Back to the Desktop

Taking the web out of the browser has been a common theme lately. As web apps become more a part of everyday productivity, accessing them outside the typical browser is more attractive.

The iPhone has a great example of this in its Maps app. The beauty of taking apps out of the browser is you can focus on the app and its functionality outside the constraints of the browser.

Not all web apps are suited to standalone status. My main criteria for deciding if I should try a standalone version of a web app are simple: do I keep the app open all the time in a tab and/or do I have any Firefox add-ons that help me manage the app?

Both of my reasons relate to my preference for Firefox as my browser of choice. Firefox 2 was the first version I can remember leaking memory in Windows. The memory consumption gets worse if you have a lot of add-ons or keep a lot of tabs open. I routinely get 20-30 tabs open without thinking, so this is an issue for me.

The beta versions of Firefox 3 seem to be slightly better, but this may also be due to the fact that very few add-ons have been forward-ported yet.

So, I’m finding ways to cut back on Firefox’s memory usage. Maybe I need a new browser. 

Anyway, the resulting list of web apps is pretty short.

  1. Twitter: I had several add-ons to Firefox (Twitterbar, TwitterFox), plus twitter.com always open in a tab.
  2. Google Reader: I keep Reader open all the time and use it probably more than any other web app.
  3. GMail: I use the GMail Notifier add-on, and I check GMail from my iPhone all the time.

Now that I’ve established demand for these as standalone apps, I had to find the best solution, and the main choices I considered were Adobe AIR clients and Mozilla Prism.

There are a bunch of AIR clients for Twitter out there, but Twhirl is my choice. I’m not sure why, maybe inertia. I tried a couple others and went back to Twhirl. I’m not alone either. It’s estimated that Twhirl accounts for 7% of tweeting, which is tops among AIR clients. They’ve released some cool new features lately, like posting tweets to Pownce and Jaiku. I’m eager tosee what comes out of the acquisition by Seesmic.

I like the standalone view of Twitter for a couple reasons:

  1. It runs ambiently in the background, collecting the stream of tweets. I can pay close attention, or none at all. Either way, it keeps a log for me.
  2. Twhirl and most other Twitter clients make it easy to use the feaures of Twitter like reply, direct message, retweet and favorite that are more hidden in the web version.

Because of Twitter’s API, AIR clients are all over the place. With Reader and GMail, this isn’t the case. So, I use Mozilla Prism for these.

Prism is essentially a stripped down browser window that doesn’t render menus or support add-ons. It just loads whatever URL you tell it to load. No rocket science involved.

The benefit is that Prism uses a lot less memory than Firefox does, and each Prism “app” has its own allocation of memory. So, if my Prism instance of Reader eats itself, my GMail and Google Docs instances will be safe. Not so in Firefox or any browser, as I’m sure you’ve experienced.

Prism has the same ambient benefits that Twirl provides, i.e. I can ignore my Prism apps and work happily, undistracted in the browser window without knowing how many unread feed items I have or how many new messages I have in GMail.

Fewer distractions is a huge benefit of using Prism, as Lifehacker documented. I’m marinating a post on distractions right now too.

One bummer is that Prism doesn’t handle icons very well, which is a bit of a drag. It captures the favicon from the URL, then blows up to icon size, all pixelated. I dug up some icon packs from Apple for my Mac. Now, my desktop looks like alphabet soup. It’s kind of fun to see what words you can accidentally spell, like a Scrabble game.

Another major bummer is that I can’t get Prism working on my work XP laptop, where I need to use a proxy server. Prism doesn’t have any configuration options, but since it’s stripped down Firefox, it has an about:config. I copied all the values from Firefox into Prism, but no dice. This is a downer because I have less memory on the XP laptop.

As web apps move out of the browser, I expect to see a lot more cool stuff. Bex riffs on the possibilities of Google Gears, Greasemonkey and Web 3.0 in a post he sent to me in email (gasp) on Friday. Coincidentally, I was playing around with Prism at the time. Weird.

I like Google Gears, but it’s gone nowhere since it was introduced last year. Sure, Reader uses it, and Zoho did something with it, and Docs now supports some offline functionality. Still, I hear it’s limited and more aimed at supporting the lose of web connectivity than offline productivity. I feel pretty underwhelmed so far.

Plus, it’s only for Firefox 2.

What do you think about taking web apps out of the browser? Sound off in comments.

AboutJake

a.k.a.:jkuramot

18 comments

  1. I like what I’ve seen of Air , even done a little development with Air and APEX. Being able to build applications that run on the desktop using just HTML and JS is awesome. Prism is really nice as well, for the reasons you mentioned.

    The main reason I like browser based software though is the browser is arguably the most ubiquitous piece of software around, whether it be FF, IE, Opera, Safari, mobile or desktop, and on whatever OS.

    If you sit down at a computer , even if it’s not yours or a public computer and it has a webbrowser and it has internet connection you can pretty much use any web based application around.

    If you sit down at a computer and need to run an Air , Silverlight , even Flash, based application whoever owns that computer might not be as happy that you installed it or depending on the your privileges you can’t / shouldn’t be able to install it.

    The main issue with browser based software being offline access but that seems to be, and will continue to become less and less of a problem.

    Sure I’m slightly biased but that’s one of my favorite things about APEX and it’s web based development. I can’t count the times that I’ve just jumped on a random computer and demoed , fixed or developed applications.

  2. I like what I’ve seen of Air , even done a little development with Air and APEX. Being able to build applications that run on the desktop using just HTML and JS is awesome. Prism is really nice as well, for the reasons you mentioned.

    The main reason I like browser based software though is the browser is arguably the most ubiquitous piece of software around, whether it be FF, IE, Opera, Safari, mobile or desktop, and on whatever OS.

    If you sit down at a computer , even if it’s not yours or a public computer and it has a webbrowser and it has internet connection you can pretty much use any web based application around.

    If you sit down at a computer and need to run an Air , Silverlight , even Flash, based application whoever owns that computer might not be as happy that you installed it or depending on the your privileges you can’t / shouldn’t be able to install it.

    The main issue with browser based software being offline access but that seems to be, and will continue to become less and less of a problem.

    Sure I’m slightly biased but that’s one of my favorite things about APEX and it’s web based development. I can’t count the times that I’ve just jumped on a random computer and demoed , fixed or developed applications.

  3. and what precisely is (gasp) wrong with email??? 😉

    trust me… if I tweeted my every thought, nobody would be happy.

  4. and what precisely is (gasp) wrong with email??? 😉

    trust me… if I tweeted my every thought, nobody would be happy.

  5. @Carl: All very valid, but I’m not arguing against the browser in general, just for certain apps in AIR/Prism.

    It’s moot if we’re talking about a computer that isn’t my own.

    @bex: In this context, it’s funny: you emailed me to point me to a post in your blog (to which I subscribe, BTW) about web apps and Web 3.0/4.5.

    I guess email is still the LCD for communicating w/people. So, touche.

  6. @Carl: All very valid, but I’m not arguing against the browser in general, just for certain apps in AIR/Prism.

    It’s moot if we’re talking about a computer that isn’t my own.

    @bex: In this context, it’s funny: you emailed me to point me to a post in your blog (to which I subscribe, BTW) about web apps and Web 3.0/4.5.

    I guess email is still the LCD for communicating w/people. So, touche.

  7. Hi.

    This is going to ramble and probably be flame-bait so I appologise in advance. 🙂

    First, we felt liberated from client-server by the net, but were lacking functionality.

    Next, we tried using Java applets or OCXs to get that client-server feel, but people objected to having to install a specific JRE version and download whole applets, or being limited to a single vendors browser.

    Now we use DHTML and AJAX to try and recreate client-server experience, but this has many browser compatibility issues and involves downloading large Javascript libraries, so it’s not disimilar to the problems with applets.

    The immediate future looks like a rise in client-side apps built on proprietry frameworks, accessing data from web services.

    Let’s take the case of email. If I have a client side app accessing my gmail account as a web service, this is no different to having a PC IMAP client accessing my server based IMAP email account. All the data is still centralized on my server, not downloaded to my PC like POP, although I can download it if I know I need it offline. So what have we invented here?

    What the net has always lacked is a standard framework for interactive applications. We shouldn’t be relying on one million and one DHTML or AJAX toolkits. This functionality should either be part of the browser, or part of a standard “web-app” component. That way, people can concentrate on writing useful apps, not worrying about implementation issues caused by the limitations of the platform.

    I have a lot of respect for all the people doing amazing things with DHTML and AJAX, but it really is trying to fix the problem with a band aid. At some point in the future we will look back at this and laugh at how ridiculous it all was.

    I know it’s been said before and it’s all a bit idealistic, but the browser in its current form is not the tool for complex interactive applications. 🙂

    Cheers

    Tim…

  8. Hi.

    This is going to ramble and probably be flame-bait so I appologise in advance. 🙂

    First, we felt liberated from client-server by the net, but were lacking functionality.

    Next, we tried using Java applets or OCXs to get that client-server feel, but people objected to having to install a specific JRE version and download whole applets, or being limited to a single vendors browser.

    Now we use DHTML and AJAX to try and recreate client-server experience, but this has many browser compatibility issues and involves downloading large Javascript libraries, so it’s not disimilar to the problems with applets.

    The immediate future looks like a rise in client-side apps built on proprietry frameworks, accessing data from web services.

    Let’s take the case of email. If I have a client side app accessing my gmail account as a web service, this is no different to having a PC IMAP client accessing my server based IMAP email account. All the data is still centralized on my server, not downloaded to my PC like POP, although I can download it if I know I need it offline. So what have we invented here?

    What the net has always lacked is a standard framework for interactive applications. We shouldn’t be relying on one million and one DHTML or AJAX toolkits. This functionality should either be part of the browser, or part of a standard “web-app” component. That way, people can concentrate on writing useful apps, not worrying about implementation issues caused by the limitations of the platform.

    I have a lot of respect for all the people doing amazing things with DHTML and AJAX, but it really is trying to fix the problem with a band aid. At some point in the future we will look back at this and laugh at how ridiculous it all was.

    I know it’s been said before and it’s all a bit idealistic, but the browser in its current form is not the tool for complex interactive applications. 🙂

    Cheers

    Tim…

  9. @Tim: That was a good read, and I wholeheartedly agree. That is all. I really have nothing to add.

  10. @Tim: That was a good read, and I wholeheartedly agree. That is all. I really have nothing to add.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.