Keep an Open Eye

Ray Kurzweil has written a series of books ( The Singularity Is Near: When Humans Transcend Biology, The Age of Spiritual Machines: When Computers Exceed Human Intelligence, and The Intelligent Universe: AI, ET, and the Emerging Mind of the Cosmos) in which the underlying theme is that intelligence will blossom unimaginably aided and abated by an equivalent blossoming of machine intelligence. And the arguments are compelling - the sheer momentum of Moore’s Law has seen an unbroken doubling of computing power (CPU speed, memory capacity, disk speed and capacity, communication bandwidth, etc) every 12-16 months for the past 40 years. And even natural physical barriers such as the speed of light, the limits to miniturization, the natural vulnerability/randomness of atomic level processes appear to fall by the wayside to new solution sets. So the progression of computing power is now at gigaflops - billions of computer operations persecond and will be at 1000 times that in 10-15 years. Ray has started to do the imagineering of what that type of computing power could bring about.

But of course the Devil is in the details.

Hardware computing power has always lead software development. And software development has not nearly kept the same relentless productivity pace of hardware development. Even software that generates software - is at best primitive, clumsy, and woefully incomplete. Read the works on Model Driven development.

The fundamental retardent to software emulating hardware’s Moore’s Law is do no evil.

Do no evil is the famously attributed motto of Google’s Larry Page and Sergei Brin. But it applies to software development in the following sense - one must suspend evil intent in the development of most complex software. Because that software, in order to be effective is granted rights and privileges that could easily disable the whole system. Think of the eval() function in JavaScript or the old fashion ability of dBase macro commands to create new commands on the fly or the privileged status OS kernel processes.

For the past 15 months I have been arguing that touchscreen computing with multi-touch capability will be along with RAIA (the ability to deliver GUI apps that run on any browser and/or locally on any platform) the two technologies that dominate the personal PC and mobile device direction. Well its nice to see that Scientific American agrees with my first argument.

In their June 2008 issue, Scientific American again peers ahead and decides that multitouch screen technology will be the next new wave on the GUI interface scene. Now the specific technologies that Scientific American looks at are multi-thousand dollar implementations. But we have already seen Apple deliver touch screen to the iPhone (its distinguishing feature) and on the touchpad of its new MacBooks. And HP delivered a very competitive direct touchscreen PC but not yet multitoch enabled. But you can get around that with Java routines or Flash ActionScript libraries that deliver multi-touch in software.

In short, multi-touchscreen may be closer to delivery than might first be expected. Mark Scientific Americans word - touchscreen will become a dominant product characteristic for GUI deliverance in the next few years. Now all GUI vendors have to do is figure out how to deliver RAIA versions of the same. Indeed this will be a test of whether the proprietary and barely browser cross platform Microsoft solution will be rejected for the cross browser, cross OS platform approach taken by Adobe Flash and Java JavaFx.

Will Microsoft be able to sell proprietary once again? Only the Shadow knows.

Joe Wilcox at Microsoft-Watch has been doing more truth-saying about Microsoft’s various latest deliveries in the PC marketplace than any other IT Trade Press pundit. He has been correct to not be kind to Redmond on Vista, aspects of Live Mesh, and the process of doing the Yahoo-Buy. But on Windows XP, Joe unexpectedly loses his truth-speak - see here.

I am reminded of the assessment of Fake Steve Jobs and what he had to say about former PC-Mag editor Jim Louderback upon his quiting PC Mag to take another post and finally saying that Vista sucked:
“Really? You think? Translation: Now that I’m no longer working at PC Mag, I can finally tell you the truth. Vista blows. Vista is a boat anchor. Vista paid my salary for the past few years, but now that I’m no longer sucking on the Redmond teat I’m going to play hero and tell you what I didn’t dare tell you before”.Unfortunately I am getting the same feel for Joe Wilcox and his comments on not extending Windows XP sales beyond June 30th while Windows Vista, XP’s replacement, continues to be unacceptable to so many users. Instead of demanding like others that Redmond should make Windows Vista as compatible as XP and as resource effieient - Joe caves into the Redmond mantra that we have to move on no matter how bad Vista is relative to Windows XP.

Joe appears to agree with Redmond - far be it that Microsoft should spend some of those 85-90cent profit per dollar of revenue on allowing Windows XP to exist until Microsoft gets a non-bloated, non-software and hardware incompatible, less security obnoxious, more performant version of Windows Vista out the door. And the cost is neglible if not revenue positive because as Microsoft works to get Vista right Windows XP brings in even more revenue because 1)it has an even higher profit margins than Vista and 2) it helps stem defections to a)Apple, b)Linux and Asus, c)Web 2.0, and d)mobile devices like iPhone, Blackberry, etc.

Joe appears to agree with Redmond that Microsoft should be allowed to shaft end-users who have not been able to buy a PC loaded with Windows XP from any major PC vendor for months while small to large businesses have been able to stall all Vista well into next two years if they so choose. Come on Joe, show some support for the little guy.

Joe appears to agree with Redmond that Infoworld got it wrong and there is no performance advantage of Windows XP over Vista and so why allow XP to exist - just get on with the forced upgrade to Vista.

And finally, Joe appears to be excusing Redmond from finding a way to make XP available after June 30, 2008 on Asus Eee and other Minibook laptops. Hey Joe, can you get the story right on the rumor that this exonerated Minibook version of Windows XP is crippled such that it can only run 2-4 independent applications ??

Really, Joe do you think a)the option of downgrades from Vista to XP is good (Steve Ballmer saying you still have a choice and there is no work involved at all) and b)"Until the Windows ecosystem moves[to devoted Vista only support], customers won’t get the best Vista experience” will make a whit of difference? Vista support is as good as Windows support- like the German use of the Umlaut “sometimes, sometimes, always, never.” Really, Joe, do you think this is all Microsoft owes to the Windows brand and long customer loyalty? Come on Joe - you and the IT trade press in general should be demanding a lot more of Microsoft in their “service” to the Windows community. It sounds like we should have somebody watching the so-called Microsoft-Watcher.

Joe Wilcox, truth-speaker on Microsoft, appears to have lost his voice.
PS: Joe, comments are open any time, any day, any minute.

Microsoft is certainly delivering a full court press on Silverlight (its RIA cross browser tool)- and really pushing its proprietary features that will work only on Windows XP and Vista but also “best” only with .NET 3.5 (2.0 .NET is what you get with XP and various versions of .NET depending on when you got your Vista).

Now of course Mono on Linux is scheduled to support Silverlight but not nearly all of the Silverlight 2.0 features. In Soccer they call this the Offside Trap, in Redmond its the Proprietary Trap. Get developers to think they are getting complete cross browser and maybe some cross OS platform RIA performance - and deliver a lot less. Remember the crucial test for an RIA is will the app run offline and online on every platform its used on. Like Microsoft’s current support for HTML/CSS/DOM standards, Redmond is the most seriously deficient of RIA providers as well. Beginning to sound like a broken record ? Well lets see exactly what Redmond is promoting.

What Developers Should be Demanding of Redmond

Steve Apiki has written a piece on Nine Silverlight 2 Features Not to Be Missed. Here is what Web 2.0 developers should be demanding of Microsoft before they even consider Silverlight.

1. HTML DOM Integration
For Silverlight applications that are deployed as a control or element of a larger Ajax application, tight integration between Silverlight and the HTML DOM means no bump between Silverlight and its hosting page, and a better user experience. The number one use case for DOM integration before Beta 1 may well have been gaining access to HTML controls, but now that we’ve got Silverlight native equivalents we can move on to enhance page integration.
As noted, IE7 and IE8 still have the worst DOM/CSS/JavaScript standards conformance by far. Don’t let Redmond make permanent those proprietary extensions and shortcoming by short- changing RIA and Silverlight too.

2. JSON Serialization
Silverlight 2 Beta 1 applications can use the new .NET 3.5 DataContractJsonSerializer to serialize .NET objects to JSON. In our example project, below, we’ll use this class to serialize our object for local storage. Since the representation is JSON, you can also use this technology to deserialize JSON strings from a web server or web service.
There are a number of Open Standards for JSON serialization, don’t let Redmond bypass these and sell you their proprietary JSON extensions is the best they can do. redmond is being given credit in the Economist for being more Open Source oriented and delivering more Interoperability. Here is yet one more chance for Microsoft to deliver - there is a lot of great Open Source code here and throughout the RIA world..

3. Styles and Templates
In the Silverlight control model there’s complete separation between control functionality and control appearance. The only thing that’s intrinsic about a button is its functionality; that is, it fires a click event when clicked. The appearance of the button and how that appearance changes for various states can be lightly or fully customized using Styles and Control Templates. It’s important that these customizations are expressed and applied in XAML, not in code, keeping them in the realm of the designer and not the developer.
The Silverlight offensive is really about selling XAML, the key to WPF and Vista, to developers. But we all know that the most serious source of bloat in Vista is XAML and the GUI interface. This is proprietary squared. Do you really want to Bloatize your RIA and Web 2.0 Apps and make them decidedly non-standard/proprietary. Redmond has seen no reason to explain why it has passed on SVG, XUL, MXML, and other XML-based GUI frameworks. Clearly Silverlight is a proprietary drive.

4. Local Storage
Isolated Storage gives your Silverlight application access to storage resources on the client. It’s “isolated” because the store is partitioned per application, meaning no other applications can access files in your storage. On the other hand, your application (application defined by its URL) always gets the same storage, even if it’s run in a different browser.
This is one of the standards that has yet to be decided. But nobody has yet codified and standardized how this most important RIA feature should be done. The Google Gears model has some cross platform support with Adobe Flash/AIR supporting most of Gears. But the last thing you want to do is have a repeat of the SQL standards mess - every vendor pledging, nobody delivering the full SQL standards yet. Hence silos information databases and “locked” stacks plague the database and app development field to this day. Even worse is to let Microsoft set the standard without reference to general trends in the field; their Local Storage for SilverLight does just that.

5. Databinding
Silverlight now supports two-way, one-way, and one-time databinding between visible controls and classes in code that represent application logic. One-way and one-time databinding are for read-only controls. Two way databinding lets the user make changes that automatically update classes in the model. You can also bind visual controls to static XAML resources, and indirectly to other visual controls, but I’ll focus on making the binding in C# code here.
Here we have another convergence on proprietary vehicles in Silverlight. If C# and XAML were completely cross portable to Linux and Mac as Java and ActionScript are then I would have no complaints. But everybody knows that C# and XAML are only fully featured and compatible on Windows Vista and NET 3.5 - a very limited computing base. And as for a Mac presence ….

6. Generics
One of the best parts of Silverlight 2 development is access to a substantial subset of the .NET 3.5 Framework and the CLR. The CLR supports generics, classes and methods that are written to work with types that aren’t known until runtime. Generics allow you to write flexible classes and methods without losing the benefits of strong typing.
Generics is complex C# and .NET coding; but also again not available with full features on Mac, Linux, Solaris, many Unix, etc.

7. ItemsControls
An ItemsControl is a UI element that displays a list of data objects. Doesn’t sound like much, but the power of the ItemsControl is the flexibility that this simple mission provides. The ItemsControl can display the items of any enumerable collection, and it can display it in any fashion you like, all set up through the declarative magic of templates.
The only complaint here is that templating for GUI is already non-standard, this feature adds to the confusion. Sort of like the many competing JavaScript frameworks in the Web 2.0 world.

8. Layout Management
Along with its new controls, Beta 1 provides flexible layout management derived from the layout management system of WPF. To position controls within your application, you use one of three types of layout controls.
This is another XAML specific capability which emulates XUL and SVG which proceeded the Microsoft proprietary “solution”.

9. WebClient
Silverlight applications can consume data from REST web services using the WebClient object. In fact, if you want to get media as a stream or need to process plain text or xml data, WebClient is your best choice as it can get anything you need from a web server. The coolest thing about WebClient is that it can operate asynchronously, pulling down data in the background without blocking the UI thread. When the download is complete, WebClient fires an event to let you know that the data is available.
This is like a broken record; but Web Client is yet another Microsoft proprietary solution which reveals the underlying intent of Silverlight - to return through RIA, Web 2.0 and GUI Service in the Cloud to a Microsoft dominated technology.

Summary

So the question becomes quite simple. Do users and developers want to reward Redmond for its absolute maltreatment of the Web interface from roughly 1999 through to 2006.? Do users and developers want to submit to pseudo-interoperability (we have Mono on Linux but not Mac, Sybian, Solaris and a host of other OS - and Mono may run Silverlight depending on Novell’s implemetation…) ?Do users and developers want to let Microsoft set the defacto and proprietary standards for Web 2.0 and RIA? Remember Redmond’s failing grades on meeting promised support for W3C HTML, CSS, DOM, This is a period when Microsoft failed to make any updates to the IE browser including long promised conformance to DOM, CSS, JavaScript and other W3C standards and just stopped cold implementation of W3C and other Web standards across not just IE but its complete Web development apps. Do users want to submit in similar fashion to the proprietary, non-cross browser, non-cross OS platform for the highly proprietary feature set of Silverlight as the RIA of choice ? There are plenty of open and cross browser/OS platform choices like Curl, Flash/Flex/AIR, Google’s Cloud tools, OpenLaszlo, Sun Java/JavaFx and others which have better feature sets, performance, and much better interoperability + Open Source commitments.

So why submit to Redmond’s Proprietary Trap ?

I am doing a review of the O’Reilly book, Google Apps: the Missing Manual, and it brings up a key point about the Google empire. It is accidental and becoming static. But it is an Empire because Google has just gotten Yahoo to sign a deal that will have Yahoo displaying Google Search ads on Yahoo’s site. I don’t understand the technical reasons for this because Yahoo has its own Search Marketing ads and Yahoo Publisher Network for publishing those ads on user websites. The argument is that the Yahoo Publishers Network version is still nascent and so Yahoo needs to borrow Google’s ad engine for awhile. And since this is a non-exclusive arrangement that Yahoo can terminate(but when and at what cost, undetermined) - Yahoo is still considered in the drivers seat on monetizing its properties and search engine capabilities.

However, the net effect of the deal is the expectation that Google share of the Search ad market will go above its current 60% - call this the apogee of Google’s Search success.

Google like Microsoft before it has an accidental empire - admittedly a self-made accidental empire. Why accidental ? Because for the first 4 to 5 years of its existence, Google did not have a model for monetizing itself. It was looking for businesses to pay for having search engine software or software+hardware added to their websites. Nice but not huge. But at the sametime Google was mastering the science of supporting monstrously huge Linux based computing centers that could deliver the very fast response time and more relevant search than any other provider. Not until late 2003 early 2004 did Google start Adsense which charged clients for clicks on clients small Google placed ads. And because advertisers didn’t pay until clicks were made on their ads - this has become one of the predominant advertisng method on the Web.

But where are those Google ads ?

On tens of millions of users websites. How much does Google pay for the click revenue it gains - well Goggle is not telling. But if you go to the other side and bid for keywords and Google ad placements, one can see the payment is not much more than 10 cents on the Google revenue dollar. This is the small amount yet highly voluminous and profitable revenue stream that Bill Gates declared he was looking for way back in the early 1990’s. Sergy and Larry, almost by accident, got their first. And for some reason, neither Yahoo nor Microsoft have been able to duplicate or find an inroad onto Google’s tens of millions of advertising sites.

But where does Google go next ?

Stronger search capabilities? Actually, Ask.com and others are leading Google in delivering new ways to do searches and get more relevant results quickly. More contxtual search and look up. I admit I use Wikipedia for Contextual Searches and that its use now dominates Google and all other search engines. And now Powerset is supercharging Wikipedia with Natural Language that still seesm to elude Google. However, Wikimedia Foundation does not accept any ads.

Use of media and collaboration as revenue streams? Google Mail, Calendar, Chat, Talk are all Google ad sponsors but otherwise free services. SaaS-Software as a Service ? Google Apps is an online Apps offering equivalent to roughly Microsoft Office and is free for up to 5 related users and then charges on annual basis. Google Media - Google images, You Tube, Google Earth, and Google Maps are all free services in which ads sponsors provide sources of revenues.

From this listing it appears that Google’s next major revenue stream, not advertising dependent, will come from SaaS. Which SaaS is hard to ascertain because Google makes so much money from such easy ad sources, it is likely to become victim of the Schumpeter Disease or The Microsoft Effect - “Easy sources of revenue are hard to displace until it becomes too late”. But the bottom line is that Google, which has largely ignored good GUI design(Google Mail like Google Apps are no great GUI shakes nor is Google Widget Toolkit. And with touchscreen and gestures coming to PC desktops well before the Web, Google has some real GUI catching up to do) and customer service(hence the O’Reilly Missing Manuals), - Google is not a shoe-in in the SaaS world.

I am amazed that a virtual company like VMware or Citrix/Xen has not teamed up with a PC vendor like Dell, HP, Lenovo or Asus to offer a packaged VM desktop that runs in Linux and then your choice of Vista, XP, or Mac as a VM machine. And VM would be preset-up and loaded on the PC so context switching would be trivial. I would be willing to pay $300 to $500 for this capability. And Microsoft would love to sell more Vista and - ooops now you understand why Microsoft has been adamant about shutting down sales of Windows XP as of June 30th 2008. And as for Apple and Steve Jobs allowing for a Mac/OS VM package - you will have to wait until HFO.

Now tell me dear reader, is this the kind of OS provision and service in the market place you had in mind when Microsoft lured you into giving them a 95% monopoly of the desktop OS market share ? Don’t you just love to pay the Windows, Office, Exchange, and other Microsoft taxes for your IT operations ? How do you like the service and support you are getting ? Just asking.

This is the second time I have picked up strong signals that Management at Microsoft is showing the stress of 7 years and counting of very mixed innovation. The results are very low at times -IE delayed almost beyond rescue as Firefox 3 just trumps even the upcoming IE8, Vista Death Watch, Zuneys, Silverlight-is -not-open-enough, versus some hits like SharePoint, HP’s fuly touchscreen PC machine, and interest in Oslo modeling(delivery still awaited). And of course this has been reflected in Microsoft stock which has has been treading water in the $30/share or thereabouts for 7 years as well.

Microsoft clearly missed the Internet. Then the search bounty. Then the power of Open Source. Now the move to the Cloud - still trying the old proprietary monopoly formula. Well, read this e-piphany from eWeek , the pleadings from a supposed Microsoft blogger, and a interesting analysis from the NYTimes, for insights into the nature of the Microsoft malaise. Clearly employees, who depend on a rising value of stock for a sizable chunk of their compensation for Herculean work schedules - they must not be happy campers. Shareholders must be a bit restive with very substantial dividends coming at random times. Bondholders are non-existent - and that is a good thing because BCE decision just confirmed that Boards of Directors don’t have to consider bondholders when making decisions other than wha is written in each bond issues covenants( that was a Yahoo you heard from Private Equity investors).

So maybe Economist Joseph Schumpeter is right on Denmark - monopoly profits drives out innovation because what can beat the ROI and low risk of existing monopoly markets versus the risky returns of new and completely different ways of doing things in emerging markets. You have to set the bar so high to match those monopoly returns, an organization ends up spinning its wheel.

This is a review worth repeating.

I would like to tell you that the new Firefox 3 web browser is blazingly fast, full of new GUI features yet with great ease of use. And FireFox 3 has one of the best set of add-ons, themes and extensions. It leaves even the upcoming IE8(early next year likely) in the dust. But I won’t - instead I will let the Microsoft evangelist, Paul Thurrott tell you.

Now if you have an 5 minutes of time (it really is that fast for download and install), upgrade to the Web 2.0 browser of choice - its that good. Don’t go browsing or blogging without Firefox 3.