large businesses paying lots for their service. They dont need to blast you with adverts and subscription reminders. A news website has very few paying customers anymore so what else are they to do. You can give lectures about keeping ui clean and using no js all day but that doesn’t bring money in.
The real issue is how these clickbait NYT articles keep getting to the top of HN when they usually have little to no substance.
One can deliver ads, without needing 10’s of MB’s of data. I’m not a fan of the ads either, but to say ads are required to be giant ram and bandwidth sucking monsters is blatantly false. That’s not a requirement for ads, that is just where we have gotten as the ad industry has infected the Internet, not a technical requirement for advertising.
But even with MB’s of ad-infested insanity plastered everywhere, the rest of the site doesn’t also need to add to the craziness with MB’s of junk for what is essentially a page of text.
This is true, we could replicate the ad infestedness of a website with a tiny fraction of the processing power needed. But I think it’s more complex than that. To understand how to fix the problem we need to know how we got to the problem.
Who is making websites slow? Is it the site developers, the ad network developers or the managers? It’s quite clear that most of the time its the ad network scripts that slow websites down as the web jumps to warp speed with an ad blocker but why do some websites (Primarily news websites) have 1000 different ad network scripts and tracking scripts? If you ask the site developers they would probably tell you they hate it and wish they could remove most of them but its the managers that request tracking script #283 be added and the devs don’t get much of a say in it so posting an article on a developer focused website telling them something they already agree with is next to useless.
This is the primary reason AMP makes websites fast. Not because there is any tech magic that makes it fast. But because it lets developers say to managers “We can’t do that. It’s impossible on AMP”
There is also another case where big websites are slow and horrible to use on mobile. Twitter and reddit are like this. I think here the reason is to make you use the mobile app so telling them to make their websites work faster will also do nothing because they don’t want you using the website.
While I basically agree here, the problem is, if you are a new developer and you search the internet on how to build a menu for my website, basically all you will get back is using giant JS frameworks that take up gobs of space, instead of the few lines of CSS and HTML5 you need(without any JS) to actually build a menu. I don’t have a good solution to this, but I see this as a major contributor for why this craziness keeps growing in size.
I think it also doesn’t help that when we do get new things like webauthn, but then only get a JS interface to use them. Somewhat forcing our hand to require JS if you want nice things. That doesn’t mean we have to shove 500MB of JS to the user to use webauthn, but we can’t do it with just HTML and a form anymore.
That’s because nobody should need to search the internet for how to make a menu. It’s a list of links. It’s something you learn in the first hour of a lecture on HTML, chapter 1 of a book on HTML.
You probably neither need nor want to use webauthn. Certainly not yet! It was published as a candidate recommendation this year. Give others a chance to do the experimenting. Web standards used to take 10 years to get implemented. Maybe don’t wait quite that long, but I’m sure you’ll do fine with an <input type="password"> for a few years yet.
I was just using both as an example, I apologize for not being clear.
Yes a menu is just a list of links, but most people want drop-down or hamburger menu’s now, and that requires either some CSS or some JS. Again, go looking and all the examples will be in JS, unless you go searching specifically for CSS examples.
This is true of just about everything you want to do in HTML/Web land, the JS examples are super easy to find, the CSS equivalents are hard to find, and plain HTML examples are super hard to find.
Anyways, I basically agree webauthn isn’t really ready for production use, but again, both of these were examples, and webauthn just because it’s something I’m currently playing with. You can find lots of new web tech that is essentially JS only, despite it not needing to be, from a technical perspective. This is what I’m saying.
I understand it’s just an example, but that’s my point really: it’s yet another example of something people way overcomplicate for no good reason. ‘It’s the first google result’ just isn’t good enough. It’s basic competency to actually know what things do and how to do things in HTML and CSS, and not accomplish everything by just blindly copy-pasting whatever the first google result for your task is.
Web authentication? Sure it’s just an example, but what it’s an example of is people reinventing the wheel. What ‘new’ web technology isn’t just a shitty Javascript version of older web technology that’s worked for decades?
Maybe check out Dynamic Drive. I used to get CSS-based effects off it for DHTML sites in early 2000’s. I haven’t dug into the site to see if they still have lots of CSS vs Javascript, though. A quick glance at menus shows CSS menus are still in there. If there’s plenty CSS left, you can give it to web developers to check out after teaching them the benefits of CSS over Javascript.
I also noticed the first link on the left is an image optimizer. Using one is recommended in the article.
EDIT: The eFluid menu actually replaces the site’s menu during the demo. That’s neat.
Thanks for the link. Those are nice demos. I’d rather them not have the editor, though, so I easily see it in full screen. Could have a separate link for the source or live editing as is common elsewhere.
It’s not JavaScript that bloats an application, not even using frameworks. Vue is 20kb minified and gzipped, if that is ‘too much’ than you are writing for a very specialized audience. Using js won’t magically make ads, trackers or popups appear on your web page. Having an ad and impression based financing model will.
Size is just one metric. Ads, trackers, and popups all generally require JavaScript, and if you restrain yourself from using it you will naturally also not be using these things. Not using JavaScript also makes you exercise restraint in your design, letting the browser do what it’s best at (rendering pages) without straining your end-user’s resources.
sure, if you never touch a knife you won’t commit a knife crime, but encouraging everyone to stop using knifes isn’t a good idea. Implementing ads and trackers without javascript is trivial as well.
Either way, I agree with you on the point that you shouldn’t use JS when it’s just as easy and nicer for everyone involved when you can skip it. But some web sites really are ‘web apps’, they have interactivity that makes it very inconvenient to not use js.
Knives have useful purposes outside of knife crime. Javascript does not have many genuinely useful purposes outside of tracking, ads and other such crap.
Implementing ads and trackers without javascript is trivial as well.
I would have no problem with ads if they were <img src="/ads/91241.jpg" />. And implementing trackers without Javascript is not at all easy. The best you can do is record the IP address of visitors to your website. You’ve always been able to do that, and it’s nowhere near the same thing as tracking.
But some web sites really are ‘web apps’, they have interactivity that makes it very inconvenient to not use js.
I’d rather just use a proper, open source programme running on my computer.
Javascript does not have many genuinely useful purposes outside of tracking, ads and other such crap.
Ehh, Javascript has a history too, and while the situation is a lot better nowadays, there were many things that could only be done using javascript. Very popular google services like Gmail or Maps only were that popular because they worked the way people wanted it to (ie. dynamically), and just dismissing this as infantile or ignorant really doesn’t do anything.
Ads would be done using metadata fingerprinting and backend ad-generation, to still generate personalized data. My feeling is that if it were necessary, this would even be done, despite meaning more work for servers.
I’d rather just use a proper, open source programme running on my computer.
Sadly, that’s not the case for most people. On the one hand, most people, if they have to wouldn’t necessary write cross platform software, and on the other hand, “less experienced” users are always weary of installing stuff on their devices. Who knows what it could do? Having a portable, seemingly universal sandbox seems like a great alternative in their eyes.
The question if this should be the case is a justified, but an entirely different one.
Use a map program. Use a mail client. The web is for documents.
I’m not opposed to having a universal secure portable virtual machine. I am opposed to it being the browser, because the browser is a terrible VM, it’s designed for documents. HTML/CSS/JS is a terribly designed and awfully inefficient rendering engine for programmes. It is designed for documents, and HTML/CSS work great for that, but for apps? Yuck.
The only reason the web is as it is is because things like the JVM were such complete failures at sandboxing programmes correctly.
On a related note, sr.ht looks great!
Thanks!
Github has large businesses paying lots for their service. They dont need to blast you with adverts and subscription reminders. A news website has very few paying customers anymore so what else are they to do. You can give lectures about keeping ui clean and using no js all day but that doesn’t bring money in.
The real issue is how these clickbait NYT articles keep getting to the top of HN when they usually have little to no substance.
One can deliver ads, without needing 10’s of MB’s of data. I’m not a fan of the ads either, but to say ads are required to be giant ram and bandwidth sucking monsters is blatantly false. That’s not a requirement for ads, that is just where we have gotten as the ad industry has infected the Internet, not a technical requirement for advertising.
But even with MB’s of ad-infested insanity plastered everywhere, the rest of the site doesn’t also need to add to the craziness with MB’s of junk for what is essentially a page of text.
This is true, we could replicate the ad infestedness of a website with a tiny fraction of the processing power needed. But I think it’s more complex than that. To understand how to fix the problem we need to know how we got to the problem.
Who is making websites slow? Is it the site developers, the ad network developers or the managers? It’s quite clear that most of the time its the ad network scripts that slow websites down as the web jumps to warp speed with an ad blocker but why do some websites (Primarily news websites) have 1000 different ad network scripts and tracking scripts? If you ask the site developers they would probably tell you they hate it and wish they could remove most of them but its the managers that request tracking script #283 be added and the devs don’t get much of a say in it so posting an article on a developer focused website telling them something they already agree with is next to useless.
This is the primary reason AMP makes websites fast. Not because there is any tech magic that makes it fast. But because it lets developers say to managers “We can’t do that. It’s impossible on AMP”
There is also another case where big websites are slow and horrible to use on mobile. Twitter and reddit are like this. I think here the reason is to make you use the mobile app so telling them to make their websites work faster will also do nothing because they don’t want you using the website.
While I basically agree here, the problem is, if you are a new developer and you search the internet on how to build a menu for my website, basically all you will get back is using giant JS frameworks that take up gobs of space, instead of the few lines of CSS and HTML5 you need(without any JS) to actually build a menu. I don’t have a good solution to this, but I see this as a major contributor for why this craziness keeps growing in size.
I think it also doesn’t help that when we do get new things like webauthn, but then only get a JS interface to use them. Somewhat forcing our hand to require JS if you want nice things. That doesn’t mean we have to shove 500MB of JS to the user to use webauthn, but we can’t do it with just HTML and a form anymore.
That’s because nobody should need to search the internet for how to make a menu. It’s a list of links. It’s something you learn in the first hour of a lecture on HTML, chapter 1 of a book on HTML.
You probably neither need nor want to use webauthn. Certainly not yet! It was published as a candidate recommendation this year. Give others a chance to do the experimenting. Web standards used to take 10 years to get implemented. Maybe don’t wait quite that long, but I’m sure you’ll do fine with an
<input type="password">for a few years yet.I was just using both as an example, I apologize for not being clear.
Yes a menu is just a list of links, but most people want drop-down or hamburger menu’s now, and that requires either some CSS or some JS. Again, go looking and all the examples will be in JS, unless you go searching specifically for CSS examples.
This is true of just about everything you want to do in HTML/Web land, the JS examples are super easy to find, the CSS equivalents are hard to find, and plain HTML examples are super hard to find.
Anyways, I basically agree webauthn isn’t really ready for production use, but again, both of these were examples, and webauthn just because it’s something I’m currently playing with. You can find lots of new web tech that is essentially JS only, despite it not needing to be, from a technical perspective. This is what I’m saying.
I understand it’s just an example, but that’s my point really: it’s yet another example of something people way overcomplicate for no good reason. ‘It’s the first google result’ just isn’t good enough. It’s basic competency to actually know what things do and how to do things in HTML and CSS, and not accomplish everything by just blindly copy-pasting whatever the first google result for your task is.
Web authentication? Sure it’s just an example, but what it’s an example of is people reinventing the wheel. What ‘new’ web technology isn’t just a shitty Javascript version of older web technology that’s worked for decades?
Maybe check out Dynamic Drive. I used to get CSS-based effects off it for DHTML sites in early 2000’s. I haven’t dug into the site to see if they still have lots of CSS vs Javascript, though. A quick glance at menus shows CSS menus are still in there. If there’s plenty CSS left, you can give it to web developers to check out after teaching them the benefits of CSS over Javascript.
I also noticed the first link on the left is an image optimizer. Using one is recommended in the article.
EDIT: The eFluid menu actually replaces the site’s menu during the demo. That’s neat.
An interesting project that shows how modern layouts can be built without JavaScript is W3C.CSS.
/cc @milesrout @zie
Thanks for the link. Those are nice demos. I’d rather them not have the editor, though, so I easily see it in full screen. Could have a separate link for the source or live editing as is common elsewhere.
It’s not JavaScript that bloats an application, not even using frameworks. Vue is 20kb minified and gzipped, if that is ‘too much’ than you are writing for a very specialized audience. Using js won’t magically make ads, trackers or popups appear on your web page. Having an ad and impression based financing model will.
Size is just one metric. Ads, trackers, and popups all generally require JavaScript, and if you restrain yourself from using it you will naturally also not be using these things. Not using JavaScript also makes you exercise restraint in your design, letting the browser do what it’s best at (rendering pages) without straining your end-user’s resources.
sure, if you never touch a knife you won’t commit a knife crime, but encouraging everyone to stop using knifes isn’t a good idea. Implementing ads and trackers without javascript is trivial as well.
Either way, I agree with you on the point that you shouldn’t use JS when it’s just as easy and nicer for everyone involved when you can skip it. But some web sites really are ‘web apps’, they have interactivity that makes it very inconvenient to not use js.
Knives have useful purposes outside of knife crime. Javascript does not have many genuinely useful purposes outside of tracking, ads and other such crap.
I would have no problem with ads if they were
<img src="/ads/91241.jpg" />. And implementing trackers without Javascript is not at all easy. The best you can do is record the IP address of visitors to your website. You’ve always been able to do that, and it’s nowhere near the same thing as tracking.I’d rather just use a proper, open source programme running on my computer.
Ehh, Javascript has a history too, and while the situation is a lot better nowadays, there were many things that could only be done using javascript. Very popular google services like Gmail or Maps only were that popular because they worked the way people wanted it to (ie. dynamically), and just dismissing this as infantile or ignorant really doesn’t do anything.
Ads would be done using metadata fingerprinting and backend ad-generation, to still generate personalized data. My feeling is that if it were necessary, this would even be done, despite meaning more work for servers.
Sadly, that’s not the case for most people. On the one hand, most people, if they have to wouldn’t necessary write cross platform software, and on the other hand, “less experienced” users are always weary of installing stuff on their devices. Who knows what it could do? Having a portable, seemingly universal sandbox seems like a great alternative in their eyes.
The question if this should be the case is a justified, but an entirely different one.
Use a map program. Use a mail client. The web is for documents.
I’m not opposed to having a universal secure portable virtual machine. I am opposed to it being the browser, because the browser is a terrible VM, it’s designed for documents. HTML/CSS/JS is a terribly designed and awfully inefficient rendering engine for programmes. It is designed for documents, and HTML/CSS work great for that, but for apps? Yuck.
The only reason the web is as it is is because things like the JVM were such complete failures at sandboxing programmes correctly.