begin oe_protect.scr
Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> espoused:
> __/ [ Jim Richardson ] on Thursday 29 June 2006 07:18 \__
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On Wed, 28 Jun 2006 11:14:22 -0000,
>> Tim Smith <reply_in_group@xxxxxxxxxxxxxxxx> wrote:
>>> In article <7or9n3-5jp.ln1@xxxxxxxxxxxxxxxxxxxxxx>, Mark Kent wrote:
>>>> Source-code, by its nature, is a superior standards-writing language than
>>>> English (or French or Spanish), as it can be compiled directly using
>>>> compilers onto multiple, different, platforms.
>>>
>>> Using source-code as a standards-writing language will mingle the standard
>>> with the implementation. Someone looking at the code won't know what
>>> things were simply implementation decisions, and what things are actually
>>> required by the standard.
>>>
>>> Source code is useful as part of a standard to provide examples of how
>>> particular parts of the standard *could* be implemented.
>>>
>>
>> Source code can tell you what the code does, not neccessarily what the
>> coder *meant* to do.
>
> To rephrase that (I hope you don't mind), source code tells you what a
> _binary_ file does, which tells you what the coder intends to do on your
> machine (e.g. does s/he phone home?).
>
Sorry - that's not right. Source code tells a compiler what it wants the
compiler to create, source is /only/ a description language, with tight
syntax, limited semantics and precisely defined grammar.
The binary will do whatever the compiler made from the source, taking
into account any errors in the compiler, as well as errors in the
intention or grammar/syntax in the source.
This is why source code is the proper medium in which to write
standards. If you like, I could bore you with the several years I spent
writing standards to explain how I know, but...
--
| Mark Kent -- mark at ellandroad dot demon dot co dot uk |
"Your mother was a hamster, and your father smelt of elderberries!"
-- Monty Python and the Holy Grail
|
|
