[LINK] Carbon Neutral Linux Computers, Canberra, 21 November 2007

Richard Chirgwin rchirgwin at ozemail.com.au
Wed Nov 21 10:27:51 AEDT 2007

Rick Welykochy wrote:
> Tom Worthington wrote:
>> Topic: Carbon Neutral Linux Computers
> [SNIP]
>> A study sponsored by the Australian Computer Society has shown that 
>> computers and telecommunications equipment in Australia generated 
>> 7.94Mt of carbon dioxide in 2005, 1.52% of national emissions. The 
>> ACS issued a Policy Statement for Green ICT, which includes 
>> suggestions on initiatives ICT professionals, government, consumers 
>> and ICT manufacturers can take to help reduce carbon dioxide 
>> emissions attributable to the use of ICT equipment. Tom Worthington, 
>> chair of the ACS Green IT Group, will discuss options available and 
>> demonstrate a new low power "thin client" computer using a "carbon 
>> neutral" processor.
> One wonders if and by how much the use of computer technology has offset
> and/or saved on carbon emissions in other sectors. Without the use of
> computer automation, bookkeeping, monitoring, control etc. would there
> be comparatively more emissions in industry and at the domestic level?
> Surely this must be the case, since microprocessors are being deployed
> in more and more consumer products to provide "greener" white goods, to
> monitor and advise on power useage, etc. 
...there's a gap between the "greenwashing" that goes into marketing 
materials and the realities of implementation. Microprocessors are 
deployed into white goods chiefly to get rid of electromechanical 
controls (eg the knob on the washing machine). And this is purely a 
bottom-line decision: the microcontroller is cheaper than the 
electromechanical control, and you get to bulldust the consumer into 
paying more for the microprocessor version.


The gotcha is that now the user is hostage to the decisions of the 
software engineer, and my experience is that software engineers are not 
washing-machine users. I am about to return from a 
microprocessor-controlled washer to one with an electromechanical 
control, because the software results in unbelievable waste of water and 
power through one design flaw: should the machine unbalance, the 
software returns it to the beginning of the preceding cycle. The impact: 
if anything goes wrong with the cycle, the machine steps backwards and 
refills. It's a truly awful outcome, since you either have to attend the 
washing machine, or it wastes both water and electricity.

Whereas an electromechanical machine simply halts and restarts.

Extrapolation: we need a new suite of consumer product tests in which 
bad software decisions are tested for their efficiency impacts...

I can believe that a refrigerator may become more electricity efficient; 
but looking around whitegoods places, I find that somehow, strangely, 
the number of "stars" achieved by new fridges differ not so much from 
older ones. So what's happening is that the processors and software are 
used to do some heavy lifting in efficiency and (IMO) letting the 
manufacturers off the hook on other design decisions.

> How does one quantify the effect
> of leveraging emissions caused by using computers against not using them
> in the first place?
> That is not to say that we can be flippant and ignore the emissions 
> caused
> by computer technology. Your typical desktop system nowadays has "super
> computer power" and could easily support a dozen thin clients and still
> have breathing space.
> One of my clients is looking into replacing full blown desktops in 
> smaller
> libraries (used for OPAC catalogue in the main) with a single host and
> perhaps six thin clients.
> cheers
> rickw

More information about the Link mailing list