Editors note:  Pilky originally published this on his personal blog on April 8th.  We have republished it here with his permission.  We have edited it to a minimal degree for formatting and minor changes like capitalization and punctuation.

Paul Graham has written a good article on the “death” of Microsoft. He argues that Microsoft, while going nowhere soon, is a shell of its former self and just doesn’t scare people any more, partly because of the new kid on the block who also has lots of money: Google (for football fans out there, this is akin to Manchester United and Chelsea).

However, this isn’t the point of this post. There is one small thing that he mentions, almost in passing, in his article and it is something I strongly disagree with:

“Everyone can see the desktop is over. It now seems inevitable that applications will live on the web—not just email, but everything, right up to Photoshop.”

He links to a site called Snipshot, which is a very cool site, allowing you to edit pictures online. However, it is about as close to an online Photoshop as Blogger is to an online Word. There seems to be a slightly delusional section of web developers who seem to believe that in a few years time all of our applications and data will be online, while our computers run little more than a browser. Of course this is complete bull.

The most obvious reason for this not happening is network speed and latency. If we were to take local networking, with gigabit ethernet, this allows for around 125MB/second of data transfer. Sounds OK, larger than most office files. Internet is slower; even with broadband, many people are just getting to around 2MBit. This is less than 1MB/second. Now for something like email or word processing or spreadsheets these speeds are okay. For the most part you’re sending and receiving text. However, when you start moving above this you have problems.

Images that are several MB in size can be transferred over the network, which for a local gigabit ethernet is fine, but for the Internet just isn’t going to happen. Snipshot claims you can “edit big pictures” up to 10MB or 25 megapixels. Now for what it does, that’s great, but to compete with Photoshop that’s pitiful. When you need to be able to deal with several image files that are 100s of MB in size all at the same time, a web application just isn’t capable. Things get worse when we move up to video. For video you could be dealing with GB’s/second of data needing processing.

Latency also limits how these applications work. If I click a button in Photoshop the data has to travel within my computer before it gets to my screen. On a web app the data would have to go back to the server to be processed before returning to your desktop. Now if the server is in the US and you are in Europe, then you are talking a few 1000 miles vs a few centimetres. Data can travel fast, but it’s still limited by the speed of light within the material it’s flowing.

Of course there are other reasons. Many web APIs (I’d be willing to say all web APIs) are nowhere near the level of completeness that desktop APIs are at. There are things you can do with desktop APIs that just aren’t feasible on the web. Take something like CoreAnimation in Leopard; I would absolutely love to see a web version of it, but it just isn’t going to happen. This leads to another point, in that desktop APIs integrate with the OS, and in turn the computer, much more closely.

One of the most obvious reasons is simply the fact that nobody is ever on the internet all the time. What if I want to read an email I was sent yesterday, but my network is down? If I use a desktop email client that’s easy, but if I use a web client then I’m out of luck. Local copies of data are vital, and there is some data people would rather stayed local.

So why, with such obvious arguments against it, do people still take the position that the web will remove the desktop? Well, simply put, web developers rely on one thing on a desktop, a browser. This, and the increasingly powerful APIs available to them, makes them believe that all a user will ever need on a desktop is also just a browser. You see the case for desktop apps making web apps irrelevant nearly nowhere. The argument doesn’t exist.

This is likely because many desktop developers understand the role the internet has to play with their applications. They don’t have to write a single piece of code that does anything to do with the internet, but chances are there are software update systems, online help systems, websites to sell the software, activation systems, etc. Look at any piece of desktop software today and it is somehow linked to the internet. The reverse just isn’t true for all web apps.

So what is the future? Well, the future is that the web won’t win. Web apps aren’t there to compete with and wipe out desktop apps, they’re there to complement them and enhance them. The prime example of this is with email. If I’m at my computer I can just download my email into Mail.app. If I’m not I can log into my webmail to view it. If anything, this is the perfect setup for a web app/desktop app eco system. The desktop app is more powerful and the web app is more accessible. Anyone who thinks otherwise is just fooling themselves and building them up for severe disappointment down the line.