Linus Torvalds Outburst Sparks Fierce Debate: Does Open Source Software Need Specs?

[David Farber <dave () farber net>]

Begin forwarded message:

From: Tom Fairlie <tfairlie () frontiernet net>
Date: October 21, 2005 1:26:01 PM EDT
To: dave () farber net
Subject: Re: [IP] Linus Torvalds Outburst Sparks Fierce Debate: Does  
Open Source Software Need Specs?


Mental note #1: Never hire Linus Torvalds to work on any
software system that I might someday want to sell or support.

Mental note #2: Never post to a message board after a night
of insufficient sleep or a morning of insufficient caffeine.

Tom Fairlie

----- Original Message -----
From: "David Farber" <dave () farber net>
To: "Ip Ip" <ip () v2 listbox com>
Sent: Friday, October 21, 2005 12:12 PM
Subject: [IP] Linus Torvalds Outburst Sparks Fierce Debate: Does Open  
Source
Software Need Specs?





> TechNews wrote:
>
> "Linus Torvalds Outburst Sparks Fierce Debate: Does Open Source
> Software
> Need Specs?"
> SYS-CON (10/03/05)
>
> In a recent posting on the Linux Kernel Mailing List, Linus Torvalds
> blasted specs as a method for developing software, claiming they
> pay more
> attention to theory than reality and introduce a needless level of
> abstraction that often falls flat when put into practice. Torvalds
> cited
> the OSI network model protocols, which he characterized as "classic
> spec-design, which had absolutely zero relevance for the real
> world." He
> admitted that specs could serve as a useful framework for
> discussion, but
> that they should have no bearing on the actual design. Feedback to
> Torvalds' posting was varied, as some immediately took him to task,
> insisting that specs, like actual code, were by definition
> approximations,
> while others supported his view that their use is limited. One
> respondent
> noted that many original Unix specs were quite good, while later
> standards
> such as HTML and SOAP were well-written, but ultimately functioned
> poorly
> due to sloppy designs. One post criticized Torvalds for being too
> narrow
in


> his definition, claiming that his view of a spec as an unchanging
> template
> incapable of adapting throughout the development process is not
> held by
> anyone in the industry. Another reader claimed specs are essential to
> commercial projects, as they serve essentially as a contract,
> defining the
> scope and function of a given body of code. By way of mediation, one
> respondent sought to clarify Torvalds' position by arguing against the
> misconception that programs can be written through the progressive
revision


> of specifications, when really specs are useful insofar as they can
> guide
a


> discussion, but must yield to technical requirements as the
> development
> process unfolds.



-------------------------------------
You are subscribed as tfairlie () frontiernet net
To manage your subscription, go to
  http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting- 
people/





-------------------------------------
You are subscribed as lists-ip () insecure org
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/