Our Electoral Sohpie's Choice: a Darwinian Perspective

I have often wondered whether our elected representatives really believe the things they say. Some seem, to me anyway, so implausible, so divorced from common sense, that my initial conclusion was that they must have been persuaded to adopt the views of the interest groups who paid for, and will be paying their campaign expenses. This made me uncomfortable, since it suggests a weakness of will, a lack of spine and integrity, of honor and commitment, values that seem increasingly rare.

There is however an alternative, slightly less depressing explanation. Given a very large number of heterogeneous candidates, those whose views most closely correspond to the largest funding sources will be elected. This model doesn't require that elected representatives adopt the views of their funders, merely that they are selected as "most fit", in a Darwinian sense, to the political environment and the resources it makes available.

Of course, the net result is the same. Electoral issues are defined by campaign donors and the views of those who we elect represent a Sophie's choice between the interests of two sets of opposing funding groups.

I imagine, not quite what the framers of the constitution had in mind.

On the Flaw of the House

Actually, by my count there are 10, so make that 'flaws'.

[Michelle Bachmann's speech against a carbon tax made on the floor of the US House of Representatives in April 2009]

Relvance?

By chance I met Mike Parker, ex-CEO of both BNFL and Dow Chemicals. He asked what I was doing in Montreal. I told him I was there for the Academy of Management meeting. "What's that?" he asked. A timely reminder of how much we appear to matter to the organizations we study.

Conspiracy #2 - Linux

If you've heard of Linux, which I imagine most people have, and you don't work for Microsoft, you probably have the impression that it's the best thing since sliced bread (a fairly low standard I should add, since sliced bread is no match for the real thing). People are falling over themselves to extol the benefits of community software development. But there are several issues that are clearly problematic.

1) Even software that is not in the "long tail" (e.g. Linux, Firefox) has components that are: dmraid, the module that supports fakeRAID seems currently to have only one developer, Heinz Mauelshagen (who is also responsible for LVM2), and in 6+ years since its launch is still a release candidate, not a stable production version. LVM2 is used in the default partitioning scheme in Enterprise Linux (including Rad Hat, CentOS, Scientific Linux)so Mauelshagen's time must be in great demand for LVM2 issues which perhaps explains the slow progress on dmraid?

2) There is the leading edge (such as Fedora 13 at the time of writing) and other 'experimental / testing' distributions, and the trailing edge (enterprise systems). The former have great hardware support, but can be flaky since the interdependence have not been debugged. The latter seem to lack breadth of hardware support - my 1920x1080 VGA /1680x1050 DVI dual displays on and old ATI Raden card were unsupported by the community developed drivers in CentOS and Scientific Linux. Using the ATI propriety driver helped a bit but there were all sorts of unexpected interactions between the community developed interface and the configuration file (xorg.conf) and the ATI supplied interface; neither write the file in a way the other understood and often the result was to completely confuse Xorg which then wouldn't start.

So my plan to avoid the problems of only partially debugged software (Ubuntu 10.04) ran up against a lack of functionality in CentOS and Scientific Lunix.

Why a conspiracy? Partly it's the frustration talking. But perhaps this is a situation that the key users and recommenders (who are in many cases also the authors) of Linux like; it makes their jobs safe. If they convince the organization for which they work to use community software, which is often not completely up to the task out of the box, their role in supporting IT becomes more pivotal, their jobs safer and their salaries higher.

Of course they won't see it that way - they probably believe they are providing an invaluable service to the organizations for which they work by managing a computing infrastructure that doesn't involve costly licensing fees. Their own time is seldom something they add into the cost-benefit analysis, since (they need to assume) this is a given, a fixed cost that is a wash whatever OS is selected.

Of course, nothing is perfect and MVS had bugs too. But I wonder how many of the large (or small) retail banks run their transaction processing systems on Linux?

Review: the Book of Eli

"Mad Max" meets "The Name of the Rose". Gary Oldman reprises his role from "The Fifth Element" with a few more wrinkles.

OS/2

So; after some fiddling and finding I needed two updated installation disks, OS/2 Warp Version 4 is up and running. It's been over 15 years since I used OS/2 V2 so I was interested to see how accurate my memory was. The thing I remember most clearly was that after it crashed (which it did fairly regularly since it was running on a PS/2 Model 55 with probably 256k, and I think the minimum requirements were at least 512k) it restarted putting every window back where I'd left it (or rather where it had left me) to the last keystroke. My hunch that the base OS had been designed by some systems programmers from the MVS labs was confirmed many years later by a senior IBM developer. I often wonder how many hours of work we (as a computer using species) would not have lost had we been using OS/2 rather than Windows these last 20 years. The backdrop by the way, is a picture of Steyning Bowl in West Sussex, looking north-east with Steyning itself in the middle distance. It was taken from here: (50.873273,-0.349354)

So it's running under Oracle (Sun's) VirtualBox. VMware gave up on OS/2 about 5 years ago from what I gather, and though I did install it under Player 3.1, networking seems to hang the boot sequence. First impressions are that the screen is tiny. Most monitors back in 1992 were 640x480 and there is no standard SVGA support in OS/2. There may be a way to get round this but for the moment I'm looking at a postage stamp. It's going to take a while to re-learn where things are. The screenshot shows IBM's browser pointed at CNN; as you can see it's having some trouble with modern websites, but at least there's connectivity.

Was it worth it? Probably not, but I do have a sense of nostalgia and some satisfaction to have it up and running again. Reminded me of attending the PS/2-OS/2 announcement in New York in 1986. My first glimpse of Lotus 123's APA GUI for OS/2; it looked really spectacular, and in my mind's eye, clearer, brighter, sharper and more colourful that any of today's offerings (OK so Excel or OpenOffice). I'm sure if I saw it again today I'd be disappointed, but it was way ahead of the character based green on black we were used to a quarter century ago.

I'd tried installing OS/2 native on various systems I've had since 2001 but all were too modern in terms of chip architecture and weren't supported, at least not without a serious investment in time poking around on the web; virtualization makes things very much easier. Indeed, back in the day, that was one of the marketing points for VM.

What next? Don't rightly know, to be honest. Perhaps I'll see what software I can install... but that's for another time.

Reflected Glory

Yesterday, Sam Palmisano, IBM's CEO, was the opening keynote at the National Governors Association annual meeting. Among the usual corporate boiler plate citations of relevant corporate involvement (a rather longer and duller version of "I fell your pain") he thought it noteworthy to mention that among IBM's accomplishments was its involvement with Apollo 13; not the Apollo program, but specifically Apollo 13.

IBM has had a long association with NASA providing onboard guidance computers for the Gemini program, the guidance computer on the Saturn 1B and the and instrument unit on the Saturn 1B and Saturn V not to mention System/360s for the Houston flight control center and systems on Apollos 14, 15, 16 and 17.

I know Apollo 13 is perhaps the best known of the Apollo flights (although I remember 8 and 11 better), but citing the one mission that was almost the most spectacular disaster in NASA's history (not forgetting Apollo 1 and Challenger) doesn't seem like the best way of basking in NASA's reflected glory.

Some computer interfaces I have known

1) Tektronix teletype terminal (300 baud) to some anonymous mainframe - 1981
2) HP basic desktop computer (model unknown) - 1982
3) IBM 5250 into System/34, System/36, System/38, AS400 - 1982-89
4) IBM 3270 into VM/CMS, 4331, 4361, 4381, 3033, 3083 - 1982-89
5) IBM 3270 into ICCF and CICS/VSE 4381 - 1982-89
6) IBM PC DOS -1982-89
6) IBM PC Windows 3.1 - 1990
7) IBM PS/2 and OS/2 -1991
8) Dell Optiplex PC Windows NT 3.5 - 1993
9) Apple Macintosh II System 4 -1996
10) Apple Mac 6100 System 5, System 6, System 7, System 7.5 -1997
11) Dell XPS 5100, Windows NT 3.5, Windows XP Professional, Fedora Core 2, Core 3 - 2000
12) Dell Laptop (model # forgotten) Windows XP Professional -2002
13) Home built Intel based PC Windows XP Professional - 2003
14) Apple Powerbook (Motorolla) OS X -2004
14) Dell 1500 server, Windows 2000 Server, Fedora Core 3, Core 7, Ubuntu 8.04 LTS -2003-09
15) Home built Intel Core 2 duo Windows XP Professional - 2005
14) Apple Macbook Air OS X - 2008
16) Dell server, Ubuntu 9.04, 9.10 - 2009
17) Hosted VM, Ubuntu 9.10 - 2009
18) Home built Intel Core 2 quad Ubuntu 9.04, 9.10. 10.04 LTS, VMware Workstation 6.2, 6.3, 6.4, VMware Player 3.01, 3.1, Windows XP Professional - 2008-2010

VMWare Player 3.1 and Ubuntu 10.04

I almost gave up on 10.04 as a host OS since Player 3.0.1 didn't get on well with it; frequent freezes and hard reset. However, last week (May 25) VMware released Player 3.1: and that has cured the problem. So far 6 days entirely problem free.

PR Disaster? Or something more machievellian?

First some observations about the Deepwater Horizon oil spill. According to 60 Minutes, the Deepwater Horizon did have a pretty good safety record until last month. BP's recent spate of safety problems may have as much to do with its takeover of ARCO as anything else. The well that the Deepwater Horizon drilled was apparently at a record-breaking depth below sea level; working in these conditions means charting new technological territory. The government can't "take over" the effort to plug the well since the only people / companies with the requisite capabilities (knowledge and physical assets) for working at these depths are BP and the companies (Transocena, and Halliburton) who are its partners in crime.

As frustration rises so does the level of inflammatory rhetoric. Even those who may know better likely feel compelled in their public utterances to ratchet up the level of accusations and demands for action. BP and the US government need each other. BP needs access to new oil leases, and US (and thus the government) needs the energy. But BP screwed up by appearing to be trying to shift blame onto its partners and the government screwed up by not setting expectations and by making illogical or impractical demands of BP, and by making promises on which it can't make good.

On side note, it's interesting that states don't want the ferderal government to spend federal tax dollars - unless it's on their state. Bobby Jindal, a one time opponent of the stimulus bill, now appears quite keen (an understatement) to accept help from the feds.

OK. So much for the boring stuff; what about a conspiracy theory. Of all the oil majors, BP was the first to embrace the idea of a transition to greener non fossil fuel energy. Its new 'flower' logo and the name change from British Petroleum to BP reflected both its increasing internationalization (if I remember this was about the time it bought ARCO) and a vision of being about more than just petroleum products.

Suppose that in the aftermath of the Deepwater Horizon disaster, the government acts to transition more quickly than before towards green energy. Of all the majors, perhaps BP is best positioned to take advantage of such a shift in policy. So given the spill (and I'm not saying that BP blew up the rig intentionally; I am sure it genuinely was an accident), does BP benefit more, relative to its competitors, by making the oil industry look good or making it look irresponsible. Since I doubt that any legislation could be directed exclusively at a single company, all the oil majors would bear the same burden. And those companies with the greatest diversity and the least dependence on oil will profit the most.

"I'm just saying...", as they say.

Another Normal Accident

The explosion on the Deepwater Horizon, the oil platform leased by BP from Transocean, seems to be another or Chuck Perrow's "Normal Accidents" - that is something catastrophic that arises when a number of smaller errors and problems compound. According to Mike Williams, an electrical engineer on board the rig when the explosion occurred, who was interviewed on 60 Minutes, there were a variety of factors, "a series of mishaps", that together seem to have lead to the disaster. The drill string was raised accidentally while the BOP (blow-out preventer) was being tested, causing damage to the "annular" rubber seal at the top of the BOP. Chunks of rubber were coming up in the drilling fluid, but this dismissed as not "not out of the ordinary - not a big deal"; presumably, this may not have been and unusual occurrence. One of the BOP control pods was partly inoperative - but then there are two for exactly that reason and the second was thought to be operating properly. Because the drilling operation had been delayed by unexpectedly tough going , the project was 200% over time, and there was increase pressure to speed up the end of project. This lead to a decision to withdraw the mud from the bore hole before the third and final plug was set. Each element on its own was clearly serious, but not catastrophic; their albeit improbable coincidence was.

Walter Taub

Walter Taub was an actor, and my uncle. This picture I took in about 1976 I think in the house I grew up in in Sussex. He and his wife Luxi visited infrequently, largely a result of the authoritarian Czech Communist government. He was fairly well know in West Germany as well as in Czechoslovakia. I saw him only twice on the stage, once in Salzburg at tne Festival in about 1975, and once at the Burgtheater in Vienna in 1977. I understood little since my German is very rudimentary, but I do remember that he seemed to have what people refer to as "stage presence". When he made his first entrance, there was an expectant and appreciative murmuring from the audience. He waited it seemed several second before delivering his first line, heightening the anticipation. Perhaps it's a combination of notoriety and fame, and reacting to that - an unspoken dialog between actor and audience. He had a deep resonant voice - somewhere I've got a recording of his portrayal of Svek which I should try and find, if only to see it corresponds to my memory of his voice from 30 years ago. I do remember he had a good sense of humor, something that he must have needed: he signed Charter 77, a document critical of the Communist regime in Czechoslovakia. To suppress what the government saw as seditious activities, it degreed that while he could continue to act on the Czech stage, his name could never be mentioned in the program or the reviews; a pretty good way to squash someone's career. Things could have been worse, were it not for his friendship with Willi Brandt, the West German Chancellor. Brandt intervened on his behalf and as a result he was allowed to travel to the West (though hie wife, Luxi, could never accompany him outside the country). He was thus able to continue to work both in films and on the stage. Both Lux and Walter were in West Germany in 1968 when the Russians invaded Czechoslovakia. They could have remained there; Walter had plenty of work. But Lux was determined for many reasons to return home to Prague. I never asked Lux about this, but my father suspected this was a decision she later bitterly regretted. Prague, when I visited in 1977, was pretty gray but the architecture was wonderful. The underground was being build then and I remember huge stretches of Narodni with the water main and electrical conduits exhumed and elevated on a huge scaffolding that stretched for what seemed like miles. Despite the political restrictions and economic mis-management (not something it appears on which communist countries have a monopoly) that meant queuing for bread and meat, Lux, as something of a celebrity, had what in relative terms was a fairly easy life. She was well known at the best restaurants in the city, and dined out regularly. But I don't think she was very happy. By the early 1990s her memory was fading and in 1995 or 96 she was taken into hospital and I went to Prague for a couple of weeks to see her; she died the day before I was due to return to France.

Oral history

My father died about two and a bit years ago: he was 91. Although I had urged him to write about his experiences, he never really had the patience, and as a result much of the richness and variety of his experiences are now lost. Much as I was moved by Umberto Eco's depiction of the loss of knowledge when the monastery library burned, I am also perturbed that without some effort to record for future generations certain profound experiences, we will be collectively the poorer. Although no one reads this bolg (that's fine - my reason for writing it doesn't include seeking fame on the internet), I hope that in some way serves as a contribution to an electronic 'oral history'.

At the end of the war in Europe (WWII), my father led a small team of doctors from Britain back into Czechoslovakia. Since there was still fighting on the Eastern front, their route was strangely circuitous. One of the team's first tasks was go into the Nazi concentration camp at Terezin. The team had brought DDT, a pesticide that was in short supply in Europe. Their task was to rid the camp of potential carriers of disease.

My father was somewhat circumspect about what they found - they were the first people to enter the camp after the Nazi's left - but I do remember him recounting the unspeakable sight of mountainous piles of human bodies. The corpses were so emaciated they assumed all were dead, but to their horror, found that some (many) were actually alive. They spent several days sorting the dead from the living and spraying them with DDT. What happened after the first day or two, I'm not sure. My memory and my father's telling of the story (which I last heard perhaps 25 years ago) are unfortunately fading from memory.

Ubuntu 10.04 LTS and VMware Player 3.0.1

After three days it's pretty clear these two don't really play well together. Interestingly I'm having the same kinds of problems as I had a while back with VMware Workstation 6.5.3 and Fedora 9: the mouse and keyboard freeze without warning, ssh from other machines is blocked and there is no alternative but a reset. I read somewhere that 10.04 has done away with HAL and I have a sneaking suspicion that Fedora did the same with Core 9. Perhaps that's the issue...

Who knows. I don't have time to look into it. If I was a good citizen I might soldier on and submit bug reports but since I have already wasted far too long recovering from the mess the crash during the 10.04 install made of my RAID array, I'm not feeling all that charitable. So everything is being backed up (twice) and I'm going back to 9.10 which has worked well for six months. Perhaps in year when the bugs are worked out I'll give 10.04 another try but frankly I'm loosing my appetite for the latest if this amount of faffing around is the cost.

This of course is the side of the open source development model that commercial software companies stress though I had studiously turned a blind eye to. There is no direct accountability for poorly written products, and so what one gains with hundreds or thousands of beta testers, one looses in a lack of discipline, control and accountability. Swings and roundabouts, and there is no free lunch.

Not so Safecopybackup

Last year I moved my offsite backup service From Mozy to Safecopybackup.com. In hindsight, a very very poor decision. Here's why:

1. While the client application appears to be tracking files in real time, and even tells you that backups are complete, they're not. A few files (and yes they were ones I wanted) I modified yesterday hadn't been backed up for three weeks.
2. Recovering data from a complete drive failure is nearly impossible because there is a limit of 200MB in the files that the web interface will allow you to zip and download.
3. Don't think that the Windows webDAV will allow you to get around this restriction; it collapses regularly and stops when files are "not found" on the server which happens frequently.
4. Files are sometimes restored with garbled file names.

So while you can see your data on their website, for all practical purposes it's not recoverable, unless you have a month and want to download files one at a time.

What I imagine most people look at when considering a backup service is the security of the store, the easy of setup and the simplicity and regularity with which files are saved. What I certainly didn't look at was what happens when you have a large volume of data to restore. Even when I had to restore the odd file here and there, something that bought me into direct contact with the restore function, the penny still didn't drop. Having spent almost a week restoring files piecemeal, it's not something I'm likely to forget any time soon.

Some thoughts about software RAID

Last week I upgraded from Ubuntu 9.10 to 10.04 LTS. I will leave it to you to decide whether this was such a good idea: But now it's done.

Somewhere in the process the system stalled (X-windows froze and there was also no way to ssh in from outside). The hard restart cause mdadm to see both my RAID10 and RIAD5 arrays as missing two of the four drives. The array wouldn't assemble. Then I may have done something stupid. I have in the past been able to recreate the array using mdadm --create. It's almost certainly not the right way to deal with the problem but experience as March notes, can be a bad teacher, and is has worked on several occasions.

When the arrays restarted the ext3 file systems on both /dev/md0 (Raid10) and /dev/md1 (Raid5) were completely destroyed.

I have three takeaways from this episode:
1) I should probably have tried to learn about mdadm recovery procedures, but there's not enough time in the world to do everything one ought, and I doubt that I will do so even now.
2) What is the weakest link in the system? The hardware (specifically the disks which haven't failed in 3 years) or the software of which mdadm is a part which has failed so many times I've lost count. If the goal is not just the illusion of security RAID offers but actual security, one has to wonder whether it's really such a panacea for somewhat naive end users like me.
3) When choosing a backup service, ask not how fast it can back stuff up but rather how quickly and easily you can restore your data.

Why I'm becoming a Luddite

"If there's a problem, technology will solve it". For much of the 20th century this has been the mantra, applied reflexively and unquestioningly. Technology, at least the bits I interact with, does have real benefits. My home is warm and well lit. I have access to more information than I know what to do with all without getting up from my desk. The list of benefits is long. So why am I worried?

This evening, I caught part of a broadcast of a panel discussion at the World Economic Forum in Davos (here is a summary). The Session PanelistsBill Gates, Co-Chair, Bill & Melinda Gates Foundation, Jakaya Kikwete, President of Tanzania, Ellen Kullman, Chair of the Board and Chief Executive Officer of DuPont, Nguyen Tan Dung, Prime Minister of Vietnam, Ngozi Okonjo-Iweala, Managing Director of the World Bank, Patricia Woertz, Chairman, President and Chief Executive Officer of Archer Daniels Midland. An impressive panel.

But I came away with serious misgivings. The three Americans all expressed a deep-rooted faith that technological innovation was the solution. Interestingly this was not a view shared by the Prime Minister of Vietnam, the President of Tanzania, and the Managing Director of the World Bank who suggested that there were other approaches that should be perused, including fixing distribution infrastructure and local markets, irritation, fertilizers and simple mechanization.

The kind of innovation Woertz Kullman and Gates were talking about is the engineering of new strains of crops that are higher yielding and pest and drought resistant. This sounds good but the economic model is less appealing (unless you are a ADM). In order to pay for this crop research, ADM has to ensure that farmers come back every year for seeds, which they will because as a by-product of some of the crop 'engineering', the plants are sterile. This has two implications. First Farmers, who have faced increasingly concentrated and powerful buyers, now increasingly must deal with powerful suppliers too. So from an industry analysis perspective, framers are in what Porter would term a relative unattractive industry (there are many of them and low barriers to entry).

But while I sympathize, that's not my main concern. What really worries me is complexity, interdependence, and tight coupling: all of which, while making the system of agricultural production more efficient, also make it more fragile and likely to fail catastrophically. Charles Perrow coined the term "Normal Accidents": we've just seen one in the financial crisis of 2008/9. entities become more interdependent and inter-linkages proliferate, perturbations that once would have caused small local ripples now generate cascading tidal waves. A full scale depression was averted with the injection of large amounts of money. But money can be found fairly quickly - food on the other hand, cannot.

Imagine a day when all the food on the planet comes from one or two 'high yield' strains, for example one strain of wheat and one of corn, those that are the most profitable for ADM to sell and for farmers to buy. Then something happens. "What can possibly go wrong?" technophiles ask. "After all, these strains have been vetted and passed by the FDA, and have been used for years without a problem". To which my reaction is "how did we get antibiotic resistant strains of staphylococcus"? We didn't see that one coming. Neither will we be able to anticipate the ways in which these strains on which we will all be dependent will fail; but they will, and since they are all genetically identical the entire system will fail at once. Injections of cash will not help if the raw material for the food manufacturing system dries up. It will ripple through the entire industrial food value chain; livestock will die, food prices will soar, staples will disappear. The Irish Potato Famine of 1845, but this time on a world-wide scale.

It may be 50 or 100 years before we get to that point: but if we continue on our current trajectory driving out requisite variety and building complex interdependent systems in the pursuit of efficiency, I predict there will be a "normal accident", and one of a scale that could do for us what a large meteor did for the dinosaurs.

One week on...

VMWare player

Shortly after my last operating system related post, I decided to try VMware Player instead of VMware Workstation. Not quire sure what prompted this but when I found that the latest version of Player (3.0) allows the creating of new guests, much of the rationale for Workstation disappeared. Six weeks later, Player is still in place. No hitches at all. Amazing.

Throw-back to the stone age

Google in China - Reprise

Anne Lawrence, in a fascinating case study, drew attention to the dilemma Google faced with involvement with Chinese language search and its decision to place servers in China. This posting suggests Google is reconsidering its position.

VMware Player

I have been a (relatively) happy VMware Workstation user for several years now. I was interested when VMware announced Workstation version 7 - until I discovered that there was a significant upgrade charge.

Since VMware Player 3.0 was announced at about the same time, I thought I'd see what Player was like: and that was three weeks ago and I haven't used Workstation since.

I'm not a real techie - I just want something that will run Linux and Windows side by side, and to be able to test new OS releases before migrating to them permanently. Sun's VBox, which I looked at and rejected as feeling like a bit of a toy a year ago, hasn't really improved as far as I could see. Player, in contrast, seems a serious piece of software. VMs seem to suspend and restart faster than with Workstation 6.5.2, and the switching bar at the top of the screen takes less space which I like. I've not found anything I need to do that I could with Workstation but can't with Player. Since version 3 supports the creation of new VMs, for people like me who simply want to have multiple OSs on one box, Player seems viable - and it's a no charge item.

Many years ago - here I'm thinking about about mainframes and IBM's VM product as much as desktop VM solutions - a big question was overhead; how much of performance hit does one take when running an operating system as a guest in a VM? With the falling price of memory and processing, the question, at least for much of my work (document reading / editing, and a small amount of C++ and PHP), has become largely moot. Only rarely do I notice any performance issues. So, thank-you VMware; sometimes technology does make our lives a bit easier.