7 ms·

What the web is increasingly unable to do today: provide text content without requiring a code execution environment. This site is another example of that. All

by majika 11y ago

What the web is increasingly unable to do today: provide text content without requiring a code execution environment. This site is another example of that. All non-application websites should provide all their content in semantic HTML at appropriate HTTP endpoints, with CSS styling (in as few requests as possible) as required per the design, and JavaScript (in as few requests as possible) that takes the semantic HTML and makes it interactive (potentially adding and removing elements from the DOM) as required per the design. The CSS should not depend on mutations resulting from the JavaScript, nor should the JavaScript assume anything of the applied styles (as the user agent should be able to easily apply custom user-styles for your site; e.g. Gmail only providing a limited set of styles that are managed server-side is laughable). Thus, all content is readable and styled properly without requiring an arbitrary code execution environment. That is what the web was meant to be. Unfortunately, most "web developers" have made the web worse over the past 10 years because simple, functional, minimal technology is not impressive, and hipsters love to show off. Nor does it help that there are few capitalist incentives for the web being open and malleable -- e.g. so users can easily use a different front-end for Facebook, or users can easily choose to avoid analytics or advertisements, or users might prefer to use the website rather than the app (providing access to personal details, contacts, location, tracking, etc). The state of the web is emergent and I'm not sure what anyone could do about it (perhaps make a better browser?), but it really irks me when web developers pretend like they're actually doing something good or useful, or that the web is actually in a healthy state. In my experience, it's the people who don't talk about web development who are the best web developers; these are the people who don't wince when they write a HTML document without a single `<script>`.

nandhp 11y ago
> What the web is increasingly unable to do today: provide text content without requiring a code execution environment. This site is another example of that. I was about to argue that this website is actually an excellent example of what you seek -- each link has a separate URL associated with it that returns a page containing that content. The links point to these real URLs so they work with "open in new tab" and "copy link" and in browsers without JavaScript enabled, while the JavaScript that runs when you click it changes the page content via AJAX (possibly saving a few round-trips) and updates the current page URL so that back/forward history and the address bar both work just like you're navigating between real webpages. And this works perfectly in Firefox (with JavaScript) and almost perfectly in Lynx (the table of contents still fills the first screenful, but that's hard to fix since Lynx doesn't support CSS). But it completely fails if you have JavaScript disabled in Firefox. Every page starts with the table of contents visible and the content collapsed (through CSS). The page then seems to assume that JavaScript will be able to immediately switch the page to the correct view (i.e. the site is broken if you have working CSS but not JavaScript). Navigation to a given page directly should start the other way by default, and to make that happen is just 21 missing characters (` class="page-feature"` on the <body> tag). However, this unfortunate error completely ruins this otherwise beautiful example of progressive enhancement.
- bhrgunatha 11y ago
  > I was about to argue that this website is actually an excellent example of what you seek > But it completely fails if you have JavaScript disabled in Firefox. I'm not sure what point you're trying to make. We've known for a long time that progressive enhancement IS possible, it's just that very few sites bother to design for that. Are you just saying it's difficult and that this site "almost" made it.
- NOtherDev 11y ago
  The issue when JavaScript is disabled is now fixed, thanks for pointing this out!
bad_user 11y ago
You're talking about "progressive enhancement". It's a romantic idea, but it never happened, probably because it's too hard and the cost is not justified given most users run with their browser's default settings. The precursor of the web made by Tim Berners-Lee dates back to 1980, but it was not based on HTML or HTTP. These happened later in 1990 and early 1991. But then CSS happened in 1994. And Javascript happened in 1995 at Netscape, but then Javascript was completely useless until Microsoft came up with the iframe tag in 1996 and then with XMLHttpRequest in 1999, which was later adopted by Mozilla, Safari and Opera. And people still couldn't grasp its potential until Google delivered Gmail in 2004 and Google Maps in 2005. Not sure what the "the web was meant to be", we should ask Tim Berners-Lee sometimes, but in my opinion the web has been and is whatever its developers and users wanted it to be, with contributions from multiple parties such as Netscape, Microsoft, Mozilla, KDE/KHTML, Apple, Google and many other contributors, being a constantly evolving platform.
- ams6110 11y ago
  It was Microsoft with Outlook Web Access that really was the first big example of what was possible. I worked with several folks around that time and we were doing "rich" web apps in IE5 with XML data delivered from the server via xmlhttp and building the presentation in the browser with XSLT. It was really slick, especially at the time, to be able to filter and sort, do summary and detail views, etc. all without a page refresh.
- jacquesm 11y ago
  CSS happened in 1996, the first couple of years of the web we had no CSS, all styling was done with inline attributes up to that point. The frustrating bit is that CSS was supposed to separate content/structure and markup, but now we have pages without content but with markup where the content is loaded after the fact. This really overshot the mark.
  - TeMPOraL 11y ago
    Also note tat the separation of content and structure is mostly nonexistent as it's currently practiced. People like to talk about it as a Golden Rule, a holy virtue, and then proceed to write even worse code than they did with tables. Generally, if you find yourself writing a tangled mess of divs in order to support CSS tricks, you're not really separating content from the structure. HTML5 semantic tags helped a little, but I still rarely see them used live. (I'm not really convinced that separation of form and content is a feasible goal anyway - some elements of form are also content at the same time - but it's still a good goal to have.) > now we have pages without content but with markup where the content is loaded after the fact This is absolutely ridiculous and it, combined with the idea of routing everything through the cloud in IoT / home automation applications, makes me wonder what happened to some good old-fashioned engineering sanity. It's like people are trying to create wasteful and insecure systems on purpose. ("And what would that purpose be," - the cynic in me asks - "maybe monetizing people's data?").
mei0Iesh 11y ago
It's the Flashification of the web. People who wanted to show off, or code web apps, used to use Macromedia Flash. People complained about it, in part because sometimes if you accessed a site without the Flash plugin you'd see a blank page. But Flash was great in many ways, and it was self-contained in objects, so websites were mostly still websites. JavaScript had been around for a long time, but there was still a cultural norm that most people respected about not requiring JavaScript. This was mainly because a lot of browsers still didn't fully support it, or people had it turned off. It was also when some people had cookies disabled. Then once Adobe bought Flash, and then Apple blocked Adobe Flash, it really killed the Flash way, and all that spilled over into HTML with HTML 5 and the new cultural norm of kids who are more concerned with showing off socially than the meat and potatoes of hypertextual information. It should've been obvious that there was a need for a new web, for code and multimedia. But in the .com boom nobody would dare try to start with something unpopulated, since it'd risk losing their chance at fortune. Today, instead, maybe we should go in the opposite direction and create a new old web; a hypertext network that specifically only works for HTML, so people can have this one to morph into an app network, and we'll not lose the text-linking place we've grown accustomed to.
- hugozap 11y ago
  > we'll not lose the text-linking place we've grown accustomed to. It's just a minority of users who are accustomed to the old web.
- initself 11y ago
  Rad idea. How would one initate a new network of this kind separately from what exists today? And what kind of interoperability would it have? Can't imagine it could go cold turkey!
BurningFrog 11y ago
Strong opinions on how things should be, but no arguments for why. That is never going to convince me.
- savanaly 11y ago
  It's because his post is entirely opinion, albeit one that frequently shows up at the top of HN comments sections because it's such a popular opinion in the hacker community that we forget it's not actually the majority's opinion.
frankensmith 11y ago
This comment reminds me of Maddox and his 90s' style html blog: http://www.thebestpageintheuniverse.net http://www.thebestpageintheuniverse.net http://maddox.xmission.com/ http://maddox.xmission.com/
userbinator 11y ago
Thus, all content is readable and styled properly without requiring an arbitrary code execution environment. That is what the web was meant to be. In other words, it was supposed to be a worldwide hyperlinked document library --- and we have mostly achieved that goal, although it is a library wherein you are constantly tracked and bombarded by books flying off the shelves at you, screaming at you to read them, and most of the books consist solely of ads with very little useful informational content. In my experience, it's the people who don't talk about web development who are the best web developers; these are the people who don't wince when they write a HTML document without a single `<script>`. Agreed completely. The ones who write information-dense HTML pages, often by hand, would not be considered "web developers" nor would they consider themselves to be; but they are what the web needs most. I've done that, and I don't consider myself a "web developer" either. it really irks me when web developers pretend like they're actually doing something good or useful, or that the web is actually in a healthy state I wouldn't doubt that they genuinely feel like what they're doing is good or useful; I've noticed the appeal of "new and shiny" is especially prevalent in the web development community, with the dozens of frameworks and whatnot coming out almost daily, proposals of new browser features, etc. Very little thought seems put into the important question of whether we actually need all this stuff. It's all under the umbrella of "moving the web forward", whatever that means. But I think we should stop and look back on the monstrosities this rapid growth has created.
- hugozap 11y ago
  Against what platform should we compare the web to decide what it should look like?. I think everyone has a different criteria.
ryandrake 11y ago
What we need is a new protocol: something that lets an author write and publish text documents, marked up with basic styling, with "hyperlinks" to other text documents. The protocol could allow embedding of simple inline figures as well. Users would run a "browser" whose function was limited to requesting and displaying these documents.
- qu4z-2 11y ago
  You mean like AMP? https://www.ampproject.org/ https://www.ampproject.org/ And yes, it's pretty silly that this is a thing.
username223 11y ago
> ... web developers pretend like they're actually doing something good or useful... Many "Web developers" are to "the web" as rapists are to sexual reproduction. They achieve short-term gratification and (perhaps unintentionally) pass their traits on to the next generation, but they couldn't care less about the longer-term consequences of their actions. The web started as "clickable citations." Now it's mostly hostile surveillance with cat pics.
ocfx 11y ago
> Unfortunately, most "web developers" have made the web worse over the past 10 years because simple, functional, minimal technology is not impressive, and hipsters love to show off No it's because when I go into work someone says to make it a certain way and if I want to get paid I have to. If you want to blame anyone blame designers who see proof of concept stuff from developers and throw it in designs.
ArleneBarlow 11y ago
The web got better according to most people. The devs responded to what people wanted, and that's flashy, interactive stuff. What the web was 'meant' to be is irrelevant, because the web has grown far beyond what it's original designer intended. On one hand you imply that it's good that the web is open, and part of being open is that people can make it what it wants, and not obey the original intention of the first designer. It's unfortunate that your vision of the web is in disagreement with the 'general users'.
Pxtl 11y ago
I tend to think rants like this are just being luddites, but all this JavaScript and external resources rest are ruining the experience of the web. Holy crap is everything slow now. On desktop we've all gotten into the habit of tabbing the new stuff and waiting for load while doing something else, but on mobile where that workflow isn't as easy and you have to watch a page load? I'd say HN is one of the very few sites I can stand anymore.
- userbinator 11y ago
  I encourage you to try the "NoScript challenge" - disable JS, enabling it only for sites you often use and which need it, and browse the Web for a week. If you tend to indulge in rich multimedia you're probably going to give up soon, but if, like me, you're just after text and image content the majority of the time, you might actually prefer it. Pages load nearly instantly (or not at all), there no more intrusive popups/popunders/slideovers/etc. and other annoyances like disabling right-click, text selection, or stuffing text into your clipboard. The "almost all sites have JavaScript and will break if you disable it" statement is common, and while I agree that the vast majority of sites do have JS on them, whether or not the stuff that will "break" on them if you disable JS is actually of any use to you is questionable.
TheAceOfHearts 11y ago
I think this is because so many websites try to shove ads and tracking crap down your throat. It's not even particularly difficult to pull off for certain websites. For example, my blog uses React and it does server-side rendering, so it'll work without JS at all. However, if you do have JS enabled, it'll let you avoid doing full-page reloads to navigate around. I totally agree that for content-focused websites, you tend to get a better experience by limiting gratuitous JS abuse. However, this isn't very feasible when you're building highly interactive applications. In the case of something like Facebook, they do have a version of their app that works without JS... But how many people can afford to maintain multiple versions of their applications? I think a good compromise is achievable by well documented APIs or even better with a public GraphQL schema! If I don't use magic APIs to build our frontend app, you can build a different frontend that's tailored for your needs.

by majika 11y ago

nandhp 11y ago

> What the web is increasingly unable to do today: provide text content without requiring a code execution environment. This site is another example of that. I was about to argue that this website is actually an excellent example of what you seek -- each link has a separate URL associated with it that returns a page containing that content. The links point to these real URLs so they work with "open in new tab" and "copy link" and in browsers without JavaScript enabled, while the JavaScript that runs when you click it changes the page content via AJAX (possibly saving a few round-trips) and updates the current page URL so that back/forward history and the address bar both work just like you're navigating between real webpages. And this works perfectly in Firefox (with JavaScript) and almost perfectly in Lynx (the table of contents still fills the first screenful, but that's hard to fix since Lynx doesn't support CSS). But it completely fails if you have JavaScript disabled in Firefox. Every page starts with the table of contents visible and the content collapsed (through CSS). The page then seems to assume that JavaScript will be able to immediately switch the page to the correct view (i.e. the site is broken if you have working CSS but not JavaScript). Navigation to a given page directly should start the other way by default, and to make that happen is just 21 missing characters (` class="page-feature"` on the <body> tag). However, this unfortunate error completely ruins this otherwise beautiful example of progressive enhancement.

bhrgunatha 11y ago
> I was about to argue that this website is actually an excellent example of what you seek > But it completely fails if you have JavaScript disabled in Firefox. I'm not sure what point you're trying to make. We've known for a long time that progressive enhancement IS possible, it's just that very few sites bother to design for that. Are you just saying it's difficult and that this site "almost" made it.
NOtherDev 11y ago
The issue when JavaScript is disabled is now fixed, thanks for pointing this out!

bad_user 11y ago

You're talking about "progressive enhancement". It's a romantic idea, but it never happened, probably because it's too hard and the cost is not justified given most users run with their browser's default settings. The precursor of the web made by Tim Berners-Lee dates back to 1980, but it was not based on HTML or HTTP. These happened later in 1990 and early 1991. But then CSS happened in 1994. And Javascript happened in 1995 at Netscape, but then Javascript was completely useless until Microsoft came up with the iframe tag in 1996 and then with XMLHttpRequest in 1999, which was later adopted by Mozilla, Safari and Opera. And people still couldn't grasp its potential until Google delivered Gmail in 2004 and Google Maps in 2005. Not sure what the "the web was meant to be", we should ask Tim Berners-Lee sometimes, but in my opinion the web has been and is whatever its developers and users wanted it to be, with contributions from multiple parties such as Netscape, Microsoft, Mozilla, KDE/KHTML, Apple, Google and many other contributors, being a constantly evolving platform.

ams6110 11y ago
It was Microsoft with Outlook Web Access that really was the first big example of what was possible. I worked with several folks around that time and we were doing "rich" web apps in IE5 with XML data delivered from the server via xmlhttp and building the presentation in the browser with XSLT. It was really slick, especially at the time, to be able to filter and sort, do summary and detail views, etc. all without a page refresh.
jacquesm 11y ago
CSS happened in 1996, the first couple of years of the web we had no CSS, all styling was done with inline attributes up to that point. The frustrating bit is that CSS was supposed to separate content/structure and markup, but now we have pages without content but with markup where the content is loaded after the fact. This really overshot the mark.
- TeMPOraL 11y ago
  Also note tat the separation of content and structure is mostly nonexistent as it's currently practiced. People like to talk about it as a Golden Rule, a holy virtue, and then proceed to write even worse code than they did with tables. Generally, if you find yourself writing a tangled mess of divs in order to support CSS tricks, you're not really separating content from the structure. HTML5 semantic tags helped a little, but I still rarely see them used live. (I'm not really convinced that separation of form and content is a feasible goal anyway - some elements of form are also content at the same time - but it's still a good goal to have.) > now we have pages without content but with markup where the content is loaded after the fact This is absolutely ridiculous and it, combined with the idea of routing everything through the cloud in IoT / home automation applications, makes me wonder what happened to some good old-fashioned engineering sanity. It's like people are trying to create wasteful and insecure systems on purpose. ("And what would that purpose be," - the cynic in me asks - "maybe monetizing people's data?").

mei0Iesh 11y ago

It's the Flashification of the web. People who wanted to show off, or code web apps, used to use Macromedia Flash. People complained about it, in part because sometimes if you accessed a site without the Flash plugin you'd see a blank page. But Flash was great in many ways, and it was self-contained in objects, so websites were mostly still websites. JavaScript had been around for a long time, but there was still a cultural norm that most people respected about not requiring JavaScript. This was mainly because a lot of browsers still didn't fully support it, or people had it turned off. It was also when some people had cookies disabled. Then once Adobe bought Flash, and then Apple blocked Adobe Flash, it really killed the Flash way, and all that spilled over into HTML with HTML 5 and the new cultural norm of kids who are more concerned with showing off socially than the meat and potatoes of hypertextual information. It should've been obvious that there was a need for a new web, for code and multimedia. But in the .com boom nobody would dare try to start with something unpopulated, since it'd risk losing their chance at fortune. Today, instead, maybe we should go in the opposite direction and create a new old web; a hypertext network that specifically only works for HTML, so people can have this one to morph into an app network, and we'll not lose the text-linking place we've grown accustomed to.

hugozap 11y ago
> we'll not lose the text-linking place we've grown accustomed to. It's just a minority of users who are accustomed to the old web.
initself 11y ago
Rad idea. How would one initate a new network of this kind separately from what exists today? And what kind of interoperability would it have? Can't imagine it could go cold turkey!

BurningFrog 11y ago

Strong opinions on how things should be, but no arguments for why. That is never going to convince me.

savanaly 11y ago
It's because his post is entirely opinion, albeit one that frequently shows up at the top of HN comments sections because it's such a popular opinion in the hacker community that we forget it's not actually the majority's opinion.

frankensmith 11y ago

This comment reminds me of Maddox and his 90s' style html blog: http://www.thebestpageintheuniverse.net http://www.thebestpageintheuniverse.net http://maddox.xmission.com/ http://maddox.xmission.com/

userbinator 11y ago

Thus, all content is readable and styled properly without requiring an arbitrary code execution environment. That is what the web was meant to be. In other words, it was supposed to be a worldwide hyperlinked document library --- and we have mostly achieved that goal, although it is a library wherein you are constantly tracked and bombarded by books flying off the shelves at you, screaming at you to read them, and most of the books consist solely of ads with very little useful informational content. In my experience, it's the people who don't talk about web development who are the best web developers; these are the people who don't wince when they write a HTML document without a single `<script>`. Agreed completely. The ones who write information-dense HTML pages, often by hand, would not be considered "web developers" nor would they consider themselves to be; but they are what the web needs most. I've done that, and I don't consider myself a "web developer" either. it really irks me when web developers pretend like they're actually doing something good or useful, or that the web is actually in a healthy state I wouldn't doubt that they genuinely feel like what they're doing is good or useful; I've noticed the appeal of "new and shiny" is especially prevalent in the web development community, with the dozens of frameworks and whatnot coming out almost daily, proposals of new browser features, etc. Very little thought seems put into the important question of whether we actually need all this stuff. It's all under the umbrella of "moving the web forward", whatever that means. But I think we should stop and look back on the monstrosities this rapid growth has created.

hugozap 11y ago
Against what platform should we compare the web to decide what it should look like?. I think everyone has a different criteria.

ryandrake 11y ago

What we need is a new protocol: something that lets an author write and publish text documents, marked up with basic styling, with "hyperlinks" to other text documents. The protocol could allow embedding of simple inline figures as well. Users would run a "browser" whose function was limited to requesting and displaying these documents.

qu4z-2 11y ago
You mean like AMP? https://www.ampproject.org/ https://www.ampproject.org/ And yes, it's pretty silly that this is a thing.

username223 11y ago

> ... web developers pretend like they're actually doing something good or useful... Many "Web developers" are to "the web" as rapists are to sexual reproduction. They achieve short-term gratification and (perhaps unintentionally) pass their traits on to the next generation, but they couldn't care less about the longer-term consequences of their actions. The web started as "clickable citations." Now it's mostly hostile surveillance with cat pics.

ocfx 11y ago

> Unfortunately, most "web developers" have made the web worse over the past 10 years because simple, functional, minimal technology is not impressive, and hipsters love to show off No it's because when I go into work someone says to make it a certain way and if I want to get paid I have to. If you want to blame anyone blame designers who see proof of concept stuff from developers and throw it in designs.

ArleneBarlow 11y ago

The web got better according to most people. The devs responded to what people wanted, and that's flashy, interactive stuff. What the web was 'meant' to be is irrelevant, because the web has grown far beyond what it's original designer intended. On one hand you imply that it's good that the web is open, and part of being open is that people can make it what it wants, and not obey the original intention of the first designer. It's unfortunate that your vision of the web is in disagreement with the 'general users'.

Pxtl 11y ago

I tend to think rants like this are just being luddites, but all this JavaScript and external resources rest are ruining the experience of the web. Holy crap is everything slow now. On desktop we've all gotten into the habit of tabbing the new stuff and waiting for load while doing something else, but on mobile where that workflow isn't as easy and you have to watch a page load? I'd say HN is one of the very few sites I can stand anymore.

userbinator 11y ago
I encourage you to try the "NoScript challenge" - disable JS, enabling it only for sites you often use and which need it, and browse the Web for a week. If you tend to indulge in rich multimedia you're probably going to give up soon, but if, like me, you're just after text and image content the majority of the time, you might actually prefer it. Pages load nearly instantly (or not at all), there no more intrusive popups/popunders/slideovers/etc. and other annoyances like disabling right-click, text selection, or stuffing text into your clipboard. The "almost all sites have JavaScript and will break if you disable it" statement is common, and while I agree that the vast majority of sites do have JS on them, whether or not the stuff that will "break" on them if you disable JS is actually of any use to you is questionable.

TheAceOfHearts 11y ago

I think this is because so many websites try to shove ads and tracking crap down your throat. It's not even particularly difficult to pull off for certain websites. For example, my blog uses React and it does server-side rendering, so it'll work without JS at all. However, if you do have JS enabled, it'll let you avoid doing full-page reloads to navigate around. I totally agree that for content-focused websites, you tend to get a better experience by limiting gratuitous JS abuse. However, this isn't very feasible when you're building highly interactive applications. In the case of something like Facebook, they do have a version of their app that works without JS... But how many people can afford to maintain multiple versions of their applications? I think a good compromise is achievable by well documented APIs or even better with a public GraphQL schema! If I don't use magic APIs to build our frontend app, you can build a different frontend that's tailored for your needs.