Changing Climates for Microsoft and Google, Desktops and Webs
Written by John Milan and edited by Richard MacManus. This is the first in a two-part series. Update: Part 2 is now available.
The most
insidious thing about global warming is that it's everywhere, but not right in front of
you. And really, it has more to do with things disappearing than appearing - things like
glaciers, ice shelves and low lying islands. The odd thing is that the progenitor of
global warming, the Industrial Revolution, was rooted in making things disappear as well.
For example, teams of horses became a steam engine or two.
Though not as pernicious as climatic change, the software environment - admittedly nowhere near as complex, but growing more convoluted every day - is tracking a similar course. Just as odd, the Information Revolution was also rooted in making things disappear. For instance, teams of typists became a shared printer or two. But now newspapers are shrinking, phone and cable companies are mutating and the distance between people is vanishing. In fact, just like the first scientists researching the rise in ocean temperatures, a group has recently been commissioned to study the ramifications of the web.
Windows is leaking, while a Google hurricane forms...
Revolutions have many components, including the seeds of their own destruction. Once the genie escapes from the bottle, he begins enforcing the Law of Unintended Consequences. Ask an industrialist in 1875 if his coal burning plant could change the earth's climate and he would have thought you're from another planet. Ask a venture capitalist a few years ago if operating system hegemony could become an albatross weighing a company down, hindering entry into new markets and emboldening competitors - and he or she would have thought the same.
And yet here we are today. The climate is changing, Windows is leaking and Google is ascending. The Industrial Revolution heralded a leap forward, but it also caused lots of carbon affecting all life. Similarly, the Information Revolution liberated data - but it resulted in lots of programming that has taken on a life of its own. All the highly touted, next generation, advanced features for Vista? Flushed away by the rising tide of complexity and backwards compatibility. The challenges for both revolutions are similar: keep the good, but fix the bad.
It hasn't happened overnight. Rather, its been a trickle - which started the very second that Windows shipped. Water actually started pooling around Windows back in the mid 90s when Netscape offered a portal to a new frontier. Bill Gates - well acquainted with weather in Seattle - recognized the moisture in his foundation, installed an Internet Explorer sump pump and weathered the storm.
Now it's the mid 00s, and it's Google gushing with ideas, cash and optimism. Many observers think a hurricane is forming in Mountain View that threatens to swamp not just Windows, but Redmond's entire empire. Well, forecasting has never been an exact science - and I wouldn't fret for Microsoft quite yet. Sure the vista may be a bit foggy, but Microsoft has a few more resources than anyone else. Even as the tide rises, they already have newer, more sea-worthy vessels well under construction. And Google? They might just miss the boat.
GoogleOS actually fogging the field
The mere mention of a new operating system competing with Microsoft reminds everyone how much fun the old days were. Apple, IBM and Microsoft continually sparred, parried and blocked each other for many years before Windows finally won the battle. It's gotten downright boring since those days, so it's easy to understand the tizzy surrounding even the prospects of another brouhaha. But Google building, releasing and supporting their own OS? I'd sooner believe them walking on water.
Bringing an operating system to market is an extremely poor idea if you want to grow a business. Indeed, recent history has plenty examples of great new operating systems - OS/2, Taligent, BeOS, JavaOS to name some prominent ones - that siphoned off huge amounts of resources and either left the company in ruins or seeking greener pastures. An ill wind blows for any OS leaving the comfy confines of its incubator. And Linux? It's not a business, and therefore not in this article - but I think Microsoft fears Linux a whole lot more than GoogleOS.
First a brief primer on why it's great to own an operating system. In short, it's a lot like owning an industrial age railroad. Railways and trains are like operating systems and applications. Operating systems are the rails themselves. Applications are the trains which, of course, run on the rails. In order for trains to travel on a railway, they must match the railway's gauge. In order for applications to run on operating systems, they must match the OS's API. If a train doesn't match a rail's gauge, or an application doesn't match an OS's API - things don't turn out well.
As the robber barons found out, when a railroad achieves critical mass you make a boatload of money. Better yet, once that track is laid it's very difficult to move aside. Like any good robber baron, Microsoft would love for Google to compete on the same field where Microsoft owns all the rails and most of the trains. In fact, that kind of GoogleOS is Microsoft's wildest fantasy come true. Microsoft would be fully capable, and would probably relish the opportunity, of roughing Google up, dragging them through the mud and sending them on their way - just like they did with Sun, Apple, IBM, Netscape, Novell and Digital Research.
What Google wants is new real estate on a higher level - a better location where they can lay their own rails. But while it may look like they've found the best location on the web, Google is in fact still traveling on other peoples rails - Microsoft's, Apple's and Linux's. They do so by offering free train rides on various implementations of the world's most successful virtual machine. But the free ride won't last forever. Google most likely will need another virtual machine to compete and grow. Otherwise, they may face a fate similar to yet another high flyer from the 1990s (see if you can guess who!).
A perfect Virtual Machine storm is breaking Windows
If you can't run on the rails, you can always take to the air. Such a notion probably would have seemed silly to Jay Gould or Leland Stanford - at least until Orville and Wilbur Wright traveled by rail to North Carolina. While Orville and Wilbur might have not have understood the business pitch of Larry and Sergey, they would have recognized kindred spirits riding along the world wide web on someone else's infrastructure.
Consider two classic applications for two platforms. One is more or less owned by Microsoft, the other more or less owned by Google. The apps are familiar to every programmer: 'Hello World' done in C++ and HTML.
In C++:
#include <iostream.h>
int main( int args, char **argv ) {
cout << "Hello World" << endl;
return 0;
}
In HTML:
<HTML>
<BODY>
<P>Hello World</P>
</BODY>
</HTML>
As always, the devil is in the details. What is not shown is the C++ compiler and linker that turns code into executable. Also not shown is the web browser which takes HTML and makes it presentable. And that's really the only difference between these two programs. Of course the ramifications are profound: the C++ application can only run on the operating system it was built for, whereas the HTML application can run on browsers, which in turn run on operating systems. In fact, the browser is really a virtual machine. The world's most successful, widely deployed, virtual machine.
There are two additional properties that add critical value to this virtual machine:
1) An explicit contract on how to install and uninstall applications; and
2) An explicit contract on how an application can affect a user's machine.
These two features have tipped the balance so far that users are willing to put up with a more rudimentary web UI than rich desktop UI.
Google's high tide?
When you visit a web page for the first time, you are installing an application. The HTML page your browser is reading is both a manifest for additional resources (images and graphics, cascading style sheets, javascript and embedded objects) and instructions for how the page should be rendered. Your browser stores as many of these resources as possible on your local disk, in order to start up faster the next time. Better yet, it checks whether these resources need updating on each visit. So why doesn't your IT person freak out when you install HTML pages, like when you install desktop applications? Because of the explicit contract that limits how an HTML application can affect your machine.
Purveyors of C++ applications realized the advantages of HTML applications immediately and even tried to address the manifest issue. But while they could create a download manifest, they couldn't enforce the necessary constraints that make IT people happy. Additionally, the applications they were downloading were much larger than piecemeal HTML pages. Throw in a few security holes and well publicized exploits and it's obvious why HTML web applications are in such favor.
These are the advantages Google enjoys today. However, two fissures exist that will force them to move:
1) Microsoft's ability to use the exact same HTML based strategy (like their current Live initiative); and
2) More threatening is Microsoft leapfrogging the current environment by solving rich application installation/uninstallation and enforcing an acceptable contract regarding what rich apps can do on a user's machine.
Unfortunately for Google, Microsoft is a lot closer to solving these two issues than people think. Microsoft has the best virtual machine with .NET, the best development tool with Visual Studio and the best access to developers with their MSDN programs. And they have a notion. Steve Ballmer himself has started touting the exact strategy they need - Click Once and Run. The final thing they need is a technical solution for .NET - similar to what a little startup in Redwood City, CA has done for Java...
TO BE CONTINUED... Join us for the next installment, when we find out just who this little Redwood City startup is!
Image by Jon Cornforth
Update: Part 2 of this series is now available.
Share3 TrackBacks
TrackBack URL for this entry: http://www.readwriteweb.com/cgi-bin/mt/mt-tb.cgi/2889
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
This is a great piece that reminds us all how the online world is a complex and tricky business. Yet there are so many similarities with the past many ignore. There is so much to learn from history.
We witness an ever increasing modification of business models where those who cannot produce significant added value are removed. But this has been happening since the stone age, it's how fast the rate of change occurs that makes it more and more challenging ...
Can't wait for the second part... Thoughtful post! But as for GoogleOS argument, I don't think GoogleOS will be yet another OS or another loser, first of all, it won't be a real OS. It will just be a shortcut to the field where Google is obviously dominant, internet. Internet itself will be an OS like platform thanks to AJAX.
In other words, what Google will do is just to cut the middle-man who dictates its own technologies and lowers down Google's market shares.
But of cours this depends on Live and Vista's performance. I don't think Google would attack if they don't need to.
I agree with my friend Emre here, this is a very interesting article...awaiting the second part.
I think we need to wait and watch the performance of the MS Adcenter advertising program, the money that MS will make from the LIVE initiatives is expected to come from here.
As for Google, they already have a highly successful advertising program - Adwords and they can leverage this in GoogleOS - online "operating system" which could be free for use but supported by ad-based revenues. As the rumor goes, if it is based on Linux then many of the current Linux users would be a direct convert to GoogleOS, hence I think Microsoft should fear it as much as it fears Linux.
I think it makes great sense for Google to release GoogleOS and increase one more source for cash flow.
Cheers,
Abhishek Sharma.
htt://www.iconfluentia.com
Good article overall but C++ vs. HTML part kinda sucks.
A well written and insightful article. My take on all of this means breaking out the consumer and the corporate space. I'm not convinced that many corporations would adopt a GoogleOS anyway. If (for many organisations who see only the running costs of a desktop realestate) this is all about deploying, running, maintaining and upgrading applications [the bane of many an IT Manager] then the GoogleOS model would be an interesting one. But everything else needs to fall into place - the planets need to align - in order for this to happen. Many business people are not always connected to the Internet, many of them still prefer islands of data and don't trust data stored by some company in Mountain View. If this were not the case then Citrix would have been far more prevalent in the business world, despite being expensive to procure and license.
As Steve Ballmer comments in the linked article Microsoft see the partnership they have with the PC manufacturers (some say stranglehold) and chip manufacturers being the key here. If browser delivered applications and content is the way, then maybe I don't need such a powerful PC? Great! Maybe I don't have to upgrade every two or three years. Great! I'm sure those people at AMD and Intel wouldn't be happy though. Unless Google buy all the servers in the world to compensate from those manufacturers the cycle that drives processor sales and software sales might be broken. So for me a GoogleOS could work for consumers, but I'd be surprised if it would work for organisations.
The thing that's really going to hurt Microsoft is their licensing. Yes, they have the premier virtual machine and the best development environment. BUT, you can write a Web application for Linux and scale the application on a computing grid turning virtual servers on and off as needed. You can't do that with Windows, and right now .NET only runs on Windows. Unless Microsoft sees the writing on the wall and open-sources either Windows or .NET so it can be ported to Linux, it will be a dead technology within three years. Even though it's easier in the short term for product development, you also have to consider long term operating expenses once the application is up and running. There's just no economical way to scale a .NET application right now, the ROI goes out the window once you need more than a couple of servers, or if demand spikes and you need to scale for that.
I am very skeptical of the .NET argument. This sounds like more MS hype to me. Even if Vista proves to have a "better" environment this hardly solves MS' problem. As long as there are applications that people want that do not run on .NET, MS will have to support the standard infrastructure or loose more market share to Firefox for example. It is very clear that Vista is going to be a very long rollout for corporate customers and consumers as well (how many consumers are going to buy a new pc just to get Vista?). How long before the majority of consumers have Vista on their pcs? 3 years? 5 years? ever?
The other problem is the word, better. MS has for decades been using the geewhiz approach to software marketing, "... our software is better because it has more features..." This has finally run out because everybody is overloaded with features, users now want more simplicity and are interested in web applications because they are serving user's needs. Features in the abstract are not having the same impact from where I stand.
Even if you like the .NET development environment, it has several things against it. First, not everyone is actually in agreement about that. MS mindshare among developers has been plummeting for years and from what I can tell, .NET has not had a big effect (so far). Part of the reason for that is probably in the cost of a .NET developer's environment. Independent developers can get as much or more power for far less money going the opensource/Java/PHP/Python/Ruby route. Since independents have a much larger impact due to the ubiquity of the net and the innovation they continue to demonstrate, it seems unlikely that .NET is somehow going to gain the upper hand. The best that MS actually offers is more of a "... gee, we can do that better..." approach. Which of the breakthroughs on the web were written in .NET? Can you name even 1? .NET is, in essence, for the stodgy oldtimers. It is an important platform in that it has and will continue to have a large base of adherents, but I fail to see how it will ever achieve dominance.
Thanks Sante, I think analyzing events with a historical eye is not only interesting, but as Mark Twain said:
"The past does not repeat itself, but it rhymes."
Maybe the C++/HTML section doesn't stand by itself in this part. It is setting up an argument in Part II though. First showing exactly the current HTML advantage, and, in Part II, showing how this can strength can turn into a weakness and offerring a potential counter .NET can/should use.
As far a scaling is better done on the web with server farms... I've never really seen this as a clear-cut, web farm only type of advantage. And, fundamentally, massive server roll should still pale in comparison to the raw computing power on everyone's desktops.
For example of the SETI program that lets individual computers solve pieces of vast problems seems like such a beautiful solution. Maybe the SETI program makes me too starry eyed, but I feel that's a more powerful blueprint for the next step in computing. But I'm not a rocket scientist. We need a better framework for the average joe's like me to build it.
Interesting article.
When I look at this from outside the box.... I see Google not so much as a traditional competitor to MS in the OS arena but more as a comapny gearing up for the next cultural shift in business needs... which will be mostly related to small to mid-sized companies. Such businesses will continue to evolve by not having the traditional unnecessary expenses that were once considered to be standards for starting and running a business (ie. large b&m space, large IT depts., large hardware investments, high License costs and slanted support contracts) .
In other words.... a period of enlightenment for sustaining a business is here and will only continue... forever. It is the time of new thinking and efficiency. Google will work to cater to this "sector", which will be greater in quantity than the traditional breeds (though they wont disappear too quickly).
GoogleOS is prob more of a vision in response to where things are heading. And will also cater to very small groups and individuals as well. The desktop OS will be used less for features and more for simplicity and stability... give or take a few application needs that desktop OS will do better for forseeable future.
Regarding .net... #7 comment notes:
".NET is, in essence, for the stodgy oldtimers. It is an important platform in that it has and will continue to have a large base of adherents, but I fail to see how it will ever achieve dominance."
I comletely agree and have seen this in my experiences working within tech industry. It will never be more than what it already has become.
OK, I just read all of Ivan's (#7) comments and I have to respond.
Fail to see how it will every achieve dominence? It has defacto dominence by shipping with every version of windows, which ships with 90% plus of all PCs. It can't help but be dominant and this, in turn, can't help but attract developers.
I do agree MS developer mindshare has been shrinking from where it was. But that speaks to my point. Right now the web *deployment* evironment is so much better for developers that a great many are indeed lighting the LAMP, so to speak. As a developer, I'll put up with alot if it makes it easier for people to try use my apps.
But actual development? I guess I'm old school-- the compiler and debugger feel so warm and safe :)
But to I'm not sure what MS makes off of the developer community, but I'm sure its miniscule compared to everything else they sell. Because they are in a battle, I think they should price Visual Studio next to nothing. It is a sign of hubris, and perhaps a little greed, how expensive their development tools are. And history has a hard lesson for things with too much of both.
John,
just a quickie, I don't want this to turn into one of those tortuous back and forths that drive people crazy.
I think you miss my point. Even though MS has 90% share or more of the pc market, that has not and I claim will not translate into .NET market dominance. As far as consumers and web users are concerned, the two are unrelated. For MS to achieve dominance with .NET it needs to get developers to make that their platform of choice. So far that has not happened.
I believe this is significant, not because the developer market is so big, but that is one of the primary ways that Windows triumphed over OS2 among others. They got the developer community to create compelling Windows applications which consumers and corporations needed or wanted. I do not see this happening currently. In fact things appear to be going in the opposite direction.
I would love to continue this conversation, but I suggest we do it in email since I think that others may not appreciate our back and forth.
Thanks for your feedback.
--Ivan (ivan.handler@gmail.com)
The comparison between a markup language like HTML and a general purpose programming language like C++ is absurd. Also a bit silly is the idea that a browser is, in and of itself, a virtual machine. I'll concede that more often than not a browser will implement a virtual machine in the form of a javascript interpreter; however, with many successful exploits targeting particular js implementations, this seems to be something of a liability.
There are two additional properties that add critical value to this virtual machine:
1) An explicit contract on how to install and uninstall applications; and
2) An explicit contract on how an application can affect a user's machine.
Microsoft was pushing the idea of a "Smart Application" onto its developers a while back, I wonder if this is what you're referring to? There doesn't appear to be much that came of it, in fact the only concrete implementation Ive seen recently is the MSSQL 2005 Reporting Services "report builder". Applications of this nature may use many of the same technologies (web services, soap, etc.) but generally run outside of the browser and require a .net runtime on the client.
Lastly, I seem to recall a rumor or two suggesting that Google would partner with a particular Linux distribution to create the GoogleOS. I don't know how realistic this is, but the idea of a stripped down Linux distro that offers little more than a modern browser (with the requisite plugins) and the necessary software to quickly get the user online seems like a pretty great idea.
Distilling the advantages of Web apps into the two points about the contracts regarding installation and machine access are very insightful. However, Microsoft faces a huge perception problem with those contracts in .NET.
Point 2 "An explicit contract on how an application can affect a user's machine" is the tricky one. Java in large part didn't get off the ground because of early concerns about that "clause" of the contract. Part of this was due to some well-publicized early exploits (ie. actual implementation faults), but the other large issue was that IT people felt that the VM, runtime and security model were too complex to fully trust until it had proved itself a long time. The issue wasn't distrust of Sun, but rather fear of potential flaws.
The same concerns apply to .NET, especially given the reputation Microsoft has built in the wake of the innumerable security breaches in its products. However, the kicker is that most web people don't just distrust Microsoft's security -- they also don't trust Microsoft. They expect that Microsoft will try to overreach to implement a "cool" feature and/or flat-out pull a fast one. We've learned how these guys do business by now -- and I'm speaking from the bias of someone in the Windows camp.
The further challenge for Microsoft is going to be coming up with an acceptable contract. Will they be able to produce a contract that doesn't have fine print smuggled in that effectively says "...we reserve the right to modify or add any 'feature' at will"? Will they be able to resist the forces that are surely calling for some kinds of monitoring, "rights management," license enforcement, etc in the platform? And will they be willing to sign away their right to add such clauses to future versions of the contract, if/when they achieve critical mass and have us by the short hairs? Given the history, I'm not betting on it.
agree with the article, but please, take it easy on the metaphors... makes it horrible to read.
You didn't like Internet Explorer sump pump? :)
Actually, be thankful Richard does the editing. He keeps me honest.
Google OS using advertising revenue for funding? Are people crazy, I had a hard enough time using Napster bcse. of Ads ... now an OS with an embedded integrated Advertising network? Does anyone think that will fly? I personally wouldn't even be open to viewing it's screen shot's ..
My bet is virtualization will kill any Google OS initiative that target's the desktop.
On the other hand, if Google OS is simply an embedded HTML window manager - that's full of ads - how many crazy people will use the technology (because obviously sane individuals won't like this).
Next thing you know "digital photo frames" will be free, but will come with 10 built in advertisments - they will be displayed in between your personal pictures ... sweet ... picture of my daughter - 10% off at Staples - picture of my son - Free printer at Best Buy ... yes ... and no doubt BECAUSE OF SOME HALO EFFECT, YOU'LL ACTUALLY SEE WIDE SPREAD ADOPTION OF MY PHOTO FRAME? Where are those crazy people I was talking about earlier ...
Great points, Miles (#13).
You're right in pointing out the difference between the .NET technical solution versus the reputable solution. And yes, it could very well be reputation that tubes .NETs growth.
Regarding Java applet adoption, yes there were concerns about a VM approach. However, I do wonder about the root of those concerns-- and how things would be different with MS backing a VM on their own platform instead of grudgingly using someone elses.
I haven't heard downloadable .NET resistance on the same scale as, say, ActiveX. But then MS hasn't really pursued this yet, either.
It probably comes down to how many suits versus sandals are in the room when deciding the .NET licensing rights.
It's an issue that, without question, can tip the balance in Google's favor.
Operating Systems are irrelevant, just as the PC is becoming more and more irrelevant. Devices and their access to the internet are what will be important in the future. Because of that, the web platform (Browser,HTML,DOM,CSS,Javascript,& server back-end) will win in the long run.
Don't think in terms of Google and Microsoft. Think in terms of devices such as: Smartphones, Sidekicks, iPods, OLPC laptops, PS3's, XBox 360's, Smart PDA's, Tivos, convergent consumer electronics...all these things have one thing in common...they will soon be hooked up to the internet. None of them have a common operating system among them. The OS is irrelevant.
Also, the idea that big business and corporations are the major indicator as to the direction of the software industry is short-sighted. There are over 1 billion people and counting online. The music and movie industry as well as print and tv news are moving to the web. Everything is moving to the web.
The fact that office buildings haven't moved their productivity applications to the web (as a platform) is almost insignificant. To make use of your references to history, the world of the PC and the office productivity app are coming to an end. The day of mobile/online everything is at hand...
The whole thing is about how secure or restrictive the (virtual) is. The more powerful the machine is the less safe it is, the safer the machine is the less useful it is. Of course, the most powerful are open collaborative systems. HTML is considered safe, CSS less safe, JavaScript even less safe, .NET not very safe, ActiveX insanely unsafe. Any powerful environment requires sophisticated sandboxing policies to keep it from becoming dangerous and destructive. Can .NET VM and IE avoid what happened to JVM and Java applets? Only if Microsoft can convince us that living dangerously is good for us. In the past, they successfully convinced us that using ActiveX is dangerous. As far as Google, they at least understand the difficulty of the issue and are responsibly taking some extra time to find the solution, so there will be no GoogleOS yet.
Odd how these things happen. Who knew it would be 'pick on Google day?'
http://www.smallbusiness20.com/Blog/tabid/6307/articleType/ArticleView/articleId/1211/Default.aspx
I *like* Google and use a few of their services. The world would be a poorer place without them.
Note for John and Ivan -
The VS Express stuff is now free and will remain so. I'm not an experienced developer (really just starting out) but it does an excellent job for me so far, allowing me to basically build whatever I want in .NET at no cost and distribute as I please. It compares pretty well to the costware versions.
So in effect, VS doesn't cost 'next to nothing' - it costs *nothing*. The LAMP stack (or Linux-Apache-MySql-{someotherlanguage) offers the added inducement of providing a free *platform* for your code to run on, and that's pretty huge. But I just wanted to correct the idea that devs have a significant buy-in cost to write .NET apps ... it ain't so!
Hi John,
Excellent article today.
I remember you folks from my days at Groove Networks. Lets try to reconnect...
Brian.
The most secure virtual machine is the .pdf system. Adobe will release Apollo in 2007, it will use the Safari html engine inside the Flash player contained by the .pdf sandbox... sort of like inventing an air force.
Microsoft has the best virtual machine with .NET
You state this as fact. I'm curious why .NET's VM is better than Java's VM especially with the soon to be release Java 6 supporting scripting languages as 1st class citizens.
While the article is well written it is falls flat with its conclusions... the conclusions are a myopic view at best...
The traditional view of the app has changed... the killer apps have all moved on from Windows. Let me explain.... what were the killer apps in the last few years?
Google, Yahoo, Amazon, Ebay, Myspace, Youtube, iTunes Store .... these are all killer applications...none of these applications run on windows...! they run on distributed servers running Linux... (or OS X server in case of iTunes...)
Windows is quickly becoming what Marc Andreesen claimed it would become... a bag of drivers...
There you go. Windows is already obsolete... Want to create the next killer app? You are better of moving to Linux... (or OS X) where all the other recent killer apps are running....
End of story.
Microsoft has the best virtual machine with .NET, the best development tool with Visual Studio and the best access to developers with their MSDN programs.
Hmmmm...I beg to differ. I've developed in both, more extensively in Java because if it's lifetime is longer than .NET's, but currently I'm consulting on a large C#/ASP.NET 2.0/AJAX web application project. Visual Studio 2005 has a lot of catching up to do before it even comes close to JetBrains' awesome IntelliJ IDEA development tool, and that's giving VS 2005 the benefit of Jetbrains Resharper add-in. Hell, both NetBeans 5.5 and Eclipse 3.2 (open source and free for all uses) are more feature-rich than VS 2005.
The CLR (the VM part of .NET) is good, but certainly Java 5 and 6 VMs perform very well also (and Java VMs work well on multiple platforms, including Solaris, Linux, Windows, etc). The MSDN access, well, you pay for it, so of course you're going to get access to your developers. But I would say that the open source community around Java is much more vibrant and allows for competing solutions to a problem. I don't see this in .NET. I see much better frameworks being produced in the Java environment (Hibernate, Spring, AspectJ, the list goes on).
Regarding .NET as the best VM, yes its my own humble opinion that I strongly believe. I didn't qualify why I think .NET is the best because the article is already a two-parter. But, here's a summary.
I gnashed my teeth orginally in classic Windows with a Charles Petzold book under my arm. Then I went through COM with Kraig Brockschmidt. After rehabilitation, I went heavy into C++ and STL. All this was a mix of GUI and server programming.
Then I did my first Java UI with the 1.1 framework. Moved on to Microsoft's java VM implementation for a server app, then back to Java for a major web app. In true .COM fashion, we spent millions of dollars of VC money on millions of dollars on Sun hardware to fix millions of problems with making EJB scale for millions of people.
Went back to C++ programming, both client and server, and finally, about two years ago, got into .NET.
Why do I think its the best?
I accept on faith the VM performance between Java and .NET is relatively equal-- though I'd give the performance edge to .NET because of its closer ties to the Windows API. And speaking of closer ties, I've never been a fan of JNI-- way to unwieldy-- whereas .NET has a tidy solution. And I like the c# syntax quite a bit.
But the biggest reasons are:
a) you can make a great argument that Visual Studio is the best IDE, and MSDN has, for the most part, excellent help and support.
b) the resulting app looks like a native windows app. This used to drive me bonkers with Java and I absolutely feel it hindered adoption.
Throw in the fact that Microsft owns the .NET and ships it with Windows now (instead of asking the user to install this great Java VM) and, in my humble opinion, .NET obviously the best VM for software publishers on the windows platform.
It is the experience VM product development should shoot for regardless of the platform.
My Money is on the Google Platformm, Vista or HastalLa Vista :-)
Could WPF/e be Microsoft's "magic bullet"?
Thanks John,
I boil it down to the following,
Microsoft = C Drive
Google = Virtual machine
The latest wave of web usage makes the c: drive less relevant, combine this with a competitive OS alternative in linux and the Brave New World looks grim for MS.
I look forward to pt.2
Hello John,
Excellent article. Waiting to read Part II.
I agree with Emre (#2). Google should stay away from the OS market and just concentrate on delivering products/services using the Internet. Providing an OS that can be adopted by the "everyday" user as well as corporations can cause more heartache that you want (not to mention dealing with PC manufacturers, government regulators etc) as Microsoft so well knows). The browser is the "pseudo OS" for Google and the HTML virtual machine you describe is capable of providing applications that can replace a number of desktop products (using HTML, JavaScript, AJAX etc).
I agree with your last comment (#27) that Visual Studio is pretty good and .NET is a good VM for windows applications. But if you had to develop a new product that could be developed just as easily using web based technologies (and nowadays web based look and feel are getting quite close to windows products), then I would recommend going with the web based approach. The accessibility, deployment and maintenance is so much easier than windows applications.
Web based products will offer great alternatives to a number of desktop apps for users (Google Docs is just the start). That's what we strive for at ZCubes.
Thanks,
Parag Mathur
http://www.zcubes.com
Well, I for one am looking forward to the competition between MS and Google. Competition will benefit us, the consumers.
Throw in the fact that Microsft owns the .NET and ships it with Windows now (instead of asking the user to install this great Java VM) and, in my humble opinion, .NET obviously the best VM for software publishers on the windows platform.
I won't argue that the CLR is the best VM for Windows. However, most of my Java deployments in the past 5 years have been on non-Windows envrionments. And the last 2 Java gigs have been exclusively on Linux. And I'm seeing a lot more Linux in the real world. Established businesses are using it.
I haven't done a rich client application in many years--almost all of the applications have been server-based applications with web front-ends. Web 2.0 is leveling the playing field. I can build a .NET Web 2.0 application as easily as I can build a Java Web 2.0 application or Ruby Web 2.0. We have choices, which is a good thing. It's good for consumers, good for Google, and good for Microsoft.
BTW, Sun is putting Java under the GPL 2.0. This open sourcing effort will allow Linux and others to include Java as a installable package within their distros. Interesting times ahead for sure.