| This 92 message thread spans 4 pages: 92 (  2 3 4 ) > > || |
|The Second Browser War|
an article from Ben Hammersley
Ben Hammersley (of recent note for his automatic RSS validation) has written a great article about the current chapter in the browser wars - and the fact that a lot more may be at stake here than browser dominance.
His analysis really helps make sense of many business moves that look odd when seen outside this big picture.
|The difference between the two - between Microsoft control or that of open standards bodies - will be the battleground for the next two years, and one that promises fireworks. |
By wrenching control of the standards for building [web-based] applications away from Microsoft today, rivals hope they can prevent another near decade of Windows domination. Microsoft, for its part, is not going to go down without a fight.
The Guardian [guardian.co.uk]
Good post, Tedster. I had a quick read of that article when it showed up in the Guardian, (I have it bookmarked, and it's on my must-read list of newspapers everyday), but it was sure worth a second read.
|However, what would happen if people's web browsers were capable of running complex applications, with code based on openly published specifications? Two things: first, the operating system would become irrelevant, so there would be no need to upgrade to the next version of Windows, and second, the playing field for everything else would be thus levelled. |
That stuck in my brain the first read through. I'd wondered why MS had neglected IE. With Bill trying to own everything, why was IE never updated, loaded down with more dodgy code, and sent out as version 7.0? The theory advanced seems a bit far-fetched at first, but maybe it's true. If so, it seems Bill isn't as smart as he thinks he is. He's blown it. The future became Firefox, and Opera, and their like, while he was bogged down trying to dig his moat deeper.
Thanks for having me take a more thorough second read of that, man.
Sorry to disagree - NOT impressed.
Internet apps will never take off for more reasons than I care to detail. As for browser apps - you have got to be kidding. Talk about a security black hole - MS might believe something of the sort is sensible, but I don't think anyone else with programming knowledge would.
Internet apps will never take off for more reasons than I care to detail.
Kaled I think that you are completely wrong and it's been my experience that browser based applications are going to out number fat clients very soon.
The advantages of thin client development are many: No software to install on users machines
Allows for remote access
Can be developed and used with open source and/or free products (PHP, Python for example)
Integration is generally straightforward (used standard HTTP protocols for example)
Easy centralisation (You can host your app in Lonon and people in New York can access it without any problem)
I can go on! All the applications that I have worked on in the last number of years have been web based. Clients that I talk to say that they are now no longer even considering fat client applications and want everything web based.
Your point about security is moot. If you are a large organisation with a web based application people will access it via a VPN (or something comparable).
You should not simply dismiss out of hand the advantages and growing popularity of browser based applications.
I'd wondered why MS had neglected IE.
Because, I believe, they see the long-term solution as one that abandons the browser altogether, so that the browser does not become the OS but the OS becomes the browser.
The idea being that the net just becomes an extension of your hard drive as far as the OS is concerned.
Why are people with multi-GHz processors and vast hard disks going to switch in droves to web-based applications with the sort of response times that I was used to 20+ years ago when I was working on 9600 baud terminal connections?
As for security, most people believe that one of the core reasons why IE is insecure is that it is integrated to the operating system. If you effectively build an operating system around browser technology you will end up once again with a fundamentally insecure system.
What do people use computers for?
Answer (in no particular order)
2) internet browsing
4) Word processing
5) Desktop publishing
6) Database apps
7) spreadsheet apps
10) Anything else I've forgotten
Just exactly which of these are suitable for web-based applications? Stretching a point - 3) email. But how many webmasters here use webmail exclusively and never use PC-based email clients such as Outlook or whatever?
Web-based applications may have niche markets and, if you include VPNs (which is stretching the definition somewhat) corporate markets, but they will never, ever, ever replace desktop applications in the mainstream, unless, of course, Governments require us to change so that they can monitor our thoughts, etc.
Kaled: You're missing the incredibly popularity of web apps in buisinesses. Its far cheaper to:
* pay somebody to write a web app
* maintain a web app
* deploy a web app
My company has switched to focus almost exclusively on web apps because of these three reasons. Writing desktop applications is too expensive and are too difficult to update.
Furthermore, web apps require fewer resources to run. You can buy cheaper computers for your employees, and gain a cost savings with a slightly larger box, or farm of boxes on the backend.
Its much easier and cheaper to maintain a farm of servers because nobody is installing 15 toolbars, spyware and screen savers on the servers.
How many people host there own website? Software not hardware is the future. Remote hosting of apps is a natural progression. Think of the possibilities. The limits of memory are greatly reduced. I see us all owning dumb boxes (television?) and accessing information not dissimilar to current cable.
It would solve but not eliminate an awful lot of copyright problems.
I am afraid I must disagree as well , I have implemented Oracle, Sap and full office suites as well as multiple other apps web based across organisations from normal user based desktop through to outlying offices and on the road staff
The advantages are fantastic re: support and roll out if you get the backend correct ( can be hard and requires some effort )
with most corporate networks running 100 switched networks and vpn and broadband , you gain so much possibly my only issue with it is Bill Gates saw the advantages 5 years ago and now once again has major controls over the backend ,
I do not believe Companies will continue to buy bigger and fatter PC's long term with associated costs of purchase and implementation and ongoing support
I focus on web based development as well. There are too many advantages to it.
There are some programs that will not work web based like photo editing software and games. But I don't plan on bringing out the next big game or competing with PhotoShop. Basically all the systems that I have worked on have been business systems, payroll, project management etc.
I just did an update to one of these systems a few weeks ago too.
Less then 1 hour and everyone was using the updated system.
I seem to recall some big-wig (possibly from Sun) thought much the same thing about web apps six or seven years ago and invested money in dumb-terminal technology. The fact that I cannot remember his name speaks volumes about the success of his enterprise.
I do not deny that there are advantages to web apps but for most people, the disadvantages will outweigh the advantages by an enormous margin.
You can buy massively powerful desktop computers for a few hundred dollars. Laptop prices will be down well below 500 dollars soon. Why throw away all that CPU power for a system that will always be slower and intrinically less reliable. In a managed, predictable, corporate environment, there may be reasons to do so, but for home users - no chance.
|You can buy massively powerful desktop computers for a few hundred dollars. |
Yeah but that is the problem. Who the heck wants to upgrade their computer every year.
So sorry Kaled but we are on the edge of a sea change and the difference will be the same as a propeller plane and going to the moon.
Kaled raises some valid points here, guys ... going to a dumb terminal mode is a step backward for most applications. It does have its place, but also keep in mind that most of the businesses that would benefit from a dumb terminal environment are already using it and have been for years.
In the typical office, different employees have to run different types of programs. If an employee's system goes down, they can't work until I.T. fixes it. In a dumb terminal environment, if a server goes down no one can work. Years ago one of my first jobs was working in a call center where our desktops were basically dumb terminals. Due to a bug in the interface, I could paste an apostrophe into the notes field during a call and if I saved it would crash the system. You couldn't TYPE an Apostrophe into the notes field, but if you typed your notes in another window you could still PASTE into the notes field. Anyway, no calls came in no one could do any work. The server took 30 minutes to fully load. Many of us knew about this bug and would use it when we felt like we needed a break. I know this is an "extreme" situation, but it's a very real one.
By the way, I DO think a dumb terminal solution is the best for a call center since you have cube farms of drone employees all doing the same thing. In the typical small business though, your employees tend to have different roles and no two employees have the exact same job so it only makes sense for their computers to contain different software and possibly even have different processing capabilities. A shared STORAGE device is a great idea, but having the applications actually run from a remote server isn't. My assistant rarely works outside of the MS Office suite, but my programmers and graphic designer would be severely hindered by a dumb terminal-style environment.
Retail Stores, on the other hand, are a great place for dumb terminals. Cash registers, Telzons (the little hand scanners stockers walk around the store with), "find the price" scanners in the isles, etc. Those don't need any real processing power, they just send and receive data from a main server anyway, so they are technically dumb terminals by design. I would never want to have to operate a retail store where it WASN'T set up that way.
Last but not least are home users - and I don't mean guys like us with 5 or more networked computers at home, I'm talking about people who have one computer. What advantage could they POSSIBLY have by going to a dumb terminal mode for their applications? They would need to be connected to the Internet to type a friggin Word document. Lame. :)
|going to a dumb terminal mode is a step backward for most applications. |
I agree completely but this assumes static software. Think how inaccessible computers were to most people until windows and the advent of the internet.
Think about how archaic the old bbs system was. Calling all of those differant phone numbers. No we are using an antiquated system. We just donít know it yet.
There is a very big difference between web-based intranet systems that can be hermetically sealed and presumed secure and such systems accessed over the internet.
The first works well and if parallel mirrored capability is built-in (rare, I know) reasonably proof against downtime.
The second is the open can of worms (and other bugs) we all know and love to hate. Until and unless security and privacy issues are addressed (and that is not on the current horizon) internet web-based software systems are dead on arrival.
Cool, some things come and go like tides.
"Microsoft is misguided; the future belongs to the network, not to the OS. (..) The thing that I think is going to be really challenging for [the Microsoft] approach is, customers don't really like upgrading their operating system all that often."
Bob Lisbonne, Netscape, August 1997
Corporate inttranets are probably the best place for hosted applications with only thin clients at the PC. I work in eneterprise application development and all our customers want web based applications. They do not want to have to install or maintain anything on their employees desktops. After all - who says what they're installing doesn't break other applications, or that it won't have security bugs, or the employees might go do something foolish like delete a file or change something that will stop the app from working right.
I don't think people will use Microsoft Word that is hosted at microsoft.com and you pay a monthly fee -- that's not going to happen. But HR systems, e-commerce stuff, etc will want better UI's and be centrally served. How nice would it be to have spreadsheet like widgets in an online inventory application?
The good news is that you can do this today with Mozilla [getfirefox.com] and XUL [mozilla.org]. XUL is Mozilla's XML dialetc for describing user interfaces, and in fact -- it's what they use to build their application. I know - "But you install Mozilla on your PC" -- well, the framework has to be there, but once that's done you're good to go. Take a look at the Mozilla Amazon Browser [mab.mozdev.org] which runs in your browser but uses thick client widgets. It looks and feels like any other desktop application.
Check out Rapid Application Development with Mozilla [mozillazine.org] by Nigel McFarlane. It's a great book, that is now available for free download!
Kaled is correct that this same thing was attempted by Sun a while back, but it was Java applets. The problem with Java applets was that Microsoft quickly had its own VM out that "broke" a lot of applets. Consumers also knew nothing of Java, VMs, Applets, etc. All they knew is that the website didn't work. Mozilla is better situated in that Mozilla has consumer recognition [with Netscape], the Mozilla web browser is available today [compared to Longhorn in 2006], and most importantly the Mozilla web browser is a more tangible application -- people know what a web browser is, and the installation is super-easy.
I'm hoping that more content authors pick up XUL and start writing awesome applications for people who visit their site with Mozilla. There is a lot of bandwith and time wasted with constantly re-painting browser windows as people navigate through "light weight" HTML user interfaces. A XUL site can make things drastically more efficient.
|I seem to recall some big-wig (possibly from Sun) thought much the same thing about web apps six or seven years ago and invested money in dumb-terminal technology. The fact that I cannot remember his name speaks volumes about the success of his enterprise. |
Sounds like Sun's Scott McNealy and Java (Applets). While they didn't live up to the hype (from their own company) Java and Java Applets do have their share of the market.
A year ago I would have agreed with you, Kaled. I am not a hardware or software expert, so I tend to gravitate to what is easy.
I was convinced to go web based with most of my business apps that were developed specifically for me.
A ton of money later, I am glad I went web-based. I have desktops that are used for development and convenience when in the office, I havent used my laptop in 3 months. Instead I now access all my business related apps with a wireless PDA. My clients see me work on the fly and wish they could do the same.
I dont see this as an all or nothing issue, but a gradual shift that will be dictated by increased productivity and ROI.
I think the future will be a mix, call it middleweight terminals :D Thin clients are nearly as expensive as a cheap desktop (last I checked), so it would be silly to buy something without a hard drive. I believe hosted apps and terminal services will be the future for client/server apps, accounting/retail management/etc, but there are a whole lot of other programs needed to run in most companies.
In 5 or 10 years we'll have the best of both worlds in most offices; 100ghz pcs that cost almost nothing connected to centrally managed services, but also running photoshop 18, outlook 2014, and firefox 2.2 ;)
|A year ago I would have agreed with you, Kaled. I am not a hardware or software expert, so I tend to gravitate to what is easy. I was convinced to go web based with most of my business apps that were developed specifically for me. |
Sounds like you are
|In a managed, predictable, corporate environment |
As I said, web apps may have a future in this area.
kaled, you are using the classical debating trick of presenting the argument as a dichotomy.
The fact is that it won't be a choice between a centrally managed thin client & stand alone fat clients. In due course the web will grow up to present a fat client based upon distributed web type architecture. In other words software publishers will be able to create feature rich clients and still use the internet as the central distribution mechanism. So, if I want Photoshop I'll be able to go to the photoshop website and enter the application area with my browser. It will then download whatever needs to be downloaded. I will then use the app as necessary, save my work either on their fully backed up servers or on my machine locally.
the success of hotmail & salesforce.com type services suggest this is happening without the need for new web standards.
my recent experience with trying to write a web based wireless network planning tool suggest that we need a much richer experience at the browser end. I hope the w3c is working on this though I am not aware of anything. we don't want to let microsoft gain control of this technology.
Both software and telecoms technologies are approaching the point where mainstream web apps might be feesible.
If anyone believes they can do so, work down the list of software classes I listed above (include paint and accountacy software and anything else I forgot) and detail the pros and cons of standard apps and web apps. Next, show me some bandwidth and disk space calculations that demonstrate the capacity of web apps to go mainstream.
Why should I use a computer that has the power of ten-year old mainframe as an advanced graphics terminal? This question wil be just as relevant ten years from now as it is today.
|My recent experience with trying to write a web based wireless network planning tool suggest that we need a much richer experience at the browser end. I hope the w3c is working on this though I am not aware of anything. we don't want to let microsoft gain control of this technology. |
To what degree are these relevant and not for what you were trying to do?
but kaled most people are already using very fast machines as just advanced graphics terminals. i'm sitting here using a dual processor G4 powermac to write this after all.
which app do i use most on my mac? the safari web browser. what is a web browser if it isn't just an advanced graphical terminal?
we can have the best of both worlds. central manageability with a rich user experience.
I don't see disk space (which is dirt cheap and getting cheaper) nor bandwidth as being barriers to this either. At work i've got a half meg ADSL line & at home a quarter meg ADSL line. In a recent Byte article called "Reykjavik: Fiber to Every Home" there was outlined the 'next level' internet experience. 100mbps to each home. more than enough for very fast internet plus video on demand etc.
now, how long is that going to take to get here? who knows but it is coming, eventually. there's plenty of juice left in DSL yet. a 2 meg line would certainly speed my downloads.
The Guardian wrote:
|Gmail (..) makes a lot more sense when you think of it as the first in a line of major web applications built to replace desktop programs. |
Gmail is a perfect example why web applications would not replace desktop programs:
* I don't want to completely rely on another company for storing my - or my company's - mails.
* While the Gmail GUI is really very handy compared to other web mail services, it is a nightmare in comparison to Outlook or Thunderbird.
* Using a desktop application, it is my decision - or that of my company - what version of a software I'm running. If Outlook 2003 doesn't please my, I can continue using 2000. And if I'm not pleased with Outlook anymore, I can even switch to Thunderbird. It may be some effort to export the mails from Outlook, but I can do it. With Gmail, I'm stuck there and have no chance to switch to Hotmail.
Of course there are a lot of possibilities where web based apps are very useful and in my company we are using several of them. But most of them are management tools for which a simple GUI is sufficient and where it's vital to have a centralized data repository.
|No one would need to keep buying Windows, or upgrading Office if all they had to do was pay Google a monthly stipend for effectively unlimited storage, guaranteed backup and an installation or upgrade process consisting of typing in a URL. |
I really would love to pay Microsoft given this alternative! Now I'm paying Microsoft for their software and that's it. If I'm not happy with a new version, I can use the old one as long as I like. Or I can switch to MacOS or Debian.
Going the "Google way" would mean I have to pay on and on. And once I stop paying, all my valuable data is still on their server. If I'm lucky, I can download it, but I have no control if it's deleted by Google - or used for whatever they may think appropriate.
Web based applicactions are taking away, desktop apps are ensuring control - as long as data formats are open. This is, what we should really worry about.
Good luck to you on your views , but if you have ever tried to roll out a new version of oracle client across 15,000 PC's running any numer of varients of S/W and Hardware then you would understand it is not the speed and power of what users have but the issues with desktop management and speed of deployment, I do not think dumb terminals are the answer they are currently expensive and poor value for money , the changes are driven by the need for greater flexibillity and cost of ownership of the desktop the cost of a PC is about 10% - 15% of the cost of ownership in a corporate environment , so thin client web based apps will continue to grow , the home user is a different animal but over time it may be cheaper and more efficient for home users to access certain apps on a cost per usage model i.e. latest version of wordprocessing etc.
Great discussion! This is a topic that I've been fascinated with for years. Years ago (about the time Windows 95 settled in) I read a book by Bill Gates about what he saw as the future of computing. He believed that within a short amount of time, desktop computers would all become dumb terminals and simply be an extension of the Internet. Local hard drives would only be used to store documents and not applications. This was accompanied by the MS move away from app-centric computing (start your app, load your document) and a move toward document-centric computing (click on a document, and the best application available will open it). This is also when the "My Documents" folder became a part of the Win OS install (Win98).
It was an hotly-debated concept then and judging from this thread, it still is. I have since seen dozens of books/articles explaining why Bill was wrong. I even recall something that resembled Bill saying himself that he had 'mis-judged' the public's desire to keep their independent mega-horsepower desktop computers.
The debate reminds me of the joke that my wife makes every time I buy a new, faster computer: After I describe to her how much faster the new computer is than the old ones that make up our network, she asks me what I'm going to be doing with all the free time that the faster computer allows for.
A more serious comparison can be made from the days when I was cutting video on some of the first 'non-linear' video editors. Back then, editing digital video required that after the edit decisions were made, the computer was left to 'render' the final digital video. This rendering often took over half a day to complete, and hopefully the PC wouldn't crash before it had completed the rendering. With those applications, a faster computer could actually trim hours of time from the process. How much time did it actually save when completing a video project? None, because the time that was saved was then spent trying different effects and improving the overall editing to make a better video. The big improvement was that a faster machine allowed the director/producer to have more options at his or her disposal before the video had to be completed.
I think the same principle applies to the desktop app/web app debate. There are some great new development tools available for thin client/web app development, and in certain situations the benefits are very compelling. But the same can be said for fat client/desktop apps. What has really improved is the scope or range of available options. This allows for greater specializations and catering to niche requirements.
The answer is that neither fat client or thin client are going away soon. Ultimately everyone wins, from the users to the developers to the companies that sell these apps. The over-used comparison that comes to mind is that we have "an abundance of colors on our palette from which our artists (developers) can choose from". Having some more non-Microsoft colors sure would be a nice addition to that palette, though.
I'd wondered why MS had neglected IE.
Because, I believe, they see the long-term solution as one that abandons the browser altogether, so that the browser does not become the OS but the OS becomes the browser.
Exactly my words [webmasterworld.com] (msg #:7) ergophobe.
The concept of "browsers" is (or should be) antiquated by now. We have web clients in stead - browsers, PDAs, cell phones, newsreaders, iTV, IM's and PIM's, office applications, and so on (even modern HI-FI equipment). All we need is (XML based) data exchange using whatever web protocol they like the best.
Also, from Kaleds 9 point list, it seems to me that all 9 are now web enabled. I do, in fact, have (real www-dot-com) web pages that i only access through a spreadsheet, just to mention one eample.
>> You can buy massively powerful desktop computers for a few hundred dollars
As for "thin clients" i agree with Kaled: That's simply not the scenario for everyday office computing. These "thin clients" are the PDAs and cell phones, they're not "thin clients" as envisioned in the late nineties.
However, all the "fat apps" that need to run on "fat clients" are increasingly web aware which is something else than web based.
|Because, I believe, they see the long-term solution as one that abandons the browser altogether, so that the browser does not become the OS but the OS becomes the browser. |
Absolutely. And I wouldn't put it past MS to make it so that either no other browser can be launched within Windows' OS or any that are will be practically unusuable:
Launching another browser + typing the URL + hitting enter = IE automatically popping up with the URL you just typed into Firefox/Opera/whatever denying you the choice (unless you want to close IE everytime you try to surf with something else).
The power that MS has over the average user's computer is phenomenally alarming. I don't expect this to do anything but increase with each release and update of MS-deployed operating systems.
Bill Gates got his start by selling an OS that, either the company he was selling it to bought it for his price, by his terms - or all they had was a piece of metal that would turn on and give you a blank screen. While, now there are other options than Windows for the modern PC - there are none that support the vast array of software (and OS specific hardware) available for the OS.
He then made his fortune off of selling software that he didn't initially exist until he had the money in hand to buy it off of someone else.
These are long-time staples in the Microsoft business model, and proof of that legacy continuing to this day are the dependencies of every Windows user who has invested their time and money on buying into and personalizing settings within windows compatible software as well as the perpetual Windows updates that get dished out by the fistful each month.
| This 92 message thread spans 4 pages: 92 (  2 3 4 ) > > |