Garbage Collection Survey

[S.Marion]

Hello,

We are interested in programmers' experiences of programming languages
supported by managed runtimes (including but not limited to Java, C#, etc).
In particular, we are interested in bugs and performance limitations that
can be ascribed to the garbage collector.

We would be grateful to receive any accounts of such problems encountered.
If you think you can help, would you please complete the form located at
http://www.cs.kent.ac.uk/people/rpg/sm244/GCForm/

Thank you for your assistance.

Richard Jones
Sebastien Marion

[Joe Seigh]

On Fri, 03 Jun 2005 10:52:06 +0100, S.Marion <sm***@kent.ac.uk> wrote:

Hello,

We are interested in programmers' experiences of programming languages
supported by managed runtimes (including but not limited to Java, C#, etc).
In particular, we are interested in bugs and performance limitations that
can be ascribed to the garbage collector.

We would be grateful to receive any accounts of such problems encountered.
If you think you can help, would you please complete the form located at
http://www.cs.kent.ac.uk/people/rpg/sm244/GCForm/

Thank you for your assistance.

Nice survey. Select for only bad experiences. Won't be too hard to imagine
what the conclusion of that survey will be.

I use GC is get massive improvements in performance in specific situations
though it's not part of a managed runtime and it isn't the GC you're thinking
of.

--
Joe Seigh

When you get lemons, you make lemonade.
When you get hardware, you make software.

[John Carson]

"Joe Seigh" <js*******@xemaps.com> wrote in message
news:opsrsk2ierqm36vk@grunion

On Fri, 03 Jun 2005 10:52:06 +0100, S.Marion <sm***@kent.ac.uk> wrote:

Hello,

We are interested in programmers' experiences of programming
languages supported by managed runtimes (including but not limited
to Java, C#, etc). In particular, we are interested in bugs and
performance limitations that can be ascribed to the garbage
collector. We would be grateful to receive any accounts of such problems
encountered. If you think you can help, would you please complete
the form located at http://www.cs.kent.ac.uk/people/rpg/sm244/GCForm/

Thank you for your assistance.

Nice survey. Select for only bad experiences. Won't be too hard to
imagine what the conclusion of that survey will be.

The OP didn't state that it was a survey (as in something designed to gauge
what proportion have good experiences, what proportion bad etc.).
I use GC is get massive improvements in performance in specific
situations though it's not part of a managed runtime and it isn't the
GC you're thinking of.

A rather unenlightening tease.

--
John Carson

[S.Marion]

Sorry I think you miss the point.
Our goal is to find the kind of problems one could experience with a GC.
We then aim to improve or develop some tools to help the developer
avoiding problems.
This has nothing to do with "GC is good or GC is bad".

Joe Seigh wrote:

Nice survey. Select for only bad experiences. Won't be too hard to
imagine
what the conclusion of that survey will be.

I use GC is get massive improvements in performance in specific situations
though it's not part of a managed runtime and it isn't the GC you're
thinking
of.

[Rapscallion]

S.Marion wrote:

Our goal is to find the kind of problems one could experience with a GC.
We then aim to improve or develop some tools to help the developer
avoiding problems.

You are in the right group here. Standard C++ has no GC and will have
no GC in the future (unless the Prince of Darkness succeeds). It's
quite easy "to improve or develop some tools to help the developer
avoiding problems" with CG in C++. :-D

[Joe Seigh]

On Fri, 03 Jun 2005 19:15:35 +0100, S.Marion <sm***@kent.ac.uk> top posted:

Sorry I think you miss the point.
Our goal is to find the kind of problems one could experience with a GC.
We then aim to improve or develop some tools to help the developer
avoiding problems.
This has nothing to do with "GC is good or GC is bad".

Joe Seigh wrote:

Nice survey. Select for only bad experiences. Won't be too hard to
imagine
what the conclusion of that survey will be.

I use GC is get massive improvements in performance in specific situations
though it's not part of a managed runtime and it isn't the GC you're
thinking
of.

The point is there is more than one kind of GC and that they don't have to be
used blindly, ie. all or nothing. Most of the problems that I'm aware of are from
not realizing that fact. Thought I suppose there are environments like Java
or .Net where you don't have any choice in the matter. C++ isn't one of those.
Not yet, anyway. For Java you can look at the rationale of JSR 166 for some of the
problems they were running into and what they did to work around them.

http://gee.cs.oswego.edu/dl/concurrency-interest/

--
Joe Seigh

When you get lemons, you make lemonade.
When you get hardware, you make software.

[Ed Jensen]

John Carson <jc****************@netspace.net.au> wrote:

Nice survey. Select for only bad experiences. Won't be too hard to
imagine what the conclusion of that survey will be.

The OP didn't state that it was a survey (as in something designed to gauge
what proportion have good experiences, what proportion bad etc.).

The bottom of the web page reads:

We shall summarise the results of this *survey* at
http://www.cs.kent.ac.uk/people/rpg/...orm/survey.php

Emphasize mine.

[Artie Gold]

Joe Seigh wrote:

On Fri, 03 Jun 2005 10:52:06 +0100, S.Marion <sm***@kent.ac.uk> wrote:

Hello,

We are interested in programmers' experiences of programming languages
supported by managed runtimes (including but not limited to Java, C#,
etc).
In particular, we are interested in bugs and performance limitations that
can be ascribed to the garbage collector.

We would be grateful to receive any accounts of such problems
encountered.
If you think you can help, would you please complete the form located at
http://www.cs.kent.ac.uk/people/rpg/sm244/GCForm/

Thank you for your assistance.

Nice survey. Select for only bad experiences. Won't be too hard to
imagine
what the conclusion of that survey will be.

I use GC is get massive improvements in performance in specific situations
though it's not part of a managed runtime and it isn't the GC you're
thinking
of.

I suspect you know not to whom you speak. ;-)

--ag

--
Artie Gold -- Austin, Texas
http://it-matters.blogspot.com (new post 12/5)
http://www.cafepress.com/goldsays

[John Carson]

"Ed Jensen" <ej*****@visi.com> wrote in message
news:11*************@corp.supernews.com

John Carson <jc****************@netspace.net.au> wrote:

Nice survey. Select for only bad experiences. Won't be too hard to
imagine what the conclusion of that survey will be.

The OP didn't state that it was a survey (as in something designed
to gauge what proportion have good experiences, what proportion bad
etc.).

The bottom of the web page reads:

We shall summarise the results of this *survey* at
http://www.cs.kent.ac.uk/people/rpg/...orm/survey.php

Emphasize mine.

Of course it is a survey in the sense that people are being asked questions,
the results of which might be summarised.

That does not make it a survey "as in something designed to gauge what
proportion have good experiences, what proportion bad etc." In other words,
it is not intended to provide statistics representative of the *population
of garbage collection users as a whole*. It is designed to identify
problems. There is nothing wrong with that, any more than there is something
wrong with a software company setting up a web site for the reporting of
bugs in its products.

--
John Carson