Russel Winder's Website


BCS Rebranded

It seems that the BCS has undergone a rebranding. Sadly rebranding generally cost a lot of money (that could be better spent elsewhere, e.g. on member services), and usually achieves precious little.

Where the BCS brand used to be very blue, it has now gone very green - I wonder why in this day and age of (often pretentious) concern for the environment?

The real question though is why rebrand, what was wrong with the old one? The BCS has a history of its academic members feeling the organization is too focused on commercial interests, and the industrial and commercial members feeling the organization is too focused on academic interests. In reality, neither was true: the BCS was the chartered body in charge of IT in the UK, but the IET (Institution of Engineering and Technology) has more finance and power, and the Information Technologists' Company more panache - though I still think calling it the Worshipful Company of Information Technologists' had even more. Perhaps though a BCS rebrand can be linked to a reenergizing of the whole organization, including its membership.

A serious problem is that membership of the BCS is not an issue of relevance to employers, this means it is not an issue of relevance to employees. Employers never seem to ask about membership of the BCS when employing staff, and until they do people seeking jobs have no real incentive to join the BCS. The BCS may be the "Chartered Institute for IT" but until IT is a profession as accountancy, architecture or medicine are then the fact that there is a"Chartered Institute for IT" will not be of relevance to people practicing in the field.

Is this in fact the beginning of a long distance campaign to introduce statutory controls on the profession of software development? Many would say"and about time too" whilst many others would say"and that will kill off all innovation". Certainly introducing more engineering approach to software development would be a good thing, but software engineering is uniquely different to all the other forms of engineering because the stuff being manipualted has no substance, it is just information without physicality.

Most people who do software engineering have no real relationship with the BCS. People in embedded systems tend to associate with IET, possibly because of the electronics and safety critical systems connections. People who consider themselves software engineers but not in embedded systems rarely associate with the BCS, and instead associate with "clubs" such as ACCU. Indeed many of these people, who the BCS really ought to value as members, see no connection at all between themselves and the BCS. And few of them see the new brand as changing anything.

So just what does IT mean to the BCS? Sadly the newly branded doesn't actually say. As we know as far as the national curriculum goes, IT means "how to use office automation software" (basically ECDL type material). Here lies the real problem: IT means too many things, and different things to different people. If the BCS is really going to grow in size to have all software developers as members then it is going to have to do something more than just a surface rebranding, it is going to have to appeal to the software professionals who currently think it is irrelevant to their lives.

A turn up for the books

Is it possible the UK government's e-petitions system can actually have some benefit? After 30,000+ people signed the e-petition about the treatment in the 1950s of Alan Turing, one of the world's greatest mathematicians and computer scientists, leading directly to his suicide, Gordon Brown made this statement. Number 10 even made it the top story on the opening page (at least on 2009-09-11).

Of course the question is will Number 10 now re-think their rather poor answer to the e-petition about Bletchley Park, which is where Alan Turing undertook all his code breaking work during World War 2. (20,000+ people signed this e-petition.)

UKUUG Summer 2009 Conference

I gave a presentation _Shared Memory Multithreading is the Wrong Way to do Pa rallelism_ at UKUUG Summer 2009 Conference. It went well - after the 45 min talk, the questions went on for 30 min and this was after the end of the day! Of course questioning whether Unix and Linux (and the other operating system) can survive the next phase of the Multicore Revolution at a Unix user group conference could be seen as a bit contentious, but it was all done in a constructive spirit. Many of the people at the session caught up with me the following day to say how much they enjoyed the session, and how it was making them think about the next generation of applications and operating systems.

I also rehashed my SCons the Builder_ from EuroPython 2009 to create a 5 min lightning talk advertising SCons as the future of build. I entitled the session If you are using Make, you are doing it wrong_.

Gant 1.7.0 released

With the formal release of Groovy 1.6.4, it seemed time for there to be a Gant release. Since there were lots of new goodies, it seemed appropriate to release 1.7.0 rather than 1.6.2.

Gant 1.7.0 is in the Codehaus Maven repository groupId:org.codehaus.gant, version:1.7.0, but remember three different artefactIds, gant_groovy1.5, gant_groovy1.6, and gant_groovy1.7 - you have to use a version of Gant that matches the version of Groovy you are using.

Distributions of Gant are in the usual place accessible from the Gant home page.

UKUUG Summer Conference 2009

I am going to be doing a talk at UKUUG Summer Conference 2009, which is being held 2009-08-08 to 2009-08-09 at Birmingham Conservatoire. Content of the talk will be parallelism, how to do it, and how not to do it. I gave the talk the title Shared-memory Multithreading Is The Wrong Way To Do Parallelism - nothing like being a bit contentious.

The conference website is at http://summer2009.ukuug.org.uk

Looking forward to EuroPython

My two proposals for presentations got accepted for EuroPython:

  • GIL isn't Evil GIL is the Global Interpreter Lock which forces a Python interpreter to be able to execute one and only one thread of Python code at a time. Many people think this stops Python being able to execute Python code in parallel. How wrong those people are. This talk is about the issues invovled and the various solutions. Communicating Sequential Processes (CSP) will undoubtedly get a mention.
  • SCons the Builder Using dynamic programming languages is changing the face of build -- which used to be dominated by Make. This talk is about how SCons harnesses Python to be able to introduce new techniques of build specification. Comparisons with Gant and Gradle will no doubt be made.

EuroPython 2009 is being held from 28th June to 4th July 2009 in Birmingham, United Kingdom.

The currently accepted talk abstracts are now up at http://www.europython.eu/talks/talk_abstracts/

The booking form is here http://www.europython.eu/registration/

The early bird rate closes on Thursday May 21st.

New Laptop

It has been time for a while now that I got a new laptop or two (I always take two laptops with me when I go to client sites or conferences, so as to avoid embarassment if one fails to work). Until now I have always had two Ubuntu laptops. I decided that now was the time to go heterogeneous, so I got a second-hand white MacBook running Mac OS X Leopard to act as my number 2 machine. I really like MacBook hardware, it looks beautiful and the keyboards are superb to use. However I still prefer Ubuntu over Mac OS X.

For my number 1 machine I decided to get a high-performance laptop with a high-resolution screen. Having looked far and wide, including at Dell who claim to sell high-performance Ubuntu machines in the UK but actually they don't, and indeed at various US companies claiming to import Ubuntu machines to the UK, I settled on Linux Emporium as the supplier, and the Lenovo T500 with Ubuntu 9.04 pre-installed as the machine.

Order placed Tuesday morning, laptop arrived Friday morning, exactly as predicted by the folk at Linux Emporium. Having spent a day personalizing and integrating it into my set-up, I really like it. Thanks to John and the others at Linux Emporium for getting exactly the machine I needed to me in the time I needed it.

Ubuntu 9.04 Jaunty Jackalope upgrade from 8.10 Intrepid Ibex is a bit of a

I have been using Ubuntu for a number of years now and stuck with it even when all my colleagues and associates abandoned it for Debian (testing or unstable rather than stable) or they bought Apple Mac OS X equipment. The Intrepid -> Jaunty upgrade has however seriously dented my confidence, and I am now thinking seriously about whether running Debian Squeeze would be a better way forward for me.

Problem 1: I upgraded my laptop which used to work fine under Intrpeid, and now under Jaunty, after a random period of time under random circumstances, it completely siezes up. The mouse pointer moves but no amount of clicking or keyboard tapping has any effect whatsoever. Reboot the machine and it is all fine again - for a while.

Problem 2: I upgraded my server and my RAID1 disc failed to be present on reboot. Fortunately this is data only, the system is on a separate non-RAID disc so boot was no problem other than fsck returning an error code so boot has to be continued manually. Very fortunately the fix reported in a bug report on Launchapd saved the day.

Do I ditch Ubuntu and switch to Debian or just suffer the 6-monthly update pains rather than risk the continuous pain of a continuously updated system such as Debian Unstable?


Copyright © 2017 Russel Winder -