Entries for Bent Rasmussen

Reactive PDC

Bent Rasmussen — Mon, 23 Nov 2009 23:12:28 GMT

Some recommendations for PDC09. There's many but maybe we should recommend highlights.

Feel free to add suggestions.

PDC 2009 [PDC 2009]

Bent Rasmussen — Fri, 20 Nov 2009 19:33:59 GMT

I have to sum up PDC 2009 and I have to do it in one word.

Awesomeness!

The new interactive live format, broadcast to the world, was immensely cool.

It may be that this years PDC was shorter but it was one hell of a ride. Thanks for the content and the format!

:-)

in reply to PDC 2009

Code Format [Code Format]

Bent Rasmussen — Wed, 18 Nov 2009 19:37:00 GMT

It's not possible to express that a string in a C# codeblock should be bolded, e.g.

[code language="csharp"]var? x = 1.0;[/code]

I realize this is not correct C# and suspect it has to do with the fact that in VB.NET you can use XML directly.

in reply to Code Format

SPDY [SPDY]

Bent Rasmussen — Sun, 15 Nov 2009 22:19:58 GMT

Anyone read about SPDY (pronounced "speedy")? It's Google's experimental HTTP replacement.

http://dev.chromium.org/spdy/spdy-whitepaper

It looks like Google is playing with the idea of having Chrome work with a SPDY server. This would give Chrome quite an advantage if Google could engineer Chrome such that it would automatically switch to SPDY upon detection (however that is done).

Maybe Microsoft, Google, Mozilla, Opera, Apple, IBM etc. could work together on a faster Web in a similar vein...

Until then, it looks like the application of embrace and extend is in process, currently by Google.

in reply to SPDY

[Side-]Effects [[Side-]Effects]

Bent Rasmussen — Wed, 11 Nov 2009 19:46:14 GMT

I thought it would be interesting to gather a collection of more or less strange effects, sort of an effect zoo. Effects can be harmful to program reasoning in many ways but they can also be fun.

in reply to [Side-]Effects

Bing + Wolfram|Alpha : It has begun [Bing + Wolfram|Alpha : It has begun]

Bent Rasmussen — Wed, 11 Nov 2009 19:19:20 GMT

Link

in reply to Bing + Wolfram|Alpha : It has begun

MSN revitalized [MSN revitalized]

Bent Rasmussen — Wed, 04 Nov 2009 23:03:19 GMT

It looks like MSN has received some "Bing"-love and now looks very useable and clean - and integrates smoothly with fb.
I may yet live to see MSN become my start page :-)

http://www.msn.com/preview.aspx

in reply to MSN revitalized

IObservable, IObserver, [Observable] [IObservable, IObserver, [Observable]]

Bent Rasmussen — Sat, 31 Oct 2009 20:24:42 GMT

I wonder what the reasoning is for not including Observable (Reactive Extensions for .Net) as part of .Net 4.0? Clearly one would want these combinators to start using LINQ or one would most likely build them oneself. I was browsing for Rx as part of .Net 4.0 beta 2 and only found the interfaces. Parallel Extensions for .Net are already part of .Net. I suppose it is still being worked out what combinators to include and specifically how to interact with the parallel extensions, hmm...

A great example of this is the Reactive Framework (RX), which is a library of extension methods (not included as part of .NET 4) that implement the LINQ Standard Query Operators and other useful stream transformation functions for IObservable<T>.

On the other hand, building this stuff sounds fun. Think of implementing task preemption using both "continuous" (hard) and discrete limits.

in reply to IObservable, IObserver, [Observable]

Code [Code]

Bent Rasmussen — Mon, 26 Oct 2009 20:39:24 GMT

Provide some padding around code blocks a la blockquotes. Difference

[code language="csharp"]class A<T> { ... }[/code]

class A<T> { ... }

So to gain the best of both one has to mix them

[code language="csharp"]class A<T> { ... }[/code]

Not too efficient.

A downside of the monospace code formatting is that it takes up more space so code can sometimes look a bit horrible in the forums.

Also, I always manually create block quotes. Is there a UI way to do this, it's a nuisance.

...And then of course Haskell code formatting ;-)

in reply to Code

Thread Pre-Emption [Thread Pre-Emption]

Bent Rasmussen — Mon, 26 Oct 2009 20:12:01 GMT

What is the best practice for thread pre-emption/cancellation? I have the following scenario: (And do not yet use Parallel Extensions for .Net)

An external service call often takes a long time to complete (as little as 1 second and as high as 10+ seconds). This brings an unacceptable latency down to the UI layer (the UI is in a different application using a web service component written in .Net using COM interop.) Anyway, the base problem is a single call which has a variable latency and the code cannot be granularized further as the latency is external.

So clearly the call needs to take place asynchronously. This is now almost implemented. The problem then arises, what to do with cancellation. If the code was fine-grained it would be possible to make it regularly read a cancellation state variable. This is not the case - the call is "monolithic" - so such a variable cannot be used. Then looking at the Thread API there is a method called Abort(). Absent a brutal sounding method like Kill() (which would obviously also be evil), I expect this to be the suitable method for this scenario. Right or?

I use a "synchronization counter" to ensure the worker thread doesn't modify the environment if it accidentally completes post-cancellation.

in reply to Thread Pre-Emption

PNRP [PNRP]

Bent Rasmussen — Tue, 20 Oct 2009 21:50:49 GMT

Anyone tried using PNRP?

I tried accessing it through the cmdl. shell, adding a registration like this

netsh
netsh>p2p pnrp peer
netsh p2p pnrp peer>add registration 0.whatever

or just

netsh p2p pnrp peer add registration 0.whatever

Using 0. as a prefix to the peer name means it becomes an "unsecure" peer name, meaning it's not really unique but it should do for testing. And obviously you substitute "whatever" with whatever you like as a name.

The registration can be checked by entering

netsh p2p pnrp peer>show registration

Resolution should happen in a different process, so in a new shell -

netsh p2p pnrp peer resolve 0.whatever

On my network and machine setup this is quite fragile - it appears to work sometimes, other times not. Sometimes can be seen immediately, other times not even after say 15 minuttes. And even when resolution occurs locally, it doesn't work outside my local network.

Once you can resolve locally, try this

ping whatever.pnrp.net

I also have that working, just not from any other machine than my local.

For those interested, see the PNRP videos uploaded on Channel 9. Search for the tag.

in reply to PNRP

Ideas for C# 5.0 [Ideas for C# 5.0]

Bent Rasmussen — Sun, 18 Oct 2009 19:12:09 GMT

What would you like to see in C# 5.0?

in reply to Ideas for C# 5.0

Pionen [Pionen]

Bent Rasmussen — Tue, 13 Oct 2009 15:25:40 GMT

This deserves a separate thread. Tour of the Pionen data center

Looks quite impressive for a civilian data center.

The data center

Facts about the data center

Originally a nuclear bunker: The data center is housed in what was originally a military bunker and nuclear shelter during the Cold War era. The facility still has the code name from its military days: Pionen White Mountains.
Located in central Stockholm below 30 meters (almost 100 ft) of bedrock: The facility has 1110 sqm (11950 sq ft) of space and is located below 30 meters of solid bedrock (granite) right inside the city.
Fully redesigned in 2007-2008: Pionen was completely redesigned in 2007-2008 to become the data center that it is today. More than 4,000 cubic meters (141,300 cubic ft) of solid rock was blasted away to make more room.
Can withstand a hydrogen bomb: The bunker was designed to be able to withstand a near hit by a hydrogen bomb.
Houses the Network Operations Center for one of Sweden’s largest ISPs: The bunker houses the NOC for all of Bahnhof’s operations. They have five data centers in Sweden, Pionen being the largest. The facility also acts as a co-location hosting center, so you can actually put your own servers here.
German submarine engines for backup power: Backup power is handled by two Maybach MTU diesel engines producing 1.5 Megawatt of power. The engines were originally designed for submarines, and just for fun the people at Pionen have also installed the warning system (sound horns) from the original German submarine.
1.5 megawatt of cooling for the servers: Cooling is handled by Baltimore Aircoil fans producing a cooling effect of 1.5 megawatt, enough for several hundred rack-mounted units.
Triple redundancy Internet backbone access: The network has full redundancy with both fiber optics and extra copper lines with three different physical ways into the mountain. Pionen is one the best-connected places in northern Europe.
Work environment with simulated daylight and greenhouses: For a pleasant working environment the data center has simulated daylight, greenhouses, waterfalls and a huge 2600-liter salt water fish tank.
Staff: 15 employees, only senior technical staff, work full time in Pionen.

Pics en masse

There's possibly a new bunker available. The nuclear-safe Norad bunker

in reply to Pionen

128-bit architechture [128-bit architechture]

Bent Rasmussen — Tue, 13 Oct 2009 12:45:33 GMT

The recent discussion/speculation around a possible Windows 8/9 with a 128 bit architechture sounds interesting if there is any truth to it, which of course none of us know. What are the possible requirements that would drive this design? It reminds me of ZFS which is also 128 bit.

So how much is 64 bit?

18,446,744,073,709,551,616 bits

2,097,152 terabytes

2,048 petabytes

That's a pretty big number. No single computer is likely to have that much space for a long time to come.

On the other hand, if we talk about this globally, we're talking about 2 million people with a 1 terrabyte disc. That threshold has already been exceeded.

128 bit on the other hand is so perverse that it can't possibly be exceeded

Then there's cloud computing.

Is this about creating a shared virtual memory address space for the whole world wide web or a really big cloud?

Sounds cool.

Is there a global memory counter for the world? A mechanical (and possibly a biological) one (although I don't believe we know the memory power of the brain yet?)?

in reply to 128-bit architechture

Silverlight Blog Software [Silverlight Blog Software]

Bent Rasmussen — Sun, 11 Oct 2009 20:34:22 GMT

Anyone know of blogging software that is actually targetting Silverlight? As a presentation format for the RSS/A. One thing that it's not a good idea to search for is "Silverlight blog" and the Silverlight Team blog is not written in Silverlight. It seems like it might be a good showcase for Silverlight as blogs are ubiquitous and the Silverlight feature set could add some much needed consistency and styling to blogs. Of course it's not always essential or pragmatic to eat your own dogfood in all possible situations but it would make an interesting case. A large and important part of the Web is the blogosphere, after all. The microblogosphere or twitersphere another (and here a classical interface is really outdated.)

in reply to Silverlight Blog Software

Soapbox Part Deux? [Soapbox Part Deux?]

Bent Rasmussen — Sat, 10 Oct 2009 18:27:11 GMT

Do you think Soapbox will ever reemerge - possibly using Silverlight? Doesn't YouTube make financial sense with all the advertisements now? Is the threat of IP lawsuits just too great? I never used Soapbox - didn't like the UX - but would have thought Microsoft could make it work as well as YouTube or Vimeo. What do you think - is a Microsoft video service for the world out of the question or is Microsoft content with not having a presence in this significant market? I wonder how they see the future of media dissemination. The traditional stores and webradio being the final word, or?

in reply to Soapbox Part Deux?

Bottoms Up [Bottoms Up]

Bent Rasmussen — Sun, 04 Oct 2009 21:25:39 GMT

Bart de Smet mentions the bottom type (⊥) in a blog post about Let. This is quite a facsinating type as it signals a type with no values whereas void is the type with one value (nothing) - but it can be used to model functions that never return, e.g. functions with infinite loops or functions that always throw exceptions; in the latter case the function does in a way "return", just not in a pure sense. This is pretty interesting for the case of LINQ and IEnumerable due to the fact that we can easily build infinite streams.

If you build an IEnumerable with a "yielded" while loop inside, the function does return immediately, but the Count extension method over an infinite IEnumerable will never return. The compiler can't always see if a function will never return but in the cases where it can statically determine that it never will, the type system could statically prevent the use of Count, for example. I don't know, but it sounds interesting. And of course one can manually assist the compiler (which will of course double-check to see if the assisted type can be proved wrong). So maybe the type of Count should be (never | int) or Either<never,int>.

What do you think? This sounds like a good topic for a future C#/VB or SomeOtherLanguage interview on C9, since I'm guessing some people are playing around with using these ideas in .Net languages.

in reply to Bottoms Up

PCP, Speed & Psychedelic LINQ [PCP, Speed & Psychedelic LINQ]

Bent Rasmussen — Sun, 04 Oct 2009 20:23:36 GMT

I'm talking about, of course, the Parallel Computing Platform team's PLINQ and the speed of LINQ in general.

Seeing Bart de Smet's blog post on IEnumerable.Let and IEnumerable memoization has me thinking. Has anyone tried measuring LINQ performance between CLR 2 and CLR 4? There is a large collection of LINQ samples floating around and maybe even real benchmarks for LINQ so it would be quite cool to see how LINQ performance is doing and how much you might pay (or gain) for using LINQ vs more traditional imperative styles (well, imperative style must win, in the extreme - but depends of course on the degree to which you're willing to play human compiler.)

PLINQ on a 4+/HT core along with memoization sounds compelling but I don't know how aggressive the C#/VB and CLR compilers are with LINQ.

There's also this product which is apparently a LINQ optimization module: LINQ on Speed. Anyone else curious about LINQ speed?

in reply to PCP, Speed & Psychedelic LINQ

vimeo.hs [vimeo.hs]

Bent Rasmussen — Sat, 03 Oct 2009 22:22:41 GMT

I find Vimeo a more smooth experience than YouTube, in particular many videos are HD, although that's not a rule and purely subjective. Anyway, I found this channel on Vimeo for Haskell. http://vimeo.com/channels/haskell

in reply to vimeo.hs

Logicomix [Logicomix]

Bent Rasmussen — Sat, 03 Oct 2009 10:42:22 GMT

Seen this? Logicomix

in reply to Logicomix

Microphones & Audio Quality [Microphones & Audio Quality]

Bent Rasmussen — Sat, 03 Oct 2009 10:09:43 GMT

Not many people have written about it but I'd like to point out that the audio quality could be improved. It sounds like many of the used microphones have poor sound quality or the post production worsens the audio quality. I think the speakers and Channel 9 deserve better.

in reply to Microphones & Audio Quality

Code & Graphics [Code & Graphics]

Bent Rasmussen — Thu, 01 Oct 2009 18:27:10 GMT

Any news on the "secret" project Brian Beckman is working on? Brian briefly mentioned he was working on an undisclosed project involving program visualization. I wonder if this is related to the new Visual Studio 2010 WPF based editor which supports mixing graphics and code seamlessly.

Hm, makes me want to build a simple extension that turns <= into ≤; != into ≠; not into ¬, etc.

in reply to Code & Graphics

Mark As ... [Mark As ...]

Bent Rasmussen — Thu, 01 Oct 2009 06:12:44 GMT

I think there's been a tendency in the past to use Mark As Spam for requesting moderation on a topic, regardless of whether it is spam, nonsense, flames or trolls. Maybe it should be called Request Moderation and later with a "popup" asking for a reason (probably multiple choice).

in reply to Mark As ...

Tracking video discussions [Tracking video discussions]

Bent Rasmussen — Thu, 01 Oct 2009 06:05:01 GMT

To better be able to track video discussions, all video posts should automatically appear as threads in a video forum. Then one can follow when new comments appear and make comments directly from the forum - and obviously these would be shared. Ideally the root post would be prefixed with the video, desc., voting, etc. This also means that very interesting videos while gradually disappearing from the front page of Channel 9, can still remain and be seen as active in the video forum.

in reply to Tracking video discussions

Strange Attractors and TCP/IP Sequence Number Analysis [Strange Attractors and TCP/IP Sequence Number Analysis]

Bent Rasmussen — Tue, 29 Sep 2009 22:37:40 GMT

This is probably old news, but I just got it today (technically yesterday),

Strange Attractors and TCP/IP Sequence Number Analysis

This is pretty cool stuff.

in reply to Strange Attractors and TCP/IP Sequence Number Analysis