Re: .NET

> That is a problem with the code, not the GC. There are trade-offs in
> everything you do.

Sure.. but the discussion at hand was about performance and GC .. .anyway thats my interpretation of it :)

> If you choose portability and stability over performance
> the so be it. Would you rather have a really fast app that crashes every
> day? Me neither.

Nope... but its certainly possible to have a super fast app that never crashes... ask some of my kbmMW users :)
Its equally possible to have a super slow app in .Net/Java that crashes all the time.

The big difference from most native solutions is imo the abstraction towards languages for .Net and binary portability
for Java (and in theory also .Net although it seems not to be MS primary goal)..

> It sounds like you've worked on some large near real-time apps written in
> Java and .Net.

Near realtime is definable I guess... specially on Java platforms :)
But its often applications with a high level of abstractions, many users and relatively relaxed requirements for
response times.
Thats actually a funny thing I have discovered... when discussing native type development with mgmt, one set of rules
apply,
but when talking .Net or Java, another set of rules often shows up... for example regarding response time requirements.

> As chief architect, why did you choose a language that
> requires a VM or JIT-compilation for a near real-time application?

Most often the choise is made for you by mgmt. In big corporations there is an IT strategy that involves specific
products from specific vendors.
If you want to go out of the box, you will have a seriously hard time.
I have done it on occations but I can tell you its really one tough job... and as always... if one choose something out
of the valid tool box every eye is on you and one single small mistake you make will be scrutinized with extreme force.
If choosing the accepted tools you have much larger freedom to make mistakes or badly performing applications.

It may sound weird but thats how it works... anyway in just about every company Ive been attached to or employed in.

> If someone else chose the language, did you inform them that there would be
> performance issues?

Ofcourse.

> If so and they said "I don't care" then what's the
> problem?

That you still get a mediocre setup with a bad performance costing from 2-10 times the money another setup would have
cost.
I could ofcourse just dont care as it wasnt my decision in the first place... but it does hurt my professional pride to
deliver something
which im not satisfied with myself for a cost that is many times higher than it needed to be.
Not all the cost is ofcourse attributed to language or platform... often its just the way big companies work that makes
it cost a fortune.
However its usually considered normal to simply add people to a Java based project in the belief that then development
will speed up.
Delphi projects are usually developed by smaller groups that act more swift and with less beuracracy.

> When the server pauses periodically for GC just remind them that
> they didn't care about performance.

Mgmts are not known to volunteered acknowledge their own wrongful decisions.... :)


--
best regards
Kim Madsen
kbm@xxxxxxxxxxxxxxxxxxxxxxxxx
www.components4developers.com

The best components for the best developers
kbmMW - RAD Enterprise class n-tier application server framework
kbmMemTable - High performance memory table
kbmWABD - RAD web development
kbmX10 - RAD house automation


.

Relevant Pages

  • Re: How to make Forth interesting?
    ... talking about JAVA and talking about Forth and Java. ... "I think it is simply a case that Chuck designed a language to ... If a specialised task works a lot better with some new primitives added ... might be multiple million dollar a year licenses with the cost of the ...
    (comp.lang.forth)
  • Re: Joel Spolsky on languages for web programming
    ... Ruby's a very nice language. ... They'd rather pay millions to get a java solution, ... The problem with job security in that circumstance only really arises if ... then the project tanked at a cost of millions. ...
    (comp.lang.ruby)
  • Re: Joel Spolsky on languages for web programming
    ... Ruby's a very nice language. ... Ruby is more likely to get you fired that Java. ... then the project tanked at a cost of millions. ... Corporate politics. ...
    (comp.lang.ruby)
  • Re: Comparing Lisp conditions to Java Exceptions
    ... All the ISO standards in the world will not make the world ... Nothing keeps you from annotating your program with exceptions based on what ... language should adhere to your theory. ... Curiously, although you don't say it, Java has the opposite problem. ...
    (comp.lang.lisp)
  • Re: casts
    ... This is why most shit programmers refuse to learn languages including ... C Sharp and Java. ... compiler in a later edition of Visual Basic, ... language for processing data. ...
    (comp.lang.c)