RE: The <blockcode> and <l> elements

Lachlan Hunt:
> Why not use:
> <blockcode>
>     <l class="h">&lt;h&gt;\chapter{introduction}&lt;/h&gt;</l>
>     <l class="p">&lt;p&gt;This is really interesting 
> stuff.&lt;/p&gt;</l>
> </blockcode>

That looks very much like using a generic <tag class=...> 
for everything.

>> For instance, in Java, I could want to use the class 
>> declaration as a blockcode heading:

>   Semantically, I don't think this would be a heading.  

I do; when I'm trying to understand code, the start of a
new class or function is a very important semantic 
breakpoint, and the name (or even full signature) of
the function is the best header.

Take a look at the online documenation for java, or
the PalmOS API.  A new public function (or class,
or datatype, or defined constant) will appear in the
table of contents, will trigger a new section, and will
be the heading of that section.

> The bottom line is that all code is plain text. 

This may be part of the disagreement.

To me, there is a fundamental (semantic) difference
between each of 

(1)  Comments  -- strictly for humans

(2)  Doc-strings -- also for the machine => more 
constraints on what should/must/must not appear.

(3)  Data -- affects the program output, but can be
swapped out without changing the logic.  Changing
it should require much less in the way of regression
testing.

(4)  Executable code -- actually does something

Yes, there are border cases such as small initializers 
or constant output strings.  The same is true in much
of life; I couldn't tell you the exact hundredth of a 
degree where "fever" starts, but doctors still treat 
"has a fever" as useful diagnostic information.

-jJ

Received on Friday, 7 November 2003 10:31:26 UTC