Hacker News new | past | comments | ask | show | jobs | submit login
The web sucks. Browsers need to innovate (sachin.posterous.com)
76 points by a4agarwal on April 29, 2010 | hide | past | favorite | 68 comments

The web is the closest we've ever come to a standards compliant universal platform that is write once - works everywhere, period.

Last time I checked you have to be a paid member and go through an oft-criticized approval process to develop for Apple's proprietary app store platform on all iDevices.

Besides, Chrome OS and webkit are supposed to usher in a new era of enhanced HTML5/Canvas support including better access to the GPU. Google is betting on Web apps and I'm pretty sure Apple is not ignoring them with a standards compliant Safari browser on all of its devices.

So the web doesn't suck.. It rocks - and it will continue to do so for the foreseeable future.

"works everywhere, period"

Hehe. This made me chuckle a lot.

We are now entering an era where the hardware is suddenly significantly different, but your app may have to target all 3. As in an iPhone sized screen, an iPad sized screen and a monitor. Often you won't be able to use the same app for each one. The UX has to be different.

So the dream of write once - works everywhere is now at an end because of the devices themselves.

Well you make a good point but I'd argue that most web apps are easily adjusted to work well using Safari even on something the size of an iPhone. Take Hacker News for example - half the time I'm reading/commenting on my iPhone even though this site is not iPhone optimized.

Hacker news is on the low end of the webapp scale.

Easier doesn't mean "write once". Sure, write less, and a more pleasant experience than years ago. Definitely not "everywhere, period".

I think what the OP means is that the webapps have changed application model from client side to a more server side approach. UX needs to change across platforms (devices) yes, but the number of versions of application "out there" is just one (or a few, unless you're doing split testing).

"The web is the closest we've ever come to a standards compliant universal platform that is write once - works everywhere, period."

I think Java applications succeeded better in this regard. I remember hearing about fewer incompatibilities between JRE versions than I currently hear between browsers; there was never a "stop supporting JRE 1" campaign like there has been for successive versions of IE.

there was never a "stop supporting JRE 1" campaign like there has been for successive versions of IE.

there isn't a "stop JRE 1" campaign but there's a good amount of "stop breaking stuff with your updates, sun".

working in helpdesk a couple years ago, we needed to keep specific versions of JRE for specific applications. if you ever needed to use two apps that required two different versions of JRE, well....good luck.

From my understanding this is always because the developers did something they know they shouldn't have, like access private unsupported features of a specific version of a Sun JVM. I've seen quite a bit of java code written and all of it could run on multiple versions of Java with no problem.

Anyone who created Java applets heard lots of gripes about the Microsoft JRE vs Sun JRE. I wouldn't be surprised if web games implemented as applets are still using the 10 year old API of the old Microsoft JRE.

I still shudder when I think about fixing Java plugin installation issues.

Why does the OP compare the web to the iPad? The iPad is really just another standalone computer running a native OS like the Mac or the PC. In this context, he's really just comparing web applications to desktop applications, which is a comparison that has been made for years. Nothing new here.

Locally installed applications are slicker and more responsive than web applications. They always have been, yet developers have moved to the web in droves despite that. Why? Because the web offers other advantages, like being accessible from anywhere, for one. It's the closest thing we have so far to the dream of write once, run anywhere.

"People use web search today because they don't know how else to find high quality information."

That's what search is. No point in searching for something if you already know where it is.

"The web is a mess of content with no organization. On an iPhone, I launch the appropriate app"

Great idea, if you can organize every bit of info on the net into its own app. I wonder how many screens that would be on an iPhone?

I agree that browsers are constrained in how they can innovate, and it is a problem. But the web is not going away. Search is not going away. And iPads are never, ever going to replace the web.

>Why do all browsers have to support the same standards?

Written like someone who wasn't around for the mid-to-late 199s.

The problems with browsers compared to web apps are well described by the OP, but I just don't understand the author's solution. Have everyone do their own thing? Make me download an app for every website I visit?

This seems like the worst possible outcome:

> Why do all browsers have to support the same standards? This only limits their innovation, and limits web developers.

Shouldn't the solution instead be a new, more powerful standard? (Ignoring how pie-in-the-sky this is.) As in, not band-aids like AJAX and Flash, but a from-the-ground-up standard? After all, 99% of iPad apps have the same basic capabilities; there's no reason you couldn't have a standard which replaces HTML which could fully cover those apps. Then, if you have a need for a really specialized service, you can create a dedicated app.

Ah yes, the new more powerful standard. Agreed by Google, Microsoft, Apple, Oracle, etc, etc, etc, etc.

Standards are for lowest common denominator and almost by design are never, ever simple enough - see e.g. SOAP vs REST, XML vs JSON (Dave Winer be damned).

I'm not sure what your point is. REST [1] has beaten SOAP as the de facto API standard, and JSON has beaten XML as the de facto standard for portable data structure.

[1] Notwithstanding: http://news.ycombinator.com/item?id=1304244

I think the point is the "winners" have something in common and this is no accident.

XML and SOAP came out of "let's invent new formats and protocols"-type standardization, while REST and JSON were products of (essentially) one guy looking at a simple, existing, well-known concept (HTTP and js object literals, respectively) and putting a new frame around it.

I understand what you're saying and broadly agree with you; though I'd point out that Roy Fielding, the guy who articulated REST, was one of the principal authors of the HTTP specification.

Article full of emotions, disconnected ideas and "facts", no real arguments :(

Good luck looking at tiny progress bars when upgrading your ABC app, I'll just press refresh.

Oh, and how do you intend to share the ABC news stories with your friends? abc://article ?

I don't think you've tried the app he is talking about. The ABC app for iPad is for watching the streaming video of recent shows. It has nothing to do with articles or sharing.

Call me old fashioned, but when I want to share something with a friend, I either tell them about it myself when I see them, or worst case scenario I shoot a quick email that says "Check out the new episode of Castle on abc.com!" The only people I know who wouldn't be able to parse that simple message and then go to the website and watch the show also don't check their email or even have an active account. Not everything has to be an automatic, shareable mini-link that can be pasted into a twitter stream.

Wow, really?

I guess hyperlinks aren't required for sharing, but they are so cool. The fact that on the web you can try a service just by following a link that a friend mails you makes things so incredibly frictionless. I supect that conversion for "just go look on abc.com" is a lot worse than "click this link".

Same thing for advertising. If there were no hyperlinks, you wouldn't be able to track conversion on your ads. That seems important.

Not to mention the fact that URLs are unique identifiers. So you can do analytics and find all the referers to your app/movie/article/whatever.

When I got to the part of the article where he wrote people are using more Apps these days than web sites, I rolled my eyes.

Web apps aren't going anywhere any time soon. Do you really want to install a Ticketmaster app and wait while it downloads just to buy tickets for one show?

This isn't true. There are numerous toolsets that allow you to have desktop grade applications. Fluid dynamic feel, drag & drop, etc. I'm the co-founder of one of these toolsets, NOLOH (http://www.noloh.com), but there are others such as Cappuccino, Sproutcore, qooxdoo, etc.

With the above tools you literally just sit down, start writing your application and forget about browser issues, or limitations. The issue isn't that you can't do it, but that people aren't.

Once we accept that the issues mentioned in the article aren't in fact issues anymore, since many toolsets have solved these problems, we can then sit down and spend a few minutes watching a video, or reading an article and help the web move forward. We don't get anywhere by saying we can't do things, because the fact is, we can. You just have to use the tools available to you.

Also, I would say that if the absence of "desktop grade" UI functions makes the web dev's focus on the app's core functionality at the expense of eye candy, then I'm all for it. But then, I don't use drag & drop too much myself.

When GMail launched in 2004, it took one step forward and 10 steps backwards from the mail application I was using. Even today, the major features GMail is releasing are simply trying to match the features I've had on the desktop for years.

Sachin, isn't that true about Posterous too?

While I agree with the points you make about the weakness of the web, you are missing one key point.

The web makes it possible for the end user to stop thinking about the OS and installing stuff. If we had to move to the model you are suggesting, people would have to start choosing browsers. It would be sad if I couldn't use Posterous because it doesn't work on the Ubuntu version of Firefox.

Besides, considering most people don't even know what a browser is, why would you want to make them install 5 different ones? It will only bring back the IT department.

When we launched posterous in 2008, yes it was missing a bunch of features. But look at us now (or actually, in a month, we have a lot of stuff about to launch). We'll have everything all other blog platforms have, and more.

But SIX YEARS after gmail's launch, they can't say the same.

For Posterous, and any other web app, we'd have to decide the tradeoff. If we focused on one browser, could we add enough value to that experience that it's worth dropping support for Ubuntu Firefox?

Each browser and each developer should be able to decide this, instead of being slowed down by the platform as a whole.

If you really wanted to head your advice you could focus on one single browser that works on all operating systems and have your users download that as their app.

Come on, are you really suggesting that gmail didn't add any value and is just playing catch up with old email software? Free unlimited email storage? Auto-saving and suggesting contacts? Threaded email conversations? Big attachments? AJAX? Email search that works?

Each browser and each developer should be able to decide this, instead of being slowed down by the platform as a whole.

You know you do have the freedom to pick one browser. But you won't because it makes no business sense, unless you are talking about innovation in a vacuum.

We are re-inventing the wheel with the web to some extent. But it is already beginning to get very good very fast and Posterous is a good parallel of that phenomenon in the blogging world.

I would go farther and suggest that GMail was designed with the intention of making the desktop version of email obsolete.

They added in some desktop features primarily (IMHO) to support users who were so ingrained in their usage patterns that they couldn't see that really good search was actually BETTER than putting your email into folders. Desktop apps are only just beginning to catch up with faster search--and they still don't do it as well as Google.

So can you tell me which email client out there works as consistently same on Linux, Mac and Windows as Gmail does?

I feared my comment would come across as a disguised ad hominem attack (thanks I didn't know there was a name for that!). But if you read the whole comment, it should be obvious that it's not in fact Tu quoque because I do address Sachin's position. I used Posterous as an example to argue that people don't make software just to innovate; there are other things to consider such as market trends and user needs.

You are right, I don't think it was an ad hominem, besides you made a valid point about Posterous. I just read the wiki stuff today and couldn't resist.

"People use web search today because they don't know how else to find high quality information. The web is a mess of content with no organization. On an iPhone, I launch the appropriate app."

How is launching the appropriate app a way to find high quality information? You would have to know which app contains the information you're looking for, in which case you already found it.

I would rather do web search and get a direct link to the information I'm looking for than search the app store for an appropriate app and then have to find the relevant information within the app.

> When I started writing this post last week, it was going to declare the end of the web. I'm not quite ready to call it dead, but it's on thin ice.

That's where I stopped reading and lol'd

IMHO one of the main reasons for the success of OS X and the iPhone/iPad platform is due to web browsers. Imagine a world where web browsers didn't exist but only specific "thick clients" that used the Internet as a web service to access and store data. Almost all these clients would be originally developed for Windows (Flash player was) with Mac and *nix as an after thought. What if you couldn't access Facbook, gmail, Twitter, etc through a web browser? Without the ability to do your day to day activities on a non Window's platform, how many users would make the jump to a Mac or an iPod/Phone/Pad?

To me, this feels like the difference between smartphone and desktop/laptop kinda boils down to the Appstore being on the homescreen.

In a way what he's asking for is already there, Adobe Air and Silverlight.

They can both do exactly as Sachin asks for, web focused, quick downloading, rich applications, just like iPhone apps. Fantastic dev environments.

But how do you find those apps? There's no Air store on every desktop. MS could never put a Silverlight store on every desktop without being slapped silly with monopoly charges. Apple are never gonna put an Adobe Air store on every Mac.

What he's asking for, in a way, is impossible, a new type of browser won't fix it. In reality only Microsoft or Apple can. Or a totally new OS. And I doubt Chrome OS is going to go in that direction.

This guy is so wrong; everything he describes is already happening; he just doesn't see it. There was stagnation for a few years which we're still catching up from, but these days we have innovation out the wazoo.

Take WebGL as an example: it started as a Firefox developer's side project, and competed for a while with other 3D APIs like Opera's simple 3D and Google's O3D. Then Khronos got involved, a standard got written, WebKit got support, and next year we'll likely have 3D games written in Javascript running on our phones.

What part of that process is broken? Seems to me as if it's working perfectly.

I've seen this argument so many times I've lost count.

If you compare an iPad app against a website, yes, the iPad app is better for the amount of effort put in. Now go and make an iPhone OS app, an Android app, a J2ME app, a Windows app, and a Mac OS X app, and you still won't reach as many users as with a single website.

Fracturing the web with browser-created SDKs will break the foremost feature of the web, and quite frankly, we've been there before and it didn't work out that great. Does any developer miss 1997, browser wars 1.0?

The author is oblivious to the irony that he is publishing his rant on the exact platform that he chooses to vilify.

The fact is that the web is a monstrous success at its purpose - to share information. The fact that it isn't rich enough for his taste just shows exactly how great it is. He is taking the entire thing for granted, and whining about what really are extra features.

Could we find something better for some apps? Sure, but that in no way invalidates the gigantor success of the web as a whole.

The author's idea essentially equates to taking the proprietary-ness of plugins and expanding them to the entire browser application. Instead of having a Silverlight plugin and a Flash plugin, we'd have an Adobe browser and Microsoft browser (that's even more proprietary than the current one). At least, that's my idea of it. If the author is suggesting a proprietary application to represent a single web site, that won't fly.

I agree on most points - if we want to see really rich web apps, this would be the way to go, and despite existing standards every browser supports proprietary stuff anyway, so we might as well stop fighting it - but the main problem is that users will no longer have a single blue E or swirly red dog on their desktop that they can rely on to bring them everything on the web.

Couple this idea with some standards about how all of these different browsers will be made available and standards that require one browser type to be able to open another browser type if the target content better supports the other browser, and maybe we could get somewhere.

The web "sucks" compared to iPad apps because those apps aren't overrun with seizure-inducing ads and crammed with superfluous SEO-optimized content.

IPad apps are designed exclusively to focus on the content or application.

The same is true for iPhone apps. Limited screen real estate forces developers to focus on the usability of their core product.

For example, media iPhone apps focus on the content (i.e., the article or the video). Media websites focus on selling display ads, driving traffic to other sites owned by the same parent company, and forcing as much content as possible onto the page to draw SEO traffic.

So the iPad appears light years more advanced but it's not. Its presentation is just more focused.

We already tried that. Remember the late 90's? IE4/5/6? Microsoft via Windows/IE are still in the best spot to push the proprietary web. They may have only 50% market share but that's quite an advantage over 15% for WebKit and 25% for Firefox unless they combine forces. As long as Windows remains dominate on the desktop ~90% of people can simply click on IE to access an IE-only website. Before too long they'll forget Chrome or Firefox ever existed. Few sane developers will exclude 50% of the market.

I highly doubt that people who have tried chrome or Firefox will forget it existed. they may not use those browsers at work right now. but corporate apps are changing rapidly towards true open web-standards-based browser solutions. and as that transition accelerates its hard to see how the fastest browser doesn't win. my money's on chrome.

But Windows only dominates on the desktop. On the mobile side, it's Opera and Safari, plus other WebKit browsers. That's impossible to ignore.

People like apps because installing and opening an app doesn't make them feel stupid. Remember the ReadWriteWeb Facebook-login debacle? Even the users who didn't make the same mistakes as those folks still feel frustrated with all the steps and layers in between them and what they want. And don't get me started on URLs.

People don't want browsers - they want Facebook. They want YouTube. We don't need better browsers; we need the Cocoa API to be cross-platform (and stop sucking).

I'd disagree, I think its more about how programmers are using the web. The plethora of easy web languages is the problem - they've encouraged programmers to develop the wrong solutions - webpages based on a towering technology stack, instead of a "proper" program to be run on client machines that makes use of the internet to support its features. Of course this is a lot harder than producing a website solution, there are platform issues etc... but its hard to deny that this allows you to produce a higher quality app - you basically get complete control of everything that a browser or Flash or whatever will do for you.

i.e. The web doesn't have to change to get the result this article seems to want - just the plethora of crappy programmers who abuse it.

I know this is heresy. but someone should start working on the next thing. Gopher got replaced.

Warm up your editor/IDE.

I'm pretty sure I am not that smart. I always thought some persistently-connected, Display Postscript like thing would do it, but it would have to be designed for the interactive from the ground up.

Then find someone to fund who can. It doesn't do any good to sit around barking at the world to do things for you.

I think my partner on my current project would have some words... :)

"on April 12, Steve Jobs said, 'Search is not where it’s at...' Full six years ago I blogged about people's over dependence on search"

Mate, unless you're the same Sachin that bats for India I doubt people are paying quite as much attention.

I think he's right that browsers need to innovate, not by innovating new ways for web apps to do things, but by innovating new ways to interact with existing web apps. There's a whole load of potential integration between browsers and web content, and while there are various browser based twitter/whatever clients and whatnot, most of them feel tacked on rather than integrated. The internet tends to feel like a giant cobweb, rather than a woven spiderweb.

I really couldn't disagree more strongly with this. The advantages of web-based apps (distribution, network awareness, persistence etc) already outweigh native apps for a lot of applications and this will only be more true as HTML5 compliant browsers become commonplace.

If anything it's the complex content creation apps that devices like the iPad seem to eschew that will be last to go online.

The OP says: "Each browser should focus on innovation, not parity."

If that's the case, developers won't bother using those shiny new proprietary features. We're in the business of making products, and there's no way most people are going to waste their time adding tons of browser-specific features.

The fact that the browsers don't have parity is probably the most time-consuming thing about web development.

I could refute almost every single point in the first set of arguments.

go for it

WebSocket, Canvas, Web Workers, HTML5 drag and drop API, WebGL, O3D, Application Cache API, Chrome/FF (and possibly other browsers) allow web apps to open mailto: links, FireBug, Web Developer Toolbar, etc. etc.

Unless you actually refute EVERY SINGLE OBJECTION, apps still win, at least in the user-experience sense. Something as small as not getting an hourglass when the app is "thinking" is a bug, which every web app has right now.

Well, that's pretty subjective. I prefer GMail's user-experience than any competitive desktop app, same for Google Calendar, Google Reader, Basecamp and FreshBooks. I guess I'm biased as I am a firm believer that the open web is the platform of the future. That being said, I totally agree that there is still a lot of room for improvement, but we're on the right track :)

It's pretty trivial to turn the cursor into an hourglass with CSS. That isn't a limitation of web applications.

Then why do nearly none of them do it? Probably because they can't judge when they will stutter and lag - so many layers of web indirection. The fact is, most web apps only a mother could love. The rest of us are sitting stunned and silent when someone shows off their latest "jewel" which stutters, lags, has buttons that don't react when pressed, takes 3-5 seconds to put up/take down a dialog based on a server database query...

I don't know what most of those things are, and I'm guessing most of the world doesn't either.

Most people don't know what HTML is, yet they can use web browsers. Not knowing HTML 5 will not prevent them from using HTML 5 -- and HTML 5 is covered by the first half of that list. The mailto extensions don't have to be -- they could as easily be built into the browser (and should be by now). Web applications can already be online/offline hybrids (Google Gears already provides the functionality, HTML 5 adds it as well).

"On the web, people still use plain text editors."

I use DreamWeaver, which I happen to think is quite nice.

This is just about the stupidest thing i have ever read.

From one of Sachin's comment 'Users would want to use your product and therefore they would switch to Chrome. Other browser would lose market share and either implement whatever makes Chrome better for you, or do even better.'

Rebuttal: http://www.youtube.com/watch?v=o4MwTvtyrUQ

nuff said

So we're back to web apps vs. desktop apps? An iPad app is installed software running against a native OS API, so I'm not sure why it's singled out when it's just another form factor for a desktop.

Browser based software sucks for a few reasons, but I'm not sure the most glaring were mentioned. It's also fantastic for many other reasons some apps are awesome on the web.

You can write a photoshop app for the browser, but why would you want to? Certain apps belong on the desktop, so let's not blame the web for that...unless you're one of those people who thought desktop apps were dead.

The world sucks. It needs to be like EPCOT.

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