Hacker News Reader

I'm sorry, but the web development experience is god awful compared to the 90s development experience using something like delphi or even (ugh) visual basic

the UI was consistent across the entire platform, all the accessibility worked out of the box

yes, the main advantage is that the web runs on "everything", but then the added complexity of needing 4 different programming languages and half a dozen build tools don't put it in a state where it has "never been easier"

even Java+Swing was better in pretty much every respect (other than providing employment for tens of "web developers")

For me, VB + Visual Studio was the highlight of late 90s programming - true drag&drop wysiwyg ui design, a clean link between code and ui, lightning fast prototyping. ahhh good times

It’s hilarious seeing people praise VB when at the time it was a running joke holding the same reluctant acceptance among devs as JS does today.

those people didn't try Borland's tools :)

(same as people who <3 visual studio today obviously haven't tried JetBrains...)

correct. as an industry we threw out battle-tested GUI toolkits and programming models for something completely backward. 15-20 years later we have excel in the browser. great. The no-deployment (for deployment is too often an after-thought of the main functionality) model was too attractive and blinded people to the costs.

The 90s were great, not because they had the best experiences possible at the time, but because there was such anticipation for the future. Sadly the web did not evolve quickly enough, and we have been paying the price ever since.

The only place where GUI apps have reigned supreme is in trading applications in finance, although I suspect that dominance is beginning to erode.

> the added complexity of needing [...] half a dozen build tools

You don't need those for the Web. You can choose to use them. You can choose not to, too.

The web enables the famed „write once, deliver anywhere“ which native desktop always aspired to be

We recommend viewing this sentiment in Chrome. It’s not been tested elsewhere.

> Browsers HTML and CSS parsers are incredibly generous in what they accept

Yes, that is part of the problem. Parsers should be fail-fast and deterministic, this is one of the many problems of the web. Developers don't notice their incorrect code due to this and without noticing rely on the implementation-specific way one browser handles the issue. Then 3 versions later or in a different browser they whole house of cards comes tumbling down.

> One developer can do with the web what 10 couldn‘t do with native

Building a non-trivial app is always faster to do native. By a huge margin.

> Parsers should be fail-fast and deterministic

Raise your hackles about the choice not to fail fast if you want, but when they fail, they do fail deterministically. It's a too-common myth that they don't. (A little perverse that people who make a big deal out of others unable to follow spec, but by doing so prove they haven't checked what the spec says—because if they did, they'd know this...)

> Parsers should be fail-fast and deterministic, this is one of the many problems of the web. Developers don't notice their incorrect code due to this and without noticing rely on the implementation-specific way one browser handles the issue.

This was XHTML documents when sent with the correct Content-Type. "Just don't display anything if the document is malformed" is worse for users if devs still create malformed documents.

The only reason developers write malformed documents was because they can get away with it. And yet, millions of lines of code are written every day in programming languages that absolutely do not accept malformed code, and that all works out just fine.

Another problem with it is that everyone who can get something to display in their browser suddenly thinks of themselves as a developer.

How does such a system reconcile the need to be opt-in to maintain backwards compatibility with the web as it exists with the likelihood that the audience you want to reach (developers who can't be bothered to Do It Right) are just the ones who will not opt in?

Ancient history now, but HTML5 was developed by WHATWG specifically as pushback against priorities of W3C at time (chief among which was push for XHTML).

That's not to say things cannot be improved, just to say this has all happened before and it will all happen again (and it is worth paying attention to what happened last time).

I think you misunderstand the comment you're replying to. He's not saying that the internet is bad, or that it doesn't make it much more convenient or profitable to deploy and sell software. he's just saying that's it's much easier to develop an application with say Windows Forms rather than HTML+CSS+JS

I'm not sure that's true if you consider the "last mile" of packaging and distribution.

The ability of web apps to be instantly installed and updated is a huge win, as is the fact that it runs on everything from a phone to a desktop with zero additional effort. Certainly it comes with tradeoffs, but if your goal is not just to write code but to put it in front of customers, it's hard to do better than Html plus JavaScript.

It's simply toxic behavior, to show up & flame something obviously quite capable like they had.

There's a lot of people receptive to this toxicity- it's fact-free rage people want to hear, it's viral & it spreads- but the truth is webdev works fine & we have many perfectly reasonable frameworks to pick from that do a very fine job. Even straight vanilla js is not unreasonable an option, especially with webcomponents.

I didn't see it as rage. It simply pointed out that the web as it exists is a document format that had self modification, and then full apps bolted on. It is a hacky mess and I don't understand how anyone just using js wouldn't find that readily apparent? There exists great tooling to hide that fact, but even then I have to agree with the gp that we still aren't to a place that is as cohesive or easy to use, or as predictable a UX for the end user as those 90s tools he mentions. There is no hate there, just hope. I am hoping that wasm and webgl can bypass html css and js for people who just want to make a webapp. Direct calls into the dom with no js wrapper should be here any minute, and webgl hopefully will follow? Once that is all sorted you'll see some development environments like delphi pop up I suspect.

> I didn't see it as rage. It simply pointed out that the web as it exists is a document format that had self modification, and then full apps bolted on

I dunno, it looked pretty aggressive to me. There was maybe a kernel of fact, but mostly opinion:

> As a platform for developing interactive applications the web is simply terrible.

Many disagree. Many would rather use Electron than native apps & find it nicer.

> There is no good way to develop a UI in HTML/CSS/JS, all attempts to do so rely on layers upon layers of wat are essentially hacks to build something resembling a UI toolkit on top of a document markup language.

I laughed out loud. There were already interactive buttons & form controls from the start. Vanilla JS is fine today. One of the web's strongest suits is that you can take a static page & sprinkle in just a little interactivity, just a little applicationization, if you want. Ultra-thin ultra-simple libraries like Enhance are doing great to provide just a little common shape & definition, tiny code but some norms & practices to adopt. It can be extremely incremental to move from document to application, with very little work at each step. We see what we want to see, and a lot of people want to see it as hard, as hugely complicated, but here's an industry heads down cranking without a ton of internal complaint, plenty of simple ways of doing things, and plenty of people doing fine.

> The result a UI that is unnecessarily complicated and fragile.

Sources wanted. Not all sites are perfect but a huge amount of things work very well. And unlike apps, users have a toolkit of capabilities they can learn & apply everywhere to help them (forward, back, bookmark, share link). Extensions let users go further than apps permit. Again, it's just, like, ya'll's downer opinion, man.

Anyhow, I like your note on hope. A lot. I don't see this ancestor post as having much hope there; it seemed set up to ruin. Trying to identify more constructive opportunities, trying to figure out new architectures we might layer onto the platform: that stuff should be exciting.

One place I lose excitement rapidly though is replacing the web with rendered canvases, as you say here:

> I am hoping that wasm and webgl can bypass html css and js for people who just want to make a webapp.

This is my nightmare. If folks want to do something else, fine, go somewhere else. But don't break the web. Don't break every user agent. Don't break view source, scraping, & accessibility tools. Just because alternate paradigms are succeeding as much as laserdisc doesn't mean they should come in & break the web next. That should be enragening.

The web is for more than pushing pixels into peoples faces. There's some value here, some shared expected experience, & coming in crashing it all because the only stakeholder you can imagine is those who "just want to make a [not-so-web]app" and damn all the rest.

I've had this bone to pick with Flutter, which is doing this imperialistic web takeover plan with CanvasKit, effectively turning the browser into a vnc session for an app. Incredibly hostile & deeply cruel way to snub hypertext & make yourself unintelligible to the rest of the ecosystem.

(The "direct calls into the DOM" wasm desires are going to be a bit chaotic & hurtful to view-source, but we've already made that hard; so ok, fine, and yeah, it is exciting to give people a post-JS web.)

I'd really like to keep both webs, but the reality is that everyone trying to monetize wants an app, and failing that a webapp that is a complete black box. So that is probably the way it will go. I like document only websites for conveying information. Know a lot about X and want to share your knowledge? Make a document. I can see room for some partial interactivity for hands on tutorials, like notebooks. But I don't really want any interactivity on my news sites, sadly they will head for the blackbox model first for the money. For a payment portal, a banking app, a multiplayer videogame, probably even a web ide or chat client, I don't want any document like behavior, I just want an app, and if it can be sandboxed in my browser, all the better.

Ideally I want security sensitive apps to turn into blackbox webapps. I want most everything else to be pure html+css. With interactivity there when you really need it to convey your meaning or to generate documents more quickly.

In reality, I think your nightmare scenario is far more likely. It just makes serving ads so much easier. And they still will keep making native apps for things that should be websites (which is my nightmare) because that makes it even easier to get valuable data.

> I've had this bone to pick with Flutter, which is doing this imperialistic web takeover plan with CanvasKit, effectively turning the browser into a vnc session for an app. Incredibly hostile & deeply cruel way to snub hypertext & make yourself unintelligible to the rest of the ecosystem.

This is nothing new. We used to have Flash, after all. (And I doubt Flutter on web will come close to reaching the ubiquity of that platform.)

Given how the web is embarking a takeover of desktop via Electron and of mobile to a much lesser degree with Ionic, to the annoyance of many users, turnabout is fair play.

Do you guys actually use the word disingenuous in face to face conversations?

Usually gets couched more softly for the person being accused of it, but when talking about statements it certainly gets used

I said “I think that argument is disingenuous, at best” to someone recently.

Their response wasn’t glowing.

I had a boss that said this in a variety of contexts at least once per week. He was fun…