suggests that its use is web programming first, database programming
second
and windows application programming a distant third with very little else.
Michael said:Definitely not. I would say Windows application programming first, web
second, and database third. It is not at all a special-purpose or
restricted language.
Jon said:My hypothesis is that C#'s use is very domain specific whereas ML is not. I
know that ML is not because I've been involved with such a broad range of
applications written in ML. I don't know about C# but everything I find
suggests that its use is web programming first, database programming second
and windows application programming a distant third with very little else.
Perhaps this is an unfair comparison because C# is a new language but ML is
a family of languages with a legacy dating back over 30 years. Even OCaml
is 11 years old now.
Jon Harrop said:Right, that's just it. C# is a completely general purpose language yet it
is
rarely used in bioinformatics, biology, physics, chemistry and
engineering.
Jon said:Well, that encompasses a massive amount of the coding landscape. I
don't think of that as being "very domain specific". Furthermore, just
because those are its *primary* uses doesn't stop it from being used
elsewhere.
Would you say that C is very domain specific as well?
Given how you seem to think that C# is "not in widespread use" I don't
see how Microsoft program managers are justifying the amount of time
and effort they've put into it.
I'd have expected it to become more mainstream in that time then,
wouldn't you?
The trick would be to do that without introducing more complexity than
it's worth for the vast majority of developers who aren't in that
field.
I notice that you didn't answer my question, however. I'll repeat it:
Are you seriously contending that there is more ML development going on
than C# development?
Michael said:I am using it in computational linguistics research right now.
www.ai.uga.edu/caspr
By "is rarely used" do you mean "is not yet entrenched and traditional"?
C# just hasn't been around that long, and laboratory science is very
tradition-bound.
Is it bad just because it's new?
Jon Harrop said:Sure. Perhaps "domain specific" is not the best phrase but I can't think of
a better one.
C was used very broadly 15 years ago but it has been largely superceded by
other languages (primarily C++) in many areas now. In scientific computing,
many more people use C++ than C. I expect the same is true of completely
unrelated domains like Windows application development. Now, I would say
that C is primarily a glue language used in foreign function interfaces and
a systems language used to write OSs and drivers.
Java is probably the most widely used language that I am aware of. Every
application area that I come to, I find people are already using Java for
it.
You can see how narrow the use of C# is by looking at Google hits. Search
for a language and a technical keyword. For example, bioinformaticians
might search for "suffix tree" and their favourite language:
As you can see, the market share for C# among technical users is tiny
compared to the market share for C# among web programmers.
Microsoft simply select the most lucrative subjects and work to dominate
them. They have absolutely no interest in working to get C# used in a wide
variety of different subjects.
Mainstream like Sudoku or mainstream like calculus?
I think they just need to add some basic libraries, like complex numbers.
No, of course not.
Jon said:And yet in all but one of your results, C# comes above OCaml, which you
are claiming *is* widely used in these areas.
Mainstream like "there are vast hordes of developers around, and it's
commonly used in the biggest development area - business".
Okay. Can you explain what on earth you meant when you said:
<quote>
While C# looks common, its use in industry extends little beyond web
programming. So I would say that use of ML is already far more
widespread and F# can only broaden its use.
</quote>
To me, C# is more widespread because there's vastly more C# development
going on than ML development.
I'm sure it's used in just about every industry on the planet, to a
greater or lesser extent.
programmer for several years now and I don't know a single C# programmer.
I was only referring to the extent. C# is a dominant force in web
programming, database programming and windows application programming but
nothing else. Java seems to dominate everything.
technique because lots of people do Sudoku puzzles but calculus is an
obscure technique because only educated people know and use it.
Jon Harrop said:I was referring to the distribution of domains rather than the size of the
largest domain. Most general purpose languages are used uniformly across a
wide range of subjects but C# has a very skewed distribution with the vast
majority of its users in a small number of application domains.
Consequently, C# appears near the top for some results (web programming) but
the bottom for others (technical computing).
Vast hordes of developers around where though? I've been an industral
programmer for several years now and I don't know a single C# programmer.
To me, that is like saying that Sudoku is a widespread mathematical
technique because lots of people do Sudoku puzzles but calculus is an
obscure technique because only educated people know and use it.
I was only referring to the extent. C# is a dominant force in web
programming, database programming and windows application programming but
nothing else.
Java seems to dominate everything.
Marc said:But since you youself say - you don't know a single C# programmer, so
perhaps your analysis is itself flawed by being unfamiliar with where
C# *is* used?
To move this forward, can I suggest that rather than saying "nothing
else", you propose some actual areas where you think C# isn't used. I
suspect that you might be surprised to find that it is, in fact, used
there - but that in these smaller areas, the visibility is lower,
exactly as a consequence of these areas *being* smaller. I will
exclude areas like low-level device-drivers, because due to the
platform this is not a reasonable development area for .NET (nor
perhaps java?).
Jon said:Well, that's not the common use of the phrase "widely used" or the word
"widespread".
I don't see why that's a problem though.
Ask recruitment agencies how many C# developers they've got on their
books, then ask how many ML developers they've got on their books. Ask
how many vacancies they've got for each. I'm sure you don't really
believe there are more ML developers than C# developers.
As for why you don't know any, I suspect that's due to the particular
areas of computing you've been working in - which are relatively
uncommon, compared with the *extremely* common areas of n-tier
development for business purposes.
Sudoku isn't a mathematical technique - but it's a more widespread
mathematical exercise than calculus is, yes.
Yes, and we all know how insignificant those three areas are, of
course.
Does Java dominate games programming?
Device drivers? OS development?
RDBMS development (not as a plugin, but as the engine)? No. Native
languages still have the edge in all those fields.
I suspect that there are three major reasons why C# isn't more widely
used in scientific computing:
1) There aren't as many scientific libraries available for .NET as for
some other platforms. This is partly due to age.
2) Scientific computing is often performed on non-Windows boxes. While
Mono is a great project, I suspect it isn't fully trusted yet, and may
not be as well optimised as .NET itself.
3) Scientific computing often involves long projects, which may have
started (in one form or another) long before .NET came on the scene.
Java has a 7 year advantage over .NET.
Notice how none of these reasons are specific to C# - they're about the
.NET platform.
Jon Harrop said:Yes. I tried using "not domain specific" in an attempt to avoid confusion.
Can you think of a better phrase?
I think it is an interesting result, not a problem.
My interest is really three fold:
. I'd like to break down market size (in $) by programming language.
. I'd like to know what languages scientists want tools for.
. I'd like to predict the future of the language landscape.
I think the first is much harder.
As you say, there are huge numbers of C# developers but their efficiency at
creating wealth is orders of magnitude lower than that of an ML developer.
Absolutely but without knowing how many C# developers can be automated by a
single ML developer that information has no practical use (unless you're in
job advertising).
Absolutely. I'm here to sell spades rather than dig for gold.
Right. Sudoku is widespread because it is heavily advertised and not because
it is valuable. There is money in the Sudoku market but nothing like as
much as in calculus.
I'm glad that web programmers consider themselves to be significant. I'm
sure they do a lot of good work, like FaceBook, but I know very little of
it.
Java is unquestionably a dominant force in games programming.
Well, that's like having a democratic vote with a single candidate.
Jon said:For example, I believe C# is not a popular language for:
. Non-Windows development.
. Scientific computing.
. Finance.
. Embedded systems.
Jon said:I was referring to the distribution of domains rather than the size of the
largest domain. Most general purpose languages are used uniformly across a
wide range of subjects but C# has a very skewed distribution with the vast
majority of its users in a small number of application domains.
Vast hordes of developers around where though? I've been an industral
programmer for several years now and I don't know a single C# programmer.
To me, that is like saying that Sudoku is a widespread mathematical
technique because lots of people do Sudoku puzzles but calculus is an
obscure technique because only educated people know and use it.
I was only referring to the extent. C# is a dominant force in web
programming, database programming and windows application programming but
nothing else. Java seems to dominate everything.
Jon said:I also suspect that the first would show a far greater dollar amount
for C# than for ML, however.
You state that as if it's a fact. Care to back it up with statistics?
Given that you're an F# evangelist, you can hardly be viewed as
unbiased.
No, the information has practical use in terms of knowing how
widespread use of C# is - using "widespread" in the common meaning of
the term rather than yours.
And the fact that you're on a C# group specifically to promote a
different language is rather interesting too. Most people here are
interested in solving C# questions rather than changing languages. Do
you regularly go to parties and say, "Hey, there's a better party
across the road"?
But the point is that it's widespread - and the introduction of
comparisons between Sudoku and calculus was to illustrate the meaning
of the term "widespread" as far as most people understand it.
Now, are you going to finally admit that in the *normal* understanding
of the word "widespread", C# is far more widespread than ML?
There are many, many systems (whether web based or not) which are
internal to companies, or only exposed within commercial contracts, but
which are nevertheless productive.
Care to stand up in a games development conference and claim that with
a straight face? When talking about games on mobile phones you could be
right, but outside that very specific market Java has very little
presence.
Care to name a single top 10 game for either the PC, Wii, XBox 360,
PS3, GameCube, PS2, XBox, DS, PSP or GBA which was primarily Java-
based?
I don't see your analogy at all - and importantly, they are all reasons
which will count against F# as well as C#. The same changes which might
drive F# adoption for scientific computing will also make C# adoption
easier.
Arne said:Fortran was strongly focused around scientific computing.
Cobol was focused around business computing.
True.
It is more the norm than the exception to have strong areas.
That only proves that you are utterly out of touch with what is going
on in the industry.
????
You made a quantitative assessment.
And when called to explain you come with a qualitative assessment.
Persistence and UI are orthogonal characteristics.
Besides I think that UI and persistence is very widely used.
Including in technical computing.
creating wealth is orders of magnitude lower than that of an ML developer.
For some reason, the big players are only just realising this. ....
how many C# developers can be automated by a single ML developer
Try using the complex number implementation from the F# stdlib in C#, or the
Extreme Optimization library's C# API from F#.
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.