Re: Checking for Modification to a Set of Files

On Sat, 25 Mar 2006 19:47:05 +0000, Randy Howard wrote:

Ben Bacarisse wrote
(in article <pan.2006.03.25.18.36.08.510189@xxxxxxxxx>):

I was not suggesting you were partisan about platforms. You seemed to be,
like me, partisan about camelCasedAPIsFromRedmond and this was obscuring
the fact that .NET applications are not Windows-only.

I've yet to encounter or hear anyone claim that they actually
use .NET on anything other than Windows for production use.

Hmm... I am not sure what answer you are reserving by using the word
"production". The mono team lists various uses that look, to me, like
production applications, but I may have missed your point.

<snip>
I've seen zero effort out of MS to push .NET as a cross-platform
standard so far.

No, I am sure they have not. They merely want it (technically CLI and C#)
to be "a standard" and have picked the easiest route (ECMA). I did not
say they would push it as cross-platform or that they would not try to
subvert it to their own ends. I wanted to move the argument away from "it
is not standard" to /why/ it is not suitable for cross-platform
development (if, indeed, it is not).

Maybe someone here has tried and can testify that writing portable .NET
applications is a nightmare?

I have not seen the phrase "portable .NET applications" used here at
all. That's telling in and of itself.

Possibly. I'd rather someone jumped in with some examples or experience.
I *know* why MS's XML document format is not, technically, as open as
ODF, but I have no hard evidence about .NET.

Inferior is a question of point of view. Despite the common mantra of
the newly minted CS grads that everything in the universe is about
performance, in the real world, you need to understand that if it
isn't a bottleneck, then it it isn't a problem.

As someone who used to be partly responsible for minting those grads, I
think the performance obsession is almost unavoidable. I think it may
have more to do with the fact that most fresh CS grads are young men
than anything to do with what, or how, they are taught.

It might also have to do with the multiple semesters of 'big O' debating
in lieu of covering broader issues that seem to be ignored for the most
part.

To paraphrase Wilde: "Education is all very well, but it is worth
remembering that not everything worth knowing can be taught".
Universities teach what can be taught and should not be criticised for
that. In other fields (medicine, law, engineering...) students learn the
"broader issues" later by other means, and CS should not be any different.

When I have been on syllabus committees, employers always want the
students to know more about "real-world" programming/applications/systems/
design/research (it varies depending on the employer) but they can not
tell you how you might teach this. When you ask which of the core courses
on programming (usually at least two languages), algorithms, data
structures, basic maths, operating systems, computer architecture,
concurrency, software engineering, compilers, networks, distributed
systems, HCI, professional issues and, yes, computability and complexity
should go to make room they can not agree (I may have missed a few).

Looking at the "homework questions" of recent years, that trend
is getting continually worse, or entrance criteria for CS programs is
bottoming out, or both.

This is called getting older. Every generation thinks this, in every
subject area. Of course, in some generations and in some subjects it
might actually be true, but the perception of it follows inevitably from
gaining experience: "how could anyone not understand that?".

<snip>
and "it is not available for OS X" is probably false (depending on what
you want to be included in the term "available").

"differencies" notwithstanding, apparently.

Indeed. You get to decide if it suits your needs based on some facts
about it. Is that not a better situation to be in than assuming it is
Windows only?

BTW, I copied an example of using the FileSystemWatcher[1], compiled it
using mcs (mono's C# compiler) and ran it on my Linux box and it did,
indeed, work. I presume the code works on a Windows box. I wonder if
anyone with a OS X or BSD would try the other mono ports?

[1]
http://samples.gotdotnet.com/quickstart/util/srcview.aspx?path=/quickstart/howto/samples/Services/FileSystemWatcher/Watcher/Watcher.src&file=CS\Watcher.cs&font=3

--
Ben.
.

Relevant Pages

  • Re: Launch .NET application through notification or windows messaging
    ... I have two .NET applications, ... A writes records to a database, and then issues a message that its ... What's the best windows technology to use to do this? ... SQL Server 2005 and above, using SQL Server .NET CLR, Service Broker and SQL Server queues. ...
    (microsoft.public.dotnet.general)
  • Re: Events on a Timer
    ... There are a variety of flavors of .NET applications, console, traditional ... Windows, ... the platform itself (the ".NET" ...
    (microsoft.public.dotnet.framework.aspnet)
  • PC World: Windows Grip on Market Share Slips
    ... Windows' share typically falls on weekends and after work hours, as users surf from home computers, a larger percentage of which run Mac OS X than do work machines. ... Net Applications also noted a small boost in market share for the open-source Linux operating system, which grew from 0.71% in October to 0.83% last month. ...
    (microsoft.public.windows.vista.general)
  • Re: Windows Forms / Avalon and VCL.NET
    ... > VCL.NET does not replace Windows Forms. ... > classic VCL style) for us to write .NET applications and assemblies. ... libraries that implement GDI user ...
    (borland.public.delphi.non-technical)
  • Re: Windows Compilation Madness
    ... another long thread like this with all the problems windows users ... There are plenty of prevalent compilers, ... If the build-chain isn't predictable then it is ... a build environment on each and every Windows box what if the ...
    (comp.lang.ruby)