Chrome 70 flips switch on Progressive Web Apps in Windows 10 – with janky results

Version 70 of the Chrome browser has begun to slither onto Windows 10, bringing with it Google's desktop take on Progressive Web App functionality. Google has been banging the drum for Progressive Web Apps (PWAs) for some time, initially seeing the tech as a way of improving the user experience of web apps before suggesting it …

[Anonymous Coward]

Sub Optimal UI?

If all it is is a web page in a frameless Chrome window, then it's still going to look like a web page. That is, it'll look pretty rubbish when a page is trying to pretend to be an application with dialogue boxes, menus, etc. Who has ever heard of a dialogue box for a proper native application that cannot be moved outside of the frame of the application's window?

And another point, boring but important, is that in company environments where there's people who aren't especially IT literate, having a consistent style and way of doing things like accessing menus, and so forth, is actually quite important. Having all that consistency tossed in the bin on a regular basis because some web developer has got the horn for some latest fancy toolbox on NMP is not helpful...

[Teiwaz]

Re: Sub Optimal UI?

Who has ever heard of a dialogue box for a proper native application that cannot be moved outside of the frame of the application's window?

Yeah, but it's safer that way.

Who really wants a return of pop-ups???

The next version of HTML will may well be more suited to webapps of this nature, but regular users are often tricked now by pages trying to pretend to be OS level dialogs...

[stephanh]

Re: Sub Optimal UI?

There are tons of applications now being developed using Electron, which means it also lives in a Chrome window. Even Microsoft does it with vscode. It doesn't seem to bother most people.

[Anonymous Coward]

Re: Sub Optimal UI?

I feel pretty negative about Electron apps, though. They lack spark. They’re a bit undercharged.

(But seriously, they’re not that great, really. And it does seem somewhat wasteful to have to bring in half a web browser just to make an app.)

[leexgx]

Re: Sub Optimal UI?

seems to have broken adblock as well (had one of my customers PC running adblock a loop in turn braking websites )

about to hit the restart chrome button now see what happens to mine

[JohnFen]

Why I'm not interested in PWA

Browser-based apps are much more difficult to create specific firewall rules for. Since making firewall rules on a per-application basis has become pretty much mandatory, this limitation is a showstopper for me.

[Anonymous Coward]

Re: Why I'm not interested in PWA

Presumably the idea is that the data being processed by such apps actually resides on some cloud, and isn't held locally at all, in which case the purpose of the firewall is somewhat different?

[JohnFen]

Re: Why I'm not interested in PWA

A major part of the sales pitch for PWAs is that they work in the absence of an internet connection, though. If they only work in connection with the cloud, then that's another showstopper.

[katrinab]

Re: Why I'm not interested in PWA

Outlook Web App works without an internet connection, however if you want to send or receive emails with it, that won't work, just like the Windows / Mac / IOS / Android versions of the app.

[DaLo]

Re: Why I'm not interested in PWA

Well a PWA is still just a web site nothing more. It can utilise hooks that can do some 'clever' os level stuff like add link to your homepage etc, but these are dependent on your browser and OS. So access to sensors and hardware has been granted by the browser so any app, whether it is a 'PWA' or a web page can access it.

Therefore your firewall ports will be as useless against PWAs as they would against any old web site. However blocking access to specific sites and to remote hosted data stores is just as easy with a PWA as another website.

As for offline/online. That is completely up to the developer - they can use web workers or service workers to allow use of a cache api or small db to do some offline work. Often this is regarded as a temporary storage state which will sync and clear down once an internet connection is achieved.

[JohnFen]

Re: Why I'm not interested in PWA

"Therefore your firewall ports will be as useless against PWAs as they would against any old web site."

Yes, this is the main reason I'm not interested in PWAs.

"However blocking access to specific sites and to remote hosted data stores is just as easy with a PWA as another website."

Correct, which is entirely insufficient.

[DaLo]

Re: Why I'm not interested in PWA

So what do you currently do with websites? Do you block browsers on the desktop or just whitelist/blacklist individual websites?

I don't fully understand what the issue is? PWAs don't get admin level control, they can't open up ports on your machine at random, any ports they send out on can be blocked, there is almost as much control over malicious websites as there are malicious programs, and more control over categorized websites, whereas I'm not sure a categorized system exists for applications.

So you can block PWAs globally or individually or block access to the web completely and restrict their remote connections and activities, this seems like quite granular control and would seem far safer than an application that has to be installed (and therefore has admin privileges at that point).

[JohnFen]

Re: Why I'm not interested in PWA

"So what do you currently do with websites? Do you block browsers on the desktop or just whitelist/blacklist individual websites?"

Neither, but I also don't allow them to run Javascript or any other client-side code by default. I do allow a few specific scripts on specific sites, but no site gets a blanket pass.

"any ports they send out on can be blocked"

And what if they're using port 80? Currently, I use my firewall to ensure that no application can use that (or any) port without me intentionally allowing it. However, if the application is a website, I can't do that (without a lot of hassle) without also blocking the browser itself.

"more control over categorized websites"

I don't know what you mean by this.

"So you can block PWAs globally or individually"

Well, I certainly will (and do!) block them globally as I don't allow client-side scripting. What method of blocking them individually do you suggest?

"I don't fully understand what the issue is?"

PWAs are difficult to block with a firewall, and my firewall is my primary defense against the misbehavior of applications. Without at least that much protection, PWAs are too dangerous to allow.

"would seem far safer than an application that has to be installed (and therefore has admin privileges at that point)."

How so? No application I install has admin privileges unless I give them such privileges, and all applications can be easily blocked from sending or receiving data through the internet.

[bombastic bob]

Wheeee.

underwhelmed indeed. I toyed with something *like* that on a droid SEVERAL YEARS AGO, but it actually displayed a web page in a way that made it _look_ like you were running an application. But it was a specialized web server. The alternative had the web browser decorations which took up screen space.

The 'droid application was something like 5 lines of Java code, and a hard-coded web server URL. Good enough for a prototype. But yeah it needed a customized web server to work. That's the REAL magic.

[J27]

Re: Wheeee.

Web apps packaged as mobile apps is a big thing now. Apache Cordova or a hybrid framework like React Native or NativeScript allows you to easily port your web application code into a single-page application-based mobile app. If the developers do their job right it can be very hard to tell the difference between one of these and a native mobile app.

[JohnFen]

Re: Wheeee.

"Web apps packaged as mobile apps is a big thing now"

Yes, and that's one of the reasons (but nowhere near the biggest reason) that I've stopped even evaluating new apps anymore. I'm really looking forward to getting off of this "mobile OS" train entirely. It's become something of a nightmare.

[Anonymous Coward]

Sounds like a new and exciting attack vector for spammers, phishers and other scumbags to use.

A malicious website pops up a malicious webapp that are disguised as your normal native apps instead of being clearly a website.

[Anonymous Coward]

Made all the easier because the code is available in source form from a public website?! It sounds even easier than lifting APKs from Android.

[MonkeyJuice]

Yes, but then it is still a website within an application frame.

[Anonymous Coward]

It can't just pop-up a malicious web app.

It's no more risky in that regard to a copy cat website or a website offering to download and install adobe-apdate-v14.5.exe

[Anonymous Coward]

In fact a lot less risky than that as the PWA will not have be allowed admin access to the system or be able to escape its sandbox.

[Barry Rueger]

Please, give me an OFF switch

I can see no way this will be a good thing, if for no other reason than the likelihood that Google will break or turn off whatever PWA you've built your work flow around.

More and more I really do prefer to keep data and significant applications as local as possible.

[J27]

Re: Please, give me an OFF switch

Nah, I'm sure it will be just as safe as relying on Chrome apps. Oh, wait...

[nematoad]

Re: Please, give me an OFF switch

"Linux and Mac Chrome users will be able to join their Windows 10 and Chrome OS..."

I think I'll pass on that "opportunity". Google are intrusive enough already so why would you trust them with all your applications and data? The question in my mind is: whose computer is it if you start using PWA? Not mine I think.

So thanks but no thanks. My PC will remain my "personal computer" and not some data syphon to Google so that they can sell me to their users.

[Milo Tsukroff]

Probably won't work when Windows is done with it

> more than once the icon in the taskbar inexplicably switched to one

> for Google Chrome

Well now, could it be that Windows doesn't play well with it? My prediction, based on decades of Microsoft practice, is that the icons won't work when Windows is done messing with it.

Oh I know, Microsoft used to claim that the phrase, "DOS is not done / Until Novell won't run" was an urban myth. Until the discovery phase of a lawsuit against Microsoft dug up the smoking-gun email that said pretty much that.

[Bomb icon because I predict that's what Microsoft will do to the PWA taskbar icons.]

[dharmOS]

Re: Probably won't work when Windows is done with it

Microsoft is one of the biggest fans of PWA. Mainly to make up for the lack of native apps on Windows 10 and its now dead W10 Mobile. UWP is dying or dead, and programmers never really bothered with it.

I have tried some of the PWAs downloaded from the Win10 store and they were underwhelming with ads that could not be blocked.

[Phil Kingston]

Re: Probably won't work when Windows is done with it

>ads that could not be blocked.

Ah, now I see why G so keen on them.

[Anonymous Coward]

Fixing a problem that no longer exists?

Web apps were mostly hated the first time around because they used Javascript, and between slow CPUs in mobile devices, and crappy software engines everywhere, they were sluggish and crash/hang prone.

Fast forward to 2018, and now mobile CPUs are much faster - pretty much on par with desktop CPUs in the case of the A12 - and Javascript engines in browsers are far more efficient and stable.

So why do we want to change to a totally new way to implement web apps, where we will have to make and then correct all the same security mistakes we make time and again every time the wheel is re-invented? Can anyone tell me why I should prefer PWAs over a Javascript web app?

Seems like it is one more thing I need to disable for a few years until they work out the kinks, and re-enable if/when it actually becomes a thing. After all, Google is behind it, and they axe about half the things they do a few years down the road. There's no reason to believe PWAs will survive.

[Teiwaz]

Re: Fixing a problem that no longer exists?

There's no reason to believe PWAs will survive.

Because there's no reason to believe Javascript won't be considered as out of date as the likes of Fortran or COBOL then either.

And we'll be on some digital 10th generation esperanto that's more efficient with agile or whatever 'new age' business coder regime that's replaced it.

[Mat Bettinson]

Re: Fixing a problem that no longer exists?

> There's no reason to believe PWAs will survive.

Actually, this isn't the first time Google have done this. Chrome Apps but those were locked to their store and exposed a number of Chrome specific APIs because web standard APIs at the time had some shortcomings, like being able to obtain unlimited storage. I wrote a substantial Chrome app for annotating speech, only to have Google make the entire platform redundant.

While that might seem that Google could just axe PWAs too, PWAs are not a Google thing, it's a web-standards thing - and lots of the key parts of these new standards are already deployed widely. Notifications, for example. All depend on the background JavaScript code called a service worker, and now you'd be hard pushed to find a popular site that doesn't use this.

The 'progressive' part of PWA means an app that progressively takes advantage of any exposed modern features. There isn't any difference between a Javascript web app and a PWA.

If I have a criticism of Google, they retired Chrome apps two years before they introduced the same basic feature for regular web-standard apps, and that sucks.

We should applaud PWAs. One-click near instant installs, inherent cross-platform compatibility, and inherent sandboxed security. They wont do everything you have a native app, but most of the apps we run for things are well within scope. The Chrome PWA proposition is really just a convenient wrapper around conventional web apps, letting you launch them directly from your OS without loading a browser bookmark. Where's the problem?

[Anonymous Coward]

Re: Fixing a problem that no longer exists?

Ah, service workers. Glad to know that, as I've already disabled those because I was sick of seeing all those stupid requests asking if a page could send me notifications.

PWA is off to a pretty bad start if it is building that on crappy foundation, because everyone with a clue will have already disabled what it needs to function.

[Anonymous Coward]

Re: Fixing a problem that no longer exists?

I'm not sure how many sites you're visiting that make you sick of seeing those requests. However,

Blame the site if you don't think that request is necessary

Just press the block button on the first request if it is a site you never want to have give you notifications

Notifications could be very useful - they're used on a vast percentage of mobile Apps and they rarely give you the option to disable in advance. Allowing PWAs to give you notifications via the OS allows you to close the web page and still get notified. So e-mail or messaging apps can work as PWAs, Alerts can come as a notification rather than bundled with all your e-mails, you can be notified if a post you subscribe to is updated,, etc. You may well not want these, hence you have the choice to block them.

But they can be useful for some, just as they are on a mobile device for normal apps.

However, maybe you can explain why it is such a 'crappy foundation' rather than the remark of "anyone with a clue, like me because I am better than most..." type remark.

[Anonymous Coward]

Re: Fixing a problem that no longer exists?

I kept getting asked by some sites on EVERY visit, so no it isn't a block once and you never have to again. The number of sites where I might possibly want such a notification (i.e. maybe at El Reg if someone responds to my comment) are dwarfed by the useless notifications sites will want to send (i.e. an evil El Reg could send a notification every time they post a new article)

I disable notifications for most apps on my phone, because of the level of nonsense they want to notify me about - up to and including the most annoying one possible - "you haven't used xxx in a while, we miss you". If I haven't used an app or visited a web site in a while, that's my choice, I don't need or want to get notifications so they can whine about my absence!

So yes, building on the foundation of something that's first use people are exposed to is something so highly annoying and mostly useless will result in people permanently disabling the source of the annoyance. Which will be a problem for PWA. Just like how early Javascript in pages was so slow and problematic a lot of people browsed with Javascript disabled for years - and some still use tools like Noscript to selectively control it.

And I still didn't hear anyone come up with a single reason why PWAs are better than current web apps using Javascript...and I notice you didn't, either.

[Anonymous Coward]

Re: Fixing a problem that no longer exists?

You didn't say which browser you are using, however it sounds like a bug in your browser if you can't block notifications and it may be worthwhile filing a bug with them. In chrome for example just visit chrome://settings/content/notifications and it will show you all the sites that you have blocked or allowed notifications from. You can also block notifications globally for sites within there.

If some site is bypassing that I would suggest reporting it either to the side admin or not visiting that site.

You don't seem to understand what PWAs are or what 'the foundation' is. If you feel the foundation is a notification function then it isn't. There is no specific foundation although service workers is sometime consider the base for it. However even that is a collection of facilities enabled by the browser and OS rather than a full library.

When you ask about why PWAs are better/worse than web apps using Javascript you seem to think they are different? PWAs are web apps using JavaScript. The term PWA is just used as a term which brings a set of standard technologies together and a roadmap for these web apps to be interoperable on multiple platforms as and when the features exist. This saves you have to use many libraries and wrappers to try to check for functionality on the thousands of combinations of browsers and OS that you might encounter.

The advantage of PWAs/WebApps for an organisation is that, in theory, they can combine a responsive website, a mobile app and a desktop application into one single platform and development environment with your standard web development team. No need to utilise multiple app stores, develop different versions for each OS and updates to your code are instantly available for test or live at the same time across all platforms. Imagine the scenario where as a company with an e-commerce platform you wanted to update some of the major technology running the backend system and you utilise Apps on Google and Android. At some point they are going to be out of sync and possibly break one or other of them. With a PWA they would all be updated at the same time (barring caching).

App vs PWA generally revolves around the extra facilities available to an App but we no longer debate creating desktop applications vs a website for almost any non-software business. The days of downloading an application to browse an electronics catalogue are gone and everyone has switched to providing that functionality through their website. Similar could be available for PWAs when compared to Apps.

It won't suit everyone for sure, but there is logic in it.

[nematoad]

Re: Fixing a problem that no longer exists?

"Where's the problem?"

Google.

[JohnFen]

Re: Fixing a problem that no longer exists?

"We should applaud PWAs."

I disagree. I think they're the next step down a path we shouldn't be walking.

[Fungus Bob]

Web apps that don't need an internet connection or visible browser window? Sounds like native apps to me.

[Lorribot]

MS spend a lot of time on their UIs and compared to the amount of time most Web designers spend it shows.

You only have to look at The Registers web site to see the problems with PWA. I like the vast majority of users have a widescreen monitor and yet the web designers only allow the web site to fil the central third of the screen .Why? Nobody remebers, maybe because someones great aunt had a 640x480 CRT monitor once. Why not go out on a limb and offer a widescreen version of web site? Or even let html flow the text to the width of the browser...just saying.

Web designers are generally rather rubbish at UIs. and when you throw in Windows, Mac, Linux, Android and iOS backdrops anything they come up with will be jarring on most at best.

What is needed is a system CSS thingy that designers can call that styles these PWAs in a way that emulates the narritive of the OS. However that would require a standards body to oversee and all OS providers to implement and given the still patchy implementation of HTML and web designers that know better i feel it is unlikely to ever take off.

[Carpet Deal 'em]

> I like the vast majority of users have a widescreen monitor and yet the web designers only allow the web site to fil the central third of the screen .Why? Nobody remebers, maybe because someones great aunt had a 640x480 CRT monitor once. Why not go out on a limb and offer a widescreen version of web site? Or even let html flow the text to the width of the browser...just saying.

There are a lot of screwed up things UI designers do, but this isn't one of them(which is to say, it's time-tested advice). Excessive width is tiring on the eyes, whereas vertical reading isn't much of a much. Having to read sentence after sentence across the width of an ultrawide monitor would just be a usability nightmare.

[JohnFen]

"but this isn't one of them"

I disagree.

[Anonymous Coward]

I have widescreen monitors too, but like an awful lot of people I have my browser sized to take up about 2/3 the width - because that's a comfortable size/aspect ratio.

Monitor size is irrelevant; it's viewport size that matters.

I have some clients who run everything full-screen, on 23" monitors, because otherwise they 'get distracted' by being able to see other things. Thankfully not everyone gets distracted so trivially.

[katrinab]

That is because if you have too many words in one line, it is very difficult to find the beginning of the next line. The text size would need to be about 100pt before it would make sense to have the text go across the full width of a full HD monitor.

[DropBear]

Bollocks, to the lot of you. Full-width is exactly how the internet used to look like back in the pre-2.0 good days when its biggest flaws to complain about were blink tags, animated gifs and page backgrounds resembling Outlook stationery. But you did give me an idea, so I tried looking at a few articles here isolating and widening the article text only - and it's simultaneously shocking and depression-inducing how massively better it looks that way, instead of the currently ubiquitous "fuck two thirds of your screen, mate!". Actually I might just have to add a few lines to my styling script to try showing the whole site permanently like that. Oh, and guess what - I could still read each and every line just fine. Turns out all you need is knowing how to read. Astonishing really...

[katrinab]

"back in the pre-2.0 good days when its biggest flaws to complain about were blink tags, animated gifs and page backgrounds resembling Outlook stationery."

Doesn't sound like the good old days to me.

And you forgot animated backgrounds, splash pages, and "Best viewed in Internet Exploder" logos.

[Kubla Cant]

MS spend a lot of time on their UIs and compared to the amount of time most Web designers spend it shows.

If anybody spent a lot of time designing the Windows 10 UI, it must have been a team of psychopaths.

My least favourite bit of Microsoft UI is their penchant for providing a fixed-size dialog for editing long text values, such as the environment tool in Control Panel.

[JohnFen]

"MS spend a lot of time on their UIs and compared to the amount of time most Web designers spend it shows."

It sure does -- considering how bad Microsoft UIs are, it shows that they're really bad at user interfaces.

[404]

hehehehe

Active desktop... Windows 98... just stahp...

[Anonymous Coward]

Re: hehehehe

I very much liked Active Desktop on Windows 9x. I used my own local webserver to serve small apps that could trigger actions via CGI from buttons embedded in my Windows desktop.

It felt like living the future!

[disgruntled yank]

Thanks.

I was in danger of getting some work done, and you informed me about "Progressive Beer".