Russel Winder's Website


Ubuntu, C++ and the lack of Boost.MPI

Having finally got libboost-mpi-dev and the version specific packages into Debian and thence Ubuntu, it seems the Ubuntu management have decided it's far too much trouble and have now removed all Boost.MPI packages from Ubuntu, cf. the notification and proof it's not in Maverick. Debian are going the other way and providing backports of Boost.MPI so all versions of Debian support it, cf. proof it's in all version in one form or another.

The conclusion is clearly that whilst Debian are supporting all tools associated with multicore systems and HPC programming, the Ubuntu management have decided that C++ and MPI are an irrelevance to the future of Ubuntu.

Conclusion will be drawn. Mostly with the blue pencil of censorship.

If you care about having Boost.MPI in Ubuntu please visit this bug report and add your comment/support.

Non-English URLs and the Tower of Babel

Given the fact that the world is partitioned by language, enforcing one script as the only script for URLs is a form of cultural imperialism. So it seems a natural evolution of the Internet that non-English URLs should be allowed. However this means the Internet is now being partitioned in a way that allows greater state control and censorship - even more than there is already.

The story of the Tower of Babel, which appears in one guise or another in many ancient books, relates how there used to be one language and that everyone could communicate with each other, but that hubris and vanity led, indirectly, to the breaking up of society into tribes each using their own language.

It may not be vanity or hubris that is leading, directly or indirectly, to the partitioning of the Internet, but breaking up it certainly is doing. One just hopes that this isn't going to lead to ghettoization.

The Internet had been a tool for allowing one language, English, to become the lingua franca (and there is indeed irony in that phrase), to create a single world of communication. It is a pity that some tribes in the world saw this as imposition rather than opportunity. Maybe hubris and vanity are involved after all. More likely though it is power and politics. Sad.

GPars 0.10.0 Now available

GPars is a framework of concurrency and parallelism constructs for Groovy programmers. GPars provides actors, CSP and dataflow constructs. Underlying GPars are the JSR166 libraries and JCSP, so the foundations are good and solid. GPars gives using these concurrency and parallelism tools a Groovily idiomatic feel.

HTC succumbs to bullying?

So HTC has signed a deal paying Microsoft royalties for HTCs use of Android in their smartphones. Microsoft are of course marketing this is part of their "Linux violates patents held by Microsoft" campaign. The question is Why has HTC really succumbed to this patent bullying tactic?

Microsoft of course claim this is because they have proved to HTC that their patent claims are real and genuine. Clearly this is just not the case, otherwise Microsoft would be chasing the really big Linux users. I guess that Microsoft have said to HTC that if HTC didn't pay up over Android then HTC would not be allowed to use Windows Mobile. But then does Windows Mobile matter any more in the smartphone market? Perhaps then Microsoft are playing a "long game": by obtaining lots of smaller organizations actually paying royalties, they build up a snowball of opinion that the patent violation is real. This way they create a perception of truth of their claim without ever having to prove their case in court.

Perhaps then the question is: Are Microsoft channeling funds back to HTC in some way to pay them back for participating in this long game patent scam?

Communicating Sequential Processes (CSP) and the Multicore World

30+ years ago Tony Hoare and colleagues developed CSP, a mathematically sound model of how to structure software so as to deal as safely as possible with concurrency. Parallelism wasn't such an issue then as it is now, post the beginning of the Multicore Revolution; at the time of development of CSP very few computers offered applications more than one processor.

Now as we enter a period of massive change in computer hardware structures (uniprocessor architectures are giving way to multiprocessor architectures), it is becoming increasingly clear that shared-memory multi-threading cannot be the software architecture of future applications. The need to correctly manage locks, mutexes and monitors, as is required in shared-memory multi-threading, is just too complicated to get right.

Erlang has been the major proponent of using lightweight processes and message passing over the last 25+ years. It's successes in the telecoms industry have generally been ignored by the wider computing industry as it rushed headlong into the quagmire that is shared-memory multi-threading. Joe Armstrong has noted in the past that although Erlang does not directly realize the Actor Model or CSP they were very strong inspirations in the development of the language.

The HPC (high performance computing) community have been working with parallel systems since almost the beginning of computing. Many models have been tried over the years, but the current winner is a combination of SPMD (single program multiple data) architecture with annotations to handle local thread management: MPI and OpenMP are the tools of handling parallelism with C, C++ and Fortran the programming languages of choice. No Actor Model, and certainly no CSP. Given the unwillingness of the HPC community to rewrite any of the codes they have developed over the last 40+ years, it seems likely that the traditional HPC community will become an increasingly irrelevant backwater of computing innovation.

On the Go (the programming language being developed by Google) programming language users mailing list, Rob 'Commander' Pike reminded us in one thread that although the goroutine concept in Go appears to have some inspiration from CSP, most of the ideas were developed independently during the development of the Squeak, Newsqueak, Alef and Limbo programming languages. He also pointed out that occam and Erlang were separate developments with there being no interchange of ideas.

Parallel and independent emergence of essentially the same idea is a strong indicator that some version of that idea is a great innovation. This has been shown many times in many different fields of human endeavour.

In the data mining arena, there is increasing dissatisfaction with SQL-based approaches; they are too slow to be useful. Employing a dataflow software architecture (an event-based approach based on lightweight independent operators, which is basically a subset of CSP), there is a revolution in the approach to data mining. SQL-based approaches gain no benefit from parallel processors, dataflow approaches gain huge benefit, and are now massively outperforming SQL-based approaches.

Thus it seems there is an "emergent property" appearing to come form all this: independent lightweight processes with no shared state is a good model of applications programming in a world increasingly dominated by parallel computers. Actor Model, dataflow and CSP are, increasingly, the future for software. The JVM world is heading that way, Scala has Actor Model, GPars provides Actor Model, dataflow and CSP to Groovy and Java programmers. Even the Python world is heading that way, cf. Python-CSP.

CSP is not a _silver bullet_, there are none in computing, but it is likely one of the best ways forward for structuring software in the post Multicore Revolution, massively parallel world of applications development.

LibConDem - Liberally Conning Democracy

Two weeks ago the Cons and the LibDems were at each others throats, trying to convince the voters how appalling and evil the other was, and that the Labs were worse. As of yesterday, the Cons and the LibDems are the best of friends, indeed they are fawning over each other. Fawning is ugly at the best of time but this definitely appears as though they are mocking the voters.

Given that the Cons and the LibDems supposedly agree on very little and disagree on so much - well at least that is what they told the voters two weeks ago - it is fairly difficult to see what there is that can form the core of the programme for this government.

Clearly a politicians need to be in power overrides any thought of principles and policies.

I suspect the coming Queen's Speech will be entertaining.

Con Lib Dem or Lib Dem Con

So Gordon Brown has resigned (as Prime Minister and leader of the Labour party) and the Queen has invited David Cameron to be Prime Minister. The Conservatives (Con) have declared a full coalition with the Liberal Democrats (Lib Dem) - whatever that means in reality. The question is whether the Lib Dems are stringing the Cons along or whether the Cons are stringing the Lib Dems along. What is truly certain is that the political, ethical and moral differences between them are far more than the similarities. A very strange collaboration. So who is more desperate for power, David Cameron or Nick Clegg?

I suspect the telling point is that the Conservative party has enough money to fight another general election, but the other parties do not. What odds on a second election in 2010 with the Conservative party going for an overall majority so as to drop the Lib Dems?

What we can predict is that all the politicians will be ensuring they get money in their and their friends pockets. The voters will lose. Whatever politicians say in public, you know they are not averse to lying. Blatantly.

UK's Hung Parliament 2010

The Conservatives threatened the markets with panic if there was a hung parliament, there is a hung parliament, and the markets did a bit of panicking. Is this behaviour built into the system, or are traders a bunch of lemmings easily fooled by FUD? Rhetorical question.

We have a hung parliament for the first time in 34 years. Now we see politicians negotiating with each other in a desperate scramble for power. Fairly ugly in itself. Interestingly though, if the UK moves to proportional representation, we will see this sort of "horse trading" after every election. The "first past the post" system has its benefits as well as its clear unfairnesses compared to proportional representation. I wonder if there will be proper debate about this, or whether the political classes will just impose the solution that is best for the political classes on the electorate. Do the political classes just see the electorate as sheep to be herded? Not a rhetorical question.

Software in Browsers

Many moons ago (about 1996), Web browsers included Java to provide client-side computation capability - applets became all the rage. Sadly most computers weren't really powerful enough at that time to properly support applets that weren't just trivial, and so client-side Java died off to be replaced by JavaScript (aka ECMAScript or whatever it is called this year). Actually client-side computation became unfashionable and seen as a "bad thing". Everything moved to being server-side, generally to the detriment of good human-computer interaction (HCI) in the Web browser.

Fashions change. Now it is cool to do things client-side: AJAX, etc. mean that client-side is now considered a "good thing" - actually this could be good for quality HCI in browser-based application, sadly it rarely is as the developers all too often have no idea about HCI and what makes a good user interface.

Now there are rumours that people who like Mono are saying that CLI should be included in all browsers. Apart from all the software patent issues because Microsoft own patents in CLI, is this just Mono folk being shills for Microsoft trying to get .NET as the standard for all Internet-based activity?

History tells us that the JVM (of which CLI is a rip-off) can already be used client-side in browsers, so why not just return to using Java applets? Computers are now vastly more powerful than they were when applets stopped being all the rage. Perhaps if they were tried again, they could become all the rage again.


Copyright © 2017 Russel Winder -