J
Jon Skeet [C# MVP]
Jon Harrop said:There's no point in speculating without any quantitative evidence.
You've not been reticent to speculate in the past.
Particularly if you are totally unaware of ML.
I'm not "totally unaware of ML". Why do you suppose I am?
Sure:
Windows once had an appalling reputation for being plagued by crippling
bugs. Microsoft did not even know where these bugs were in their operating
system. A team led by Gordon Mangione pioneered the use of automated error
throwback to Microsoft and discovered that 80% of the bugs were actually in
third-party driver code.
As a consequence, Microsoft stopped pouring money into hundreds of testers
trying to maintain their low-level code and built a group to perform static
verification of drivers using tools written in OCaml. They succeeded in
removing many of the most important bugs and the reliability of the whole
platform improved enormously as a consequence.
How many developers were displaced by the ML developer and how what is the
monetary value of stability to Microsoft?
That's a single point of data, about a single area (bug detection).
Yes, it's an important area in a large system, but it can't be assumed
to be representative of computing as a whole.
My job wouldn't exist if my statement weren't true.
Your job wouldn't exist if there weren't people who believe your
statement is true. Big difference.
What use?
Um, not sure what you mean. By the common definition of "widespread"
the information about how many C# developers there are vs how many ML
developers there are is indicative of the overall comparison.
Better than standing outside the C# party, making myself extra visible and
acting as a high-class spammer for scraps. Apparently MVPs are legal in
Australia. ;-)
I have absolutely no idea what you're referring to.
Actually I introduced Sudoku and calculus to this conversation to illustrate
the difference between common and valuable.
That certainly wasn't clear at the time.
On the contrary, I think calculus is widespread and valuable but Sudoku is
common but domain specific.
I'd agree that calculus is widespread and valuable, but I'd say that
Sudoku is *more* widespread.
Sure. XenSource do that and its written in C#. Well, the intermediate
language is C#. It is actually generated by OCaml code. ;-)
You seem to be completely ignoring my point.
You've changed your tune. You don't want to compare languages my monetary
value but you want to compare games by subsidy rather than units sold.
I'd be happy to compare languages by monetary value, but I suspect
you'll find that pretty tricky. However, if you're going to claim that
Java is unquestionably a dominant force in games programming, I'd
expect at least one very popular game to be primarily developed in
Java, wouldn't you?
Perhaps you could elaborate on how you believe it's such a dominant
force.
Try using the complex number implementation from the F# stdlib in C#, or the
Extreme Optimization library's C# API from F#.
Why don't you just save us the effort of investigating the libraries,
and explain how it's relevant to my point. How is F# (which runs on
..NET, right?) going to get round the "scientific computing is very
often not running on Windows" issue, for instance?