Re: .NET - end user reality check

Jim Slade wrote:

> Sit down and ask from a technical perspective point does .NET serve?

It actually serves a number of purposes.

> It comes down to NOTHING. It is yet another Marketecture in the long
> history of Marketectures in this industry.
> The virtual machine concept is not knew. Anyone remember UCSD
> Pascal????

Or IBM mainframes.

> The problem with run time interpreted code is "What happens when
> there is a bug in the interpreter?"

*It's NOT INTERPRETED* The JITter emits real, honest to goodness,
machine language, which gets executed by your processor of choice,
which might be buggy, too. Remember the '486 recall? How about the
Pentium recall? How about the Pentium CPU bug that would lock up NT
requiring a hard reset that didn't get a recall? Fortunately, with the
P3 and P4, Intel introduced the capability to patch the CPU on boot.

> The answer is you're screwed. At
> least when your compiler has a bug you can rewrite the code to avoid
> the problem or maybe change your compile switches.

Yeah, but how do you know that my CPU will do what yours does?

> At least with Java there is some pretense at multi-platform
> capability. .NET is purely for the Kool-Aid drinking morons who do
> whatever M$ tells them to do.

I can see several nice things coming out of .net.
1) Improved security/reliability. One of the points that the Multics
team made about how Multics was more secure than other OSs was that
ther was hardware support for preventing a lot of errors that you see
in modern environments. Windows and Linux both target multiple CPU
families, so they don't do much to help you out. (For example, Multics
caught buffer overruns at the CPU level.) Both .net and Java try to
handle this better.
2) It provides a better interop model than COM. Yes, there's still the
possibility of DLL hell via multiple frameworks. It's still a whole lot
better than having every little control that you use modifying the
global system state so it can be used. ARGH.
3) ASP.Net's actually fairly good. While it's not as nice as doing a
desktop app in Delphi, it's not that bad. It's certainly better than
ASP.
4) It's put pressure on other groups to move forward. Without .net, I
doubt that Java would've gotten generics. Both Java and .net have
pushed the C++ community to improve their libraries. (For that matter,
it's doubtful that we'd have .net if it weren't for Java.)

--
Mike Swaim swaim@xxxxxxxxxx at home | Quote: "Boingie"^4 Y,W & D
MD Anderson Dept. of Biostatistics & Applied Mathematics
mpswaim@xxxxxxxxxxxxxx or mswaim@xxxxxxxxxxxxxxxxxx at work
ICBM: 29.763N 95.363W|Disclaimer: Yeah, like I speak for MD Anderson.
.

Relevant Pages

  • Re: HMAC and timing
    ... Benchmarking in Java is a bit tricky, ... If you want a measure with about 1% accuracy, you have to hash ... kernel and its cache subsystem. ... the CPU cache. ...
    (sci.crypt)
  • Re: Java is slow!
    ... Do you really think Java event dispatcher is where all those CPU cycles ... AMD or Intel wins? ... The current AMD design seems to be simpler, so a compiler writer ...
    (comp.lang.java.advocacy)
  • Re: Stupidest game of the month. 11 move Mate.
    ... Peter Osterlund writes: ... 100% of one CPU, so whatever you did it appears that it works. ... The problem that the web page says that java is not installed is also ... So those problems were not caused by GetClub chess. ...
    (rec.games.chess.computer)
  • Re: Game Company- Java Server Thread Priority
    ... I know that there is CPU Throttling in Windows 2003... ... > for the application that was running on the old server. ... This machine has one job, run this Java program, and ...
    (comp.lang.java.programmer)
  • Re: Game Company- Java Server Thread Priority
    ... I know that there is CPU Throttling in Windows 2003... ... > for the application that was running on the old server. ... This machine has one job, run this Java program, and ...
    (comp.lang.java.help)