RE: [aus-dotnet] [OT] Joel Pobar talk in Perth

[Woods, Keith]

While I enjoy all the Perth presentations I think this was my favourite. Thanks to Nick, Alastair and Mitch (and anyone I’ve missed on the Perth .Net COP  team) for making it possible. It was a real interesting discussing right from how the processor interprets and optimises instructions to practical .net tips on how to effectively make use of locks then followed by an overview of the DLR.

 

IL if your looking for more information on locking you may find these links useful:

http://msdn2.microsoft.com/en-us/library/1c9txz50.aspx

http://msdn2.microsoft.com/en-us/library/ms173179.aspx

 

Joel even mentioned not to lock on private strings, I remember reading about this (http://odetocode.com/Blogs/scott/archive/2007/03/26/10628.aspx) a while back and was going to point it out but didn’t have to as he mentioned it.

 

Locking on literal strings is especially risky because literal strings are interned by the common language runtime (CLR). This means that there is one instance of any given string literal for the entire program, the exact same object represents the literal in all running application domains, on all threads. As a result, a lock placed on a string with the same contents anywhere in the application process locks all instances of that string in the application

(from the second link above).

 

Another great take away was the ‘Shared Source CLI’ (also known as Rotor):

http://www.microsoft.com/downloads/details.aspx?FamilyId=8C09FD61-3F26-4555-AE17-3121B4F51D4D&displaylang=en

 

Overview

The Common Language Infrastructure (CLI) is the ECMA standard that describes the core of the .NET Framework world. The Shared Source CLI is a compressed archive of the source code to a working implementation of the ECMA CLI and the ECMA C# language specification.

The current release builds and runs on Windows XP only. It is released under a Shared Source initiative. It is released under a shared source initiative. Please see the accompanying license.

The Shared Source CLI goes beyond the printed specification of the ECMA standards, providing a working implementation for CLI developers to explore and understand. It will be of interest to academics and researchers wishing to teach and explore modern programming language concepts, and to .NET developers interested in how the technology works.

Features
The Shared Source CLI archive contains the following technologies in source code form:

·         An implementation of the runtime for the Common Language Infrastructure (ECMA-335).

·         Compilers that work with the Shared Source CLI for C# (ECMA-334) and JScript.

·         Development tools for working with the Shared Source CLI such as assembler/disassemblers (ilasm, ildasm), a debugger (cordbg), metadata introspection (metainfo), and other utilities.

·         The Platform Adaptation Layer (PAL) used to port the Shared Source CLI from Windows XP to other platforms.

·         Build environment tools (nmake, build, and others).

·         Test suites used to verify the implementation.

·         A rich set of sample code and tools for working with the Shared Source CLI.

New in this Release

·         Full support for Generics.

·         New C# 2.0 features like Anonymous Methods, Anonymous Delegates and Generics

·         BCL additions.

·         Lightweight Code Generation (LCG).

·         Stub-based dispatch.

·         Numerous bug fixes.

 

 

From: peter@xxxxxxxxxxx [mailto:peter@xxxxxxxxxxx] On Behalf Of ILT
Sent: Friday, 8 June 2007 2:45 PM
To: dotnet@xxxxxxxxxxx
Subject: RE: [aus-dotnet] [OT] Joel Pobar talk in Perth

 

Points well made, Nick. I apologise if it seemed to reflect on the Perth .NET C of P organizers – it was not meant to.

While writing the post, I was thinking about my invitation to an informal discussion group of the Western Australian Division of Engineers Australia, by an acquaintance of mine who coincidentally was the head of IBM for the Asia region (9 countries).

That group is called the Software Engineering Forum.

I’m sure that he would have enjoyed Joel’s talk. I should have invited him.

IL Thomas
GeoSciSoft  - Perth, Australia

---

From: peter@xxxxxxxxxxx [mailto:peter@xxxxxxxxxxx] On Behalf Of Nick Randolph
Sent: Friday, June 08, 2007 2:21 PM
To: dotnet@xxxxxxxxxxx
Subject: RE: [aus-dotnet] [OT] Joel Pobar talk in Perth

 

You hit on a bit of a sore point IMHO in your last point – the reality is that it was opened to a wider audience and yet we still only had marginally (approx 1.5 times) more than a usual user group meeting.  Extending the reach of the user group here in Perth has always been difficult and despite the support from the DPE team on the east coast we have little (in fact none might be more accurate) support from the local branch.  This year we have been attempting to get the local branch to be more involved – we are looking to follow the other states in forming a .NET cluster - but despite knowing the event was happening I’m almost positive no one in the local MS branch was out there promoting it to their customers.

 

Anyhow that’s my 2 cents worth (I’ll put some other comments on my blog at some stage)

 

Nick Randolph

S:         nick_randolph

B:         http://community.softteq.com/blogs/nick

 

 

---

From: peter@xxxxxxxxxxx [mailto:peter@xxxxxxxxxxx] On Behalf Of ILT
Sent: Friday, 8 June 2007 10:21 AM
To: dotnet@xxxxxxxxxxx
Subject: [aus-dotnet] [OT] Joel Pobar talk in Perth

 

I must say that Joel Pobar’s enthusiastic presentation yesterday in Perth was impressive, covering a wide gamut and displaying (for me) a nice balance of the broad overview and the code-specific. It’s great to see the person behind the blogs and the reputation. And he’s just a young guy from Brizzie.

Joel gave a disturbing view of the pitfalls of the brave new world of multi-processor computing that is almost upon us, way beyond the considerations of threading and deadlocks that are (sort of) known to all of us in the audience yesterday. Also, I liked the “Intel has screwed us” theme, and smiled at the screw graphic adorning a couple of the slides.

I found the “best practice” section of his concurrency session – use of locks, etc – a really useful (but very rushed) tutorial and I would like to see this amplified into a paper that it pitched at the average programmer (and I’m a very average one). Perhaps it is already – if others at yesterday’s talk can point me in the right direction, I would be grateful.

I’m just waiting for the FxCop for threads and parallelism, to make all the brain work totally superfluous.

I chatting with Joel in the break, I mentioned (but forgot its project name: “Singularity”) a Microsoft Research project in which a new experimental OS was taking the sorts of “precautions” that Joel was discussing just one step further – to isolate applications in their own singular memory space – a bit like running a Virtual PC for every application. I last looked at this in January, and now see that in April or May this year it came to its v1.0 milestone, with v2.0 being roadmapped. Check out the website at MSR Operating Systems group if you’re interested – it may be the OS to replace Windows entirely!

Also, this MSR paper is worth reading for an overview (PDF) -

Singularity: Rethinking the Software Stack,
Galen Hunt and James Larus. Operating Systems Review, Vol. 41, Iss. 2, pp. 37-49, April 2007. ACM SIGOPS

I got quite carried away with this in January, and watched the several Channel9 webcasts by Jim Larus and Galen Hunt (two “old guys” at MSR). It’s actually quite an old project – begun in 2003 or 2004 I think.

Re the (further off) topic of old guys, Joel mentioned in his talk a side issue that’s a real milestone in database computing, which is transactional processing. This was in the context of SQL. What he omitted was that a really old guy who was in recent years at Microsoft Research (in San Francisco) – Jim Gray – was one of the key people involved in this work.  

When Jim Gray and his colleagues Gianfranco Putzulo and Irving Traiger did groundbreaking work on concurrency control for databases in the late '70' / early '80s, they were young guys actually – a little older than Joel Pobar, perhaps.

(Some of you will know that Jim Gray was tragically lost at sea off San Francisco while sailing alone, at the end of January this year. He was 63.)

If anyone has followed Jim Gray’s work in recent years, you will know that he had really wide interests which included the world-wide telescope, TerraService and Sky Server, massively parallel processing, grid computing, work on building fast networks, building huge web servers with “CyberBricks”, and building very inexpensive and very high-performance storage servers. All of this was in the context of relational databases. His webpage still exists at MSR, and the downloadable papers are really very accessible (ie, easy to read, interesting).

Joel Pobar’s interests seem to have a similar breadth. I overheard him talking about AI and gambling at dinner at 9 Mary’s last night.

Many thanks to the Perth .NET group for organizing this. It was a good turn-out, but it would be nice to open this up to a wider audience.

IL Thomas
GeoSciSoft  - Perth, Australia