[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Documentation is/is not software [was: NEW ...]

Matthew Palmer wrote:

>On Tue, Mar 22, 2005 at 12:32:30PM +0000, Matthew Wilcox wrote:
>>On Tue, Mar 22, 2005 at 09:06:19AM -0300, Humberto Massa wrote:
>>>And I believe that the Vancouver proposal, if implemented as intended
>>>up to now, will not only affect what Debian really *is*, but in some
>>>ways will *destroy* what Debian is.
>>Debian has already decided to destroy what it is by giving in to the
>>crackpots who insist that everything is software.
>Way to set the tone for a productive debate.

Yeah, we are seeing a lot of this lately.

>At any rate, the problem with trying to treat different types of
>bitstreams differently is to classify them, and identify a different
>set of freedoms which are appropriate -- and, more pretinently, why
>those different set of freedoms is important.  The "crackpots" won more
>or less by default, because nobody was able to come up with either of
>these two pre-requisites.  This suggests to me that either (a) it can't
>actually be done, in which case the "crackpots" are, after all, right;
>or (b) Debian is so filled with "crackpots" that there is nobody who
>actually wants to see documentation treated differently to executable

IMHO the problem is that there is not a clear distinction. Period. Why?
Because source code *is* documentation. The set of freedoms we want to
Free Software (AFAICT) is: freedom to study, modify... for all this we
need access to the documentation, part of which is the source code.

>I used to sit in the "documentation requires different freedoms" camp,
>but eventually just couldn't support my feelings with logical argument.
>But there are significantly more powerful minds than mine out there; I
>look forward to hearing their arguments in favour of different freedoms
>for documentation.

The problem with hearing arguments in favour of different freedoms for
documentation is that people will have to define what is -- and what is
not -- documentation. And I don't really think this is possible.

One example: are Debian-changelogs documentation? They contain
instructions on what version of a package is to be built, and which
debbugs should be closed...

>If someone can come up with a bright-line test for differentiating
>executable materials and documentation, or executable materials and say
>firmware, and can produce a "DFDocumentationG" or "DFFirmwareG" with
>effective reasoning, I will be most impressed, and will most likely
>support their position.  Until then, however, I am firmly in the "all
>things we ship are software, and the DFSG applies to all of that" camp.
>- Matt crackpot and proud


Reply to: