RE: Standards for developing secure software

From: Glynn Clements (glynn.clementsat_private)
Date: Thu Jan 23 2003 - 19:42:44 PST

  • Next message: Andre Mariën: "Re: PGP scripting..."

    Gustaf Bjorksten wrote:
    
    > i disagree. The programmers i work with are well aware that they
    > /should/ be programming with a 'secure' mindset, but they just have no
    > idea where to start. The topic is outside their sphere of knowledge.
    
    Here's my suggestion for where to start:
    
    	Don't assume that It Won't Happen
    
    If someone can obtain a benefit by Making It Happen, then there's a
    good chance that It Will Happen.
    
    What will happen if the string is too long? What if it contains
    "unusual" characters? What if the integer is negative, or zero?
    
    Simply asking oneself these questions certainly won't prevent all
    vulnerabilities. But several years of reading BugTraq suggest that it
    would have prevented a significant number of them.
    
    Even if you don't achieve perfection, the extent to which you fail may
    affect your reputation, at least amongst those who do understand the
    issues. Failing in the presence of an obscure signal-handling race may
    be viewed as a simple mistake. But if your software segfaults at
    address 0x41414141 when someone pastes 64K of 'A's into a text field,
    you may as well use "we don't give a damn about security" as your
    corporate slogan.
    
    -- 
    Glynn Clements <glynn.clementsat_private>
    



    This archive was generated by hypermail 2b30 : Fri Jan 24 2003 - 10:59:07 PST