[Bioperl-l] Quote

Chris Fields cjfields at uiuc.edu
Sat Jan 13 20:00:42 EST 2007


Not to wax philosophical on a very off-topic issue (but it is the  
weekend)...

I would also argue that software itself isn't to blame.  Software,  
particularly open-source software, is only as good as the people  
involved with its development, either the actual developers or users  
who contribute back in some way (filing bugs, making suggestions,  
etc).  People aren't perfect, so why expect software to be?  To make  
a completely lame analogy : if you sat on a faulty chair which  
collapsed, would you blame the chair or the carpenter?

As for making sure nothing is missed or is defect-free, how can one  
prove a negative?  Of course something will be missed, or a defect  
eventually found.  People are still finding new and exciting things  
(riboswitches, epigenetic regulatory mechanisms, etc) years after  
genomes have been completed and released.  A huge number of predicted  
proteins have no known or characterized function.  Security holes  
have been consistently found (and patched and sometimes repatched) in  
some of the best OS's out there.  Something changes beyond the  
control of a developer (a sequence format, or a server change) thus  
causing a bug in the software.

Whatever you do, don't confuse success with perfection.  Success at  
least is attainable; perfection, meh, not so much.

Sorry you had frustrations with something along the way.

chris

On Jan 13, 2007, at 3:24 PM, Stephen Gordon Lenk wrote:

> Hi,
>
> I'm a bit puzzled about what "actually pretty good" means. My
> understanding about Mars Rover is:
>
> "The current Mars rovers may rely on proven computer technology, but
> for Spirit the journey has not been glitch-free.
>
> After a promising start to its mission, the Spirit rover -- the first
> of the MER twins to land on Mars -- stopped sending proper data to JPL
> scientists 18 days into the mission and later baffled ground
> controllers by rebooting itself over and over again. Since then,
> mission controllers were able to regain reliable communications with
> the rover and continue to study what may have caused the malfunction."
> [http://www.space.com/businesstechnology/technology/ 
> mer_computer_040128
> .html]
>
> Most industries know that "rebooting itself over and over again" is
> not desirable. Evidently Mars is out of sight - out of mind. Software
> developers have excessive pride in deliverables that are all too often
> grossly dysfunctional. If a product released to the market failed
> by "rebooting itself over and over again," there would be no end of
> deserved criticism. Why is software immune from reasonable scrutiny?
> Who believes their community is immune from the release of defects?
> Maybe we should all ask ourselves how long it took for defects to be
> found after the release of our last piece of wonderware, whatever it
> was.
>
> As far a Google or genome software - how do you know you haven't
> missed something in ALL the web pages of the world or in a huge
> database being searched heuristicly. Are you saying defect free -
> never misses anything - perfect? Or just darn good - if so, how good,
> and how is that determined? Is that aspect of quality openly measured,
> quantified, and available or is it just brushed under the rug.
>
> I do not claim to be flawless. This is *-->>NOT<--* a screed against
> Osborne/Stroustrop or anyone else who takes pride in a job well done.
> I am simply not convinced that the software community is as defect
> free as they claim. Sorry to be 'aggressive' (if this be such) but I
> am completely sick of defective software propelled crap.
>
> Steve
>
>
> ----- Original Message -----
> From: Brian Osborne <bosborne11 at verizon.net>
> Date: Saturday, January 13, 2007 1:06 pm
> Subject: [Bioperl-l] Quote
>
>> Technology Review January/February 2007 ³Bjarne Stroustrup The
>> Problem with
>> Programming²
>>
>> TR: Why is most software so bad?
>>
>> Bjarne Stroustrup: Some software is actually pretty good, by any
>> standard.Think of the Mars Rover, Google, and the Human Genome
>> Project. Now, that¹s
>> quality software!

Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign






More information about the Bioperl-l mailing list