C
Chris Mullins [MVP - C#]
But, personally, I would never
choose .NET for a horizontal, mass distributed application. That
would have to be something more standalone - like C++, Delphi, or
maybe PowerBasic (personally, I'd probably choose C++).
I agree. Completly.
Nobody? Really. Do you have stats on that.
In relative terms, very few people do (I don't have numbers, or I would
provide them). It's not installed by default with any O/S or major product,
nor is it pushed by Windows Update.
To have it installed the user either had to manually install it (via an
optional Windows Update), or be running a program that's already installed
it. Both are unlikley scenarios in the general case. Especially for the
non-tech savy users, that small "cute" apps typically target.
Why? Because you install a dependency? Please.
Because, the installation process for .Net sometimes takes over an hour. It
also frequently fails to install at all. In the coffee-house scenario, where
users are trying to download and install your app via the free wireless,
it's not usable.
Also, for some reason I don't understand, .Net is signifigantly slower to
install via a bootstrapper than via the stand-alone MSI.
The end result is a bunch of people who were willing to install the app now
think it sucks. They don't know it was the .Net framework that had issues,
they just know they tried to install SoapBox, it took an hour, and ultimatly
failed. This makes up look bad, and our software takes 100% of the blame.
They have pushed via windows update - but as an optional update.
Optional updates don't really do anyone any good. I would love to see the MS
numbers on the this. I would really like to see the reason behind why they
havnen't pushed it wider. I suspect it's due to the high failure rate on
install - especially for the .Net 3.0 stuff.
If it's a problem for you, then I respectfully have to say your using
the wrong toolset.
I agree with you. It's sad, but .Net isn't a suitable toolset for building
widely deployed desktop applications.
The problem for me was a failure to check my/our assumptions. We assumed the
..Net Framework was widely deployed - via Office / Windows update / IE7 / XP
SP2. This turned out to be incorrect, and has caused much pain. Nobody to
blame there by oursleves for such a poor assumption.