Please. How much of the Google search stack does the “web application” part represent? Almost none. All of the heavy lifting is done on the servers, and it’s hard, performance-critical, CS-heavy programming of exactly the sort that Braude is talking about. Unless you’re somehow including that as “web programming” because it’s accessible from the web, you’re way off base here.
Obvious troll is obvious. And astroturfing.
In all seriousness, you guys don’t really buy that these are the true beliefs of the author, do you? Just another guy trying to drive traffic to his blog by saying something ridiculous and riding the wave of attention, good or otherwise, to more hits and notoriety.
Stop paying attention and he will go away.
It’s ok. Bashing Microsoft employees is still cool. It will soon be trendy to bash Google, since they are getting big now.
I think a very, very big part of the future programming is small applets for small devices, kind of the eBay or Facebook applications for the iPhone.
Silverlight 3 will run on Windows Mobile 7 natively. That will bring thousands of applications into the next generation of Windows Mobile phones.
Is an email client a niche application? Because I don’t know anyone who prefers using Gmail over Outlook or Thunderbird. Actually, I know one guy, but he’s a Mac user so his opinion doesn’t count 
Is a chat application like Skype a niche app? Is a file browser a niche app?
@codinghorror, you rock!!
Oh also…
Bloss’s Law: People who name their own laws and quote themselves on it are usually egotistical arrogant bastards.
Also, most web apps that replace desktop apps have the features we’ve seen in desktop apps 10+ years ago. See that 3D engine in JavaScript? Looks like Wolfenstein.
Must we really set up a desktop v. web false dilemma? Obviously they are both going to be around for some time yet, Web 3.0 be damned.
It’s moronic to assert that all web developers are bad programmers; it’s moronic to assert that all desktop programmers are dinosaurs. There are good programmers and bad dinosaurs in both; there are programs that work and there are those that don’t; there are things that people use and there are things that don’t.
“Web v. desktop” are not useful categories when talking about quality of programmers or whether programs actually get used.
Must we really set up a desktop v. web false dilemma? Obviously they are both going to be around for some time yet, Web 3.0 be damned.
It’s moronic to assert that all web developers are bad programmers; it’s moronic to assert that all desktop programmers are dinosaurs. There are good programmers and bad dinosaurs in both; there are programs that work and there are those that don’t; there are things that people use and there are things that don’t.
“Web v. desktop” are not useful categories when talking about quality of programmers or whether programs actually get used.
It’s not so much a new transition from desktop to web (which to some extent will probably merge anyway), as it is the continuation of the ever-present transition from harder, faster languages to easier, slower languages. The exponential increase in computing power means that runtime speed becomes ever less important than the time it takes to write the code, which has led from a transition from assembler->C/Cobol->C+±>VB/Java/C#->PHP/Python/Ruby.
In areas where speed of execution is particularly important (e.g. embedded systems, OSs, databases) the transition is slower, but even then C++ and Java are increasingly more likely than assembler and C.
Hand in hand with this, easier languages which allow code to be written more quickly mean less need for UML diagrams and designing everything up front and more ability to just figure things out as a project progresses.
But if the actual programming gets easier, that allows people to be clever in other ways rather than spending all their time writing UML diagrams and tracing memory leaks. MapReduce? Specializing compilers? Artificial intelligence? Machine learning? Computational statistics? Facial recognition? Photo realistic rendering? Any of these are a hundred times more challenging and interesting than learning how to squeeze an extra 1% performance out of some Java code.
Michael Braude’s superiority complex is evident by his need to define technical complexity as a measure of one’s worthiness as a developer. If it’s not complex, it must be crap.
Business developers excel at their jobs because they have the personal, artistic, and technical skills that other developers couldn’t conceive as being important for the job. They migrate to this environment because they are good at it. They don’t use C++ for the web, because it’s the wrong tool for the job, not because they don’t understand it.
Many C++ guru’s would fail if they had to interface with an actual human.
Jeffs ‘reaction’ is justified… albeit a bit childish and actual made me rolling on the floor laughing (ROFL)…
The Web is BUILT on top of the following:
- hardware circuits
- software DRIVERS
- Machine Manguage
- ASM
- C (or other)
- PHP/ASP
There are MANY layers to programming… and yes I thought web programming was going to be a BREEZE when I was coming from C++ game programming.
But, it’s all the same logic. If this than this… Interfaces and Abstract classes verse SQL and PHP (which PHP actually now supports Interface Classes and all that Jazz)
Still Jeff’s react was ‘only human’. Albeit a bit ignorant and childish…he IS a web programmer.
Couldn’t agree more, except the “niches” bit of the argument. There are a lot of vendors pushing web solutions for business software, but there is no compelling case to make business software web-only. Web-enabled, yes. But web only is a very different pile of hurt.
Web apps are definitely popular, but there will always be a place for desktop application. IMO, these are moving to be web delivered and will continue to do so (Silverlight Out-of-Browser and AIR). But they are still desktop applications. And even this will not cover 100% of desktop application (embedded apps, batch processes, etc).
I do like ready Jeff’s Blog, but in this instance both blogger are showing a little bit of arrogance. Plus, IMO whenever you get personal and start making personal attacks you lose all credibility. Both bloggers should have had some restraint in their response. I’d like to see a intelligent debate on this issue between the 2 original bloggers. Maybe Carl Franklin can do a podcast covering both sides “Web vs Desktop” and moderate the discussion.
Carlos: The following statement show how much you do not know. “No serious developer (web or not) uses .NET . If you can’t understand why, you’re not developing serious software/applications.” The .Net framework is one of the best frameworks to write web and desktop applications and your comment is one the same level as those who argue C# is better than VB.
This Michael Braude guy is spot on: at least, that’s the very reason I code in PHP, and I spent 5 years getting my degree in Computer Science…
Some days, I wish I weren’t so scared about abstract classes…
This person comments that “Javascript is a functional / ‘prototyped’ language, not an OO language” while constantly emphasizing how smart he is. Let’s just leave him alone. Beating up mentally challenged people is not nice.
This article makes about as much sense as a chef saying “I like bread. All my friends like bread. Therefore, in the future, all food will be bread.”
If there is one thing that the history of computing should tell you, it is that you have NO IDEA what will happen 5 years down the road. All we can do is forcast general trends. Sure, the web is becoming more popular. But the desktop isn’t going away any time soon. And the web and the desktop are only part of the programming world. Just because you have a hammer in your hand doesn’t mean every problem is a nail.
Don’t get me wrong, as I love StackOverflow and even your blog, but in this topic you’re just plain wrong.
I am Brazilian and at my university most of my colleagues tend to write webapps as a career, but that’s just not what the market is demanding. Embedded programming is booming. No, really booming. Booming to the point that far more CPUs are sold in other architectures than x86 ones. Now, those chips must be programmed. And the program, since is so close to hardware, don’t have a great reuse rate, which means, pretty custom programs for custom hardware. That simply will never have anything to do with web programming and certainly won’t be done in JavaScript.
On the other hand we have databases, OSs, rendering software and others that doesn’t make any sense to become web apps, and there is brilliant, perhaps the most brilliant, people working on those areas.
So while we move to the web in programs where sharing makes sense, I’ve never seen a web text editor that can do what Textmate can. Until then, I guess I’ll just continue to flip my bits here.
Atwood’s Law is wrong; it won’t be written in JavaScript… it’ll be written in JQuery. 
(Sorry to not be constructive, but I’m borrowing from an earlier quote in your blog that I still joke about…)