Slashdot Mirror


Progressive Web Apps Moving Mainstream As Twitter Makes Its Mobile Site the Main (arstechnica.com)

An anonymous reader quotes a report from Ars Technica: Twitter is showing some users of its desktop website a new user interface that is designed to be faster and to feature support for the recently added bookmarks feature (supported in the iOS and Android clients but not, currently, the main website), a data-saver mode, and a night mode. These users have been selected at random and moved over to the new interface so they can test the interface and provide feedback. The new interface isn't all that different from the old one: it's organized a little differently, with a two-column layout instead of the three columns currently used, but overall it will feel familiar to anyone who has used the microblogging platform before. What makes this move interesting isn't the specifics of the interface itself, but the technology it's built on.

The new interface isn't actually new at all. It has been available for some time now as mobile.twitter.com, Twitter's mobile-friendly Web interface. In turn, that same Web interface is used to drive the Windows 10 app, the KaiOS platform for "smart feature phones," and the recently released Twitter Lite app for Android. This is why it has the data-saver mode; it has been designed with an eye on those users who suffer from poor or expensive bandwidth or have underpowered devices. This mobile site is perhaps one of the most prominent instances of what could be a new breed of Web application: the Progressive Web Application (PWA). PWAs are Web applications that build on certain modern browser features to provide an experience that is much more like that of a traditional application. For example, PWAs can support offline operation using service workers (a way of running JavaScript in the background that can respond to events and make network requests that degrade gracefully if the network is unavailable); they integrate with platform features such as notifications; they're also designed so that they could be pinned to app launchers and home screens and treated as if they were "real" applications rather than merely webpages.

24 of 76 comments (clear)

  1. Tick tock... by fragMasterFlash · · Score: 1

    How long until site owners start howling about their ads not rendering before the content the user actually cares about?

  2. Re:Just don't post anything on twitter by Zontar+The+Mindless · · Score: 2
    --
    Il n'y a pas de Planet B.
  3. pwa good but basic things like DNS bad... by johnjones · · Score: 4, Interesting

    twitter actually has some nice compiler guys and does some good work on web standards and things like PWA... they are let down by their networking team it seems...

    here are some basic DNS failures at twitter :

    No DNSSEC (allows nation states to spoof twitter on HTTPS connections )
    Name Servers are on the Same Subnet
    Serial numbers do not match across servers
    SOA Serial Number Format is Invalid on some servers
    outside provider (oracle) has failed many performance targets

    regards

    John Jones

  4. Isn't AJAX Similar? by buravirgil · · Score: 1

    How is this not simply Twitter asserting a new initialism to avoid potential patents and licensure? And simultaneously seeking growth in less developed economies? Platforms based on social connectivity "harvest" data by XML that is useful to what enterprises? What agencies of security? Metadata, its frames and categories, is a developing comprehension and its assignments blur the lines between enterprise and politics.

    A growing number of people advocate to step away from being the "product" of these enterprises-- so long as a western corporations seek growth in developing economies, how can the information they sell to business interests outside that developing economy not be seen as advantage seeking and extractive?

    --
    Would were! Should is! Could be! And live a hundred times three.
  5. Re:Just don't post anything on twitter by Desler · · Score: 2

    Why is InfoWars?

    If you violate these rules, your posts and/or user name will be deleted.
    Remember: you are a guest here. It is not censorship if you violate the rules and your post is deleted. All civilizations have rules and if you violate them you can expect to be ostracized from the tribe.

    https://www.infowars.com/terms...

    Hypocrisy, thy name is Alex Jones.

  6. "Progressive Web Apps" by Gravis+Zero · · Score: 1

    AKA insecure programs executing on your computer. We seriously need to ditch this kind of shit.

    --
    Anons need not reply. Questions end with a question mark.
    1. Re:"Progressive Web Apps" by tepples · · Score: 1

      Would you instead prefer a full page reload every time you click to expand or collapse anything that isn't simple enough to be a checkbox hack? Would you instead prefer a full page reload if you're entering values into a form and a value two screens up happens to be out of range?

    2. Re:"Progressive Web Apps" by Gravis+Zero · · Score: 1

      No, I would prefer CSS to be able to load a page of subelements from the server but only when clicked. You don't need a Turing complete programming language to solve dynamic loading.

      --
      Anons need not reply. Questions end with a question mark.
    3. Re:"Progressive Web Apps" by tepples · · Score: 1

      I would be interested in reading your proposed specification to allow more declarative interactivity in HTML+CSS documents.

  7. Maintaining one "traditional app" isn't enough by tepples · · Score: 1

    It is *far* *easier* to maintain a traditional app in this respect rather than do it via a PWA and some PITA broken javascript interface.

    The problem is that maintaining one "traditional app" isn't enough. You must maintain five, one each for Windows, macOS, X11/Linux, iOS, and Android. Qt doesn't save you from having to buy code signing certificates (which are more expensive than TLS certificates) and build, test, and distribute for all five platforms.

    1. Re:Maintaining one "traditional app" isn't enough by tepples · · Score: 1

      First, you can't buy Photoshop anymore for any platform.

      Second, by my standard, Adobe has failed where developers of web based photo editors have succeeded.

    2. Re:Maintaining one "traditional app" isn't enough by tepples · · Score: 1

      Then would you prefer that each user buy five computers, one to run applications exclusive to each operating system?

    3. Re:Maintaining one "traditional app" isn't enough by tepples · · Score: 1

      Even with "smart application design," a native application still requires a substantially bigger budget ($$$) for porting than a web application.

  8. Five different products, one for each OS by tepples · · Score: 1

    They should just transform their "web platform" into a set of libraries, like e.g. KDE, but less frameworky.

    In your vision, would end users compile an application from source code? Or would developers be responsible for shipping five different products, one built for Windows, the second for macOS, the third for X11/Linux, the fourth for iOS, and the fifth for Android? At least a program written in JavaScript or WebAssembly can be executed without modification on multiple platforms. "Inner-platform effect" on Wikipedia states that "portability and privilege separation reasons" are valid goals, and in some cases they may outweigh the drawbacks of an inner platform.

    1. Re:Five different products, one for each OS by lgw · · Score: 2

      Web developers developed for "five different platforms" for most of the web's history, you know. Cross-platform testing is always a bigger cost than development, and you have to do that regardless. One way or another, you end up using libraries or a framework that abstracts away platform differences.

      It would be nice if Xamarin would extend its tools to Mac and Ubuntu, though. They do a good job with Windows/Android/iThing.

      --
      Socialism: a lie told by totalitarians and believed by fools.
    2. Re:Five different products, one for each OS by tepples · · Score: 1

      You appear to claim that building, testing, and distributing for Google Chrome, Safari, Edge, and Firefox is no less costly than building, testing, and distributing a native application in Windows, macOS, X11/Linux, iOS, and Android editions.

      As for testing: The subsets of the HTML5 platform accepted by Blink, Apple WebKit, EdgeHTML, and Gecko are more similar than (say) Cocoa, Win32, Xlib, Cocoa Touch, and Android API. Qt can cover some but not all of these differences. (I haven't used Qt myself, as my employment is in a field of programming other than native GUI applications, so I'm going by what others have told me.)

      Early on, you could target Blink and Gecko and leave EdgeHTML and Apple WebKit in second-class support until you accumulate enough capital to expand your testing to browsers whose pack-in browser is non-free (namely Windows, macOS, and iOS). Both a Blink-based browser and a Gecko-based browser are available for all major platforms other than iOS, and Microsoft considers any behavior differences between EdgeHTML and Blink to be "bugs that we're interested in fixing." (Source: "Building a more interoperable Web with Microsoft Edge" by Microsoft Edge Team) But unlike for web applications, operating system publishers have not committed to any sort of interoperability with respect to native GUI applications.

      Once you have developed and tested your application, the next step is distribution. You'll need a domain and web hosting whether you are publishing a web application or a native application. In the case of a web application, users access the application through this domain. In the case of a native application, users download the application's installer or are redirected to the correct platform's mobile app store listing through this domain. And either way, you'll need some sort of dynamic capability, whether to run a web application or to process payment from users buying a license to a native application. This rules out just putting a static site on Amazon S3.

      But if you are developing a native application, a domain-validated TLS certificate obtained from Let's Encrypt will not be enough. You have to buy an Apple developer ID to publish on iOS at all and to keep Gatekeeper from recommending that macOS users delete your application without running it. And you have to buy an Authenticode certificate to keep SmartScreen from recommending that Windows users delete your application without running it. And you have to keep it renewed as long as you continue to maintain your application. And it's still five different SKUs you have to ship rather than one.

    3. Re:Five different products, one for each OS by lgw · · Score: 1

      You appear to claim that building, testing, and distributing for Google Chrome, Safari, Edge, and Firefox is no less costly than building, testing, and distributing a native application in Windows, macOS, X11/Linux, iOS, and Android editions.

      I was talking about the pre-Chome days of poor standards adherence.

      Sure, if you're a one-man crew, those licensing costs suck. But if you're an actual software shop, the cut the stores take dominates, and that's the argument to make these days. It's only getting easier to develop native applications cross-platform, and cross platform dev costs keep shinking (it's very close to "just use Xamarin" for non-games). Cross-platform testing costs will always remain, though, regardless of approach.

      Obviously, for games the cheap approach is Unity if you're small, and Lumberyard if you're big, and the tech problems for going cross platform are small compared to the problems of marketing and user expectation being do different between platforms.

      --
      Socialism: a lie told by totalitarians and believed by fools.
    4. Re:Five different products, one for each OS by tepples · · Score: 1

      Sure, if you're a one-man crew, those licensing costs suck. But if you're an actual software shop, the cut the stores take dominates

      Say a 2-man startup develops its minimum viable product as a web-based service to avoid annual and percentage fees of the OS publishers' stores. Should it decide to expand to "an actual software shop", it'll probably stay web-based.

  9. PWAs, How are these new? by shayd2 · · Score: 1

    How is this different from thin clients?

    Not necessarily the methods, but the overall goals (+ ads)

  10. Re: Just don't post anything on twitter by Desler · · Score: 1

    And yet it's still within their right to do so even when it's capricious and arbitrary despite all the snowflakes going apoplectic. Alex Jones and his band of snowflakes are not being censored and they are free to move to Gab where they'll fit right in.

  11. That's why I'm staying with the Web. by Qbertino · · Score: 1

    I've been doing web development for nearly 20 years now and have finally decided to actually stay in the field despite the douchebag quota in the industry being through the effing roof. Stuff like PWAs and browser vendors finally getting their shit together and bring mostly standards compliant keep it interesting for me. Plus an abundance of new and neat technologies to keep things interesting. I'll just be looking for better teams in the future. You develop a thick skin and a acute sense of smell for shitty gigs and crappy web-shops.

    --
    We suffer more in our imagination than in reality. - Seneca
    1. Re:That's why I'm staying with the Web. by Billly+Gates · · Score: 1

      I've been doing web development for nearly 20 years now and have finally decided to actually stay in the field despite the douchebag quota in the industry being through the effing roof. Stuff like PWAs and browser vendors finally getting their shit together and bring mostly standards compliant keep it interesting for me. Plus an abundance of new and neat technologies to keep things interesting. I'll just be looking for better teams in the future. You develop a thick skin and a acute sense of smell for shitty gigs and crappy web-shops.

      Hey QBertino. My old XP box with IE 6 looks funny on your website. Can you fix?

      If it can be fixed before Monday great because Grandma just may happen to stumble across your website and we can't have that.

      Thanks as it should be a simple fix.

  12. Re: Just don't post anything on twitter by Highdude702 · · Score: 1

    So basically "we don't like you so you're banned!!11!!!!1!twelve"

  13. Re: Just don't post anything on twitter by eaglesrule · · Score: 1

    Yeah, Infowars maybe has tens of thousands of users, and Twitter has hundreds of millions. They're practically the same thing.

    Because tech titans colluding to silence personas non grata as they will it is nothing to be concerned over. We promise you'll like it. It's for your own good.