On Fri, Apr 25, 2008 at 5:49 PM, Ian McWilliam <[EMAIL PROTECTED]> wrote:
>  Can anybody explain why is it acceptable to modify a "standard" for "ports"
> but not not for "base"?

I'm going to guess that the core reason is "what helps more users?":

- A pdf spec that's written to sell the illusion that you still have
control over data that you've given to someone else? That's not
helpful, may even be harmful. Insert bad-crypto vs. no-crypto
argument...

- A pdf reader that gets in the way of you accessing content or a pdf
reader that lets you get your job done with the minimum of hassle? The
latter. I'm going to assume that if you're looking at a PDF you have
at least some license to access it. At least the password protected
pdfs make you work for it if you're trying to read a pdf without
permission. A "non-printable" pdf can be trivially screencapped and
printed... more illusions for sale. And what if i'm writing a driver
based on something in a de-permitted pdf? How is the magic "0x8000" I
typed any different from the "0x8000" I could've just copied from the
datasheet?

- An ssh implementation that tries to avoid known (possibly
security-related) bugs or an ssh implementation that just hopes the
other guy got it right too. The former is more helpful. Maybe you'll
be lucky and negotiate secure settings, maybe you'll be slightly
unlucky and fail to connect or maybe you'll be very unlucky and
negotiate "cipher none".

Also, (and this may seem a little shallow) the original implementers
(ssh, xpdf, whatever else) probably try to adhere to the spec as
closely as they can. They produce a program with predictable behaviour
- that's helpful. Then they give away the source in hopes that someone
finds it useful... after that, there's nothing stopping their users
who've legitimately acquired the source from saying "this is 99.9999%
of what i need, i just need one more tweak" and hacking it up to suit
their own ends.

CK

-- 
GDB has a 'break' feature; why doesn't it have 'fix' too?

Reply via email to