Skip to main content

Flogging a Dead Horse

Posted by javakiddy on June 30, 2008 at 3:57 PM PDT

Today is apparently Bill Gates' first day away from Microsoft. As he leaves, some have suggested Microsoft's star is now in the descent, as Google's star climbs ever higher. Is this really the case, is Google destined to become the next Microsoft? When a company attains a certain dominance in the market, isn't it hard to unseat them? After all, they can afford to hire all the best people!

Cast your mind back to IBM's nervous toe-dipping when it came to the fledgling micro computer market in the Seventies, or Microsoft's initial head-in-the-sand attitude towards the internet in the Nineties — being big doesn't always make you right. Indeed the larger the organisation, the better it gets at sustaining incorrect assumptions in the face of mounting contradictory evidence. (One wonders, for example, whether a concept like transubstantiation could ever have survived in a religion with only a handful of members?)

There's safety in numbers, for sure, but only by way of passing the buck for a bad idea. Shared responsibility can often mean no responsibility at all. In the right environment bad memes can survive unchallenged, and humans seem particularly good at creating those environments. We believe because the people around us believe, not because we have given an idea careful contemplation or scrutiny. What's important is that the group has clear goals; how well those goals stand up to reality is of secondary concern. As the song says: "any dream will do!"Google prides itself on hiring only the select few, the brightest of the bright. But a youthful company with an allegedly youthful workforce, all recruited from similar stock, doesn't leave much room for diversity of outlook. Those familiar with tale of Apple Computer's first decade will know how a group of supposedly intelligent people can divorce themselves from reality when they're distracted by working on new and exciting technologies.

With so many developers recruited straight out of university, one wonders how many Google engineers really remember programming before the arrival of web? Perhaps unsurprisingly Google is wedded to the web as a platform for Rich Internet Application development, yet is there any strong evidence that this is a fruitful avenue to pursue?

Sure, Gmail is used by many, but what of Google's other web-app offerings? Docs? Spreadsheet? Does anyone, aside from the occasional curious soul, seriously use these applications? I don't think so! Yet Google continues to develop and promote the likes of GWT (Google Web Toolkit) and Gears, technologies designed to smooth over or work around the obvious shortcomings in the web platform.

If I was being cruel I might wonder whether the true genius of Google lies in finding more inventive ways to flog a dead horse! When you consider the promise of Adobe's AIR, Microsoft's Silverlight, or indeed Java's own JFX, you wonder why we aren't seeing evidence of significant investment in these technologies — or at the very least in one of them. Silverlight and JavaFX may still be raw, but AIR is mature enough to start producing applications, if only exploratory beta releases.

Perhaps behind the scenes there are indeed moves to examine these alternative RIA technologies? Maybe, as I write, the latest build of a JFX based Google Docs has just finished compiling, or an Adobe AIR Gmail has been handed off for internal testing...? But there's very little evidence of that from the itinerary of the recent Google I/O conference.

The problem with these non-web RIA platforms is they rely on a foundation of software already being installed on the user's desktop — in the case of JavaFX, for example, it's the JRE. Having to stop and install a plugin creates a hiccup in the user experience; by contrast the web browser is guaranteed to be present. So the future of non-web RIAs depends upon breaking an old chicken/egg scenario: RIAs won't start being written until runtimes are ubiquitous, and runtimes won't be ubiquitous until there's enough RIAs to drive demand for them.

Microsoft has the option of simply pushing Silverlight out like it did with IE 7, such are the perks of owning the OS which runs on 90% or so of the World's computers. While the roll out wouldn't reach every Windows user, it would cover enough to make Silverlight a lot more attractive. Can Google afford to wait for Microsoft to do this?

The only vendor of RIAs which has the clout (in terms of brand recognition and trust) with the common user is Google. If Google announced an enhanced Adobe AIR version of Gmail you can bet your bottom dollar it would generate plenty of headlines. Yet Google seems content to keep prodding the web, in the hopes it will somehow transform itself, in the best Cinderella fashion, into an effective RIA vehicle.

There's plenty of paranoia about Google around at the moment. People are whispering "the next Microsoft" every time Mountain View announces a new product. Yet I wonder if, by shackling themselves only to the web and not dabbling with these other technologies, Google hasn't committed itself to something it will later regret?

Of course Google could be right — maybe the future of the RIA is inside the browser. And even if they're wrong, there's still plenty of time to do a U-turn (although by waiting they could forfeit the opportunity to champion their preferred platform.)

For the sake of us all, I sincerely hope they are wrong, and they'll start dabbling with JavaFX or AIR sooner rather than later!

Related Topics >>

Comments

some writers have said that Google (rightly) believes in open standards ie. that the internet and their apps should be built on open standards, and not beholden to particular vendor, eg. Adobe, Sun, Microsoft, or whoever. DHTML/javascript/ajax foots that bill. If Flex or whatever is opensource, and it's basic tools are opensource, and it becomes a dinkum open internet standard, then that surely that makes it a more compelling option that DHTML for richer, more responsive, more beautiful web apps...

Do you have any usage data on the Google apps that you assert are being ignored (docs, spreadsheet...)? I know that I use them and think that they are great. I am a huge fan of GWT... Why would you say that Java Swing has never gained much popularity for commercial use? I think that it is because it is annoying to have to download a runtime environment. GWT apps are great because you can start use them in any browser without any additional inconvenience. I think that AJAX frameworks like this are going to reign king.

@radd9er :Do you have any usage data on the Google apps that you assert are being ignored (docs, spreadsheet...)?

No, admittedly I don't. Anecdotally though they don't seem to be making much of an impact -- most of the noise surrounding them seems to be about the fact that Google managed to create such applications within the browser. I hear nothing about people actually using them in a practical sense. I wonder what the usage statistics actually are, eg what market share Docs has versus Word or OO ?!?

@tompalmer: It might not seem like the emperor has any clothes to you, but until you can explain why anything exists in the first place, it might be worth giving people the benefit of the doubt. Just my own two cents (if it's even that much).

But by that token we'd have to give tentative default credence to the Flying Spaghetti Monster too, and those Spaghetti worshipers are a bunch of crazy argumentative @$*#holes! :)

Also, I don't personally believe in transubstantiation, but I think it might be more polite to be careful on such subjects. It might not seem like the emperor has any clothes to you, but until you can explain why anything exists in the first place, it might be worth giving people the benefit of the doubt. Just my own two cents (if it's even that much).

Google does use Flash for some things, at least the street view in maps and the charts in finance. And then there's that video thing. But I think they are right not to put everything on that boat when Adobe runs it alone. Depending on Silverlight is equally risky (and not as mature). Java is happily open source (mostly), but it's not really a solid cross platform solution, even after a decade. Nothing but DHTML really gets the job done yet. That's why many people focus on making DHTML better rather than trying to make another solution work. I'd _love_ to see Java get there (and FX, while nice, isn't really what's _needed_). I really would.

@pbw: The religion started with only a handful of members, so I guess your thesis (whatever it is) is in trouble.

Unless Transubstantiation wasn't recognised in the form it is today when the religion first started, but grew from the fringes to the mainstream as part of the evolution of the religion. It's the kind of social phenomena Daniel C Dennett (the famous American philosopher) has written about, and his theories seem fairly reasonable, to me at least.

"Indeed the larger the organisation, the better it gets at sustaining incorrect assumptions in the face of mounting contradictory evidence. (One wonders, for example, whether a concept like transubstantiation could ever have survived in a religion with only a handful of members?)" The religion started with only a handful of members, so I guess your thesis (whatever it is) is in trouble.

The horse that Google picked was Java. The fact that Java wasn't open source and that applet performance and deployment had largely been ignored was the reason they had to compile to AJAX/Javascript. I'm sure Google would have loved to have a better choice of deployment runtimes, but they didn't have one at the time (and they didn't have a time machine to get one from the future -- which they would need even now since 1.6u10 is still in beta). I think this community is off base if they are painting them as the bad guy for choosing the most open and available way of using Java at the time. They could have chosen Flex, Silverlight, or some javascript library thing, but they chose something that helps java adoption/tools.

Keep in mind the big picture, which is all about money. Google makes huge amounts of money from adwords and adsense, and makes zero money from anything else they do. That includes Maps, GWT, gmail, Android, and everything else besides web search. So the big picture is: google makes money from web search, and is dabbling in lots of little other software stuff. Microsoft is the same way. Their Windows and Office divisions have always made huge amounts of money, and none of their other divisions have ever made any money (in fact, they've lost a lot of money). The XBOX division may be in the black now, but will take many years to make up all their loses. So again, the big picture is that MS makes WIndows and Office, and dabbles in a lot of other areas. So why aren't we seeing any investment by google in anything besides search? Why would they? It may help their branding a little to put out cool stuff, but it's really just a money sink.

"The problem with these non-web RIA platforms is they rely on a foundation of software already being installed on the user's desktop" Uh.. Sun has been fairly successful at getting OEM's to bundle Java runtimes in windows installations. This means a large portion of existing windows desktops have a modern JRE. What if Sun were to distribute JavaFX as part of a 6uX update? What if Sun were to get these same OEM's to also distribute JavaFX? The problem with relying on web-browser RIA's is it entraps the internet within the web. So long as RIA's are dependent on the browser then the internet will be only as capable as web browsers. But the internet has far more capabilities than HTTP.

"Yet Google continues to develop and promote the likes of GWT (Google Web Toolkit) and Gears, technologies designed to smooth over or work around the obvious shortcomings in the web platform." An interesting exception is Anroid. It's much more like J2ME/JME-like (*wink* *wink*) and not at all like a web application.

"The problem with these non-web RIA platforms is they rely on a foundation of software already being installed on the user's desktop" So is true of the web browser. Do you not remember when operating systems didn't have web browsers in them and you had to go get one somehow? You are right that part of what makes web applications attractive today is the fact that a renderer for such applications is broadly deployed. But that wasn't always the case. Something about web was attractive enough that people sought it out, eventually giving it enough traction to become as proliferated as it is today. That same magic could happen again if you can get a critical mass of interest in "the next best thing." But people won't start gravitating toeards something new if its only moderately better than what they have today. There needs to be a compelling reason -- for developers, users, companies alike -- to adopt it.

"When you consider the promise of Adobe's AIR, Microsoft's Silverlight, or indeed Java's own JFX, you wonder why we aren't seeing evidence of significant investment in these technologies — or at the very least in one of them." I don't think its something to wonder about. They're all limited technologies in that they are closed, or locked to their own platform (e.g. my .NET loving neighbor will never partake in Java FX while I will never partake in Silverlight). What we need is an open standard that all languages can leverage. One I've seen that has promise is OpenLaszlo. It is browser-like in that it is a markup language with event handlers. I *wonder* if there are other such standards that could foot the bill.