[Pmwiki-users] More on attachments

[Patrick R. Michaud]

> From john.rankin at affinity.co.nz Wed Nov 20 22:28:50 2002
> 
> I agree, but ...
> I'm not sure it's a big deal. [...]
> The key question which I think you are suggesting is: "Would a naive 
> author use this feature?" 

Nope, that's not my key question at all.  My key question is and always
has been "what markup characters will make sense and look reasonable for 
style-based markup?"  So far the primary suggestions have been {{...}} and 
%style%...%% -- neither of which I'm terribly fond of.

I believe that part of what makes wiki accessible to newbie authors
is that a document written in wiki markup still has some basic resemblance
to what it will look like when displayed as a web page.  This is in sharp
contrast to HTML where the markup document often looks nothing like
the rendered document.  Furthermore, wiki markup is usually still readable 
as a document (with the possible exception of my double bracket abuses),
whereas it's very difficult to "read" an HTML document because the markup
tags greatly interrupt the flow of reading.

You don't have to convince me that it would be easier, better, and more
flexible if markup customizations can be made within a wiki document
instead of a configuration script.  I've always known that, and I even 
have a fairly good idea of how I'll implement it once I have a markup
I'm comfortable with.  The ONLY reason I haven't implemented colored
text or other style attributes directly into PmWiki is that I haven't 
found markup characters that I feel are "clean" enough to preserve the
readability of markup text so that newbie authors aren't discouraged
from contributing.  As soon as someone can propose a markup that
I think works, or convince me that one of the existing proposed markups
is in fact "good enough", the whole thing will be implemented because I
have a lot of sites that could use this feature also.

It's as I've said many times before--the hardest part of incorporating
new features into wiki is coming up with markup to support it that makes
sense.

Pm


P.S.:  On a related note, I have long believed that one of HTML's 
greatest shortcomings is that it didn't provide the capability for
authors to define new "content-based" tags in terms of existing 
primitive "rendering-based" ones.  CSS starts to give authors more 
control over the rendering, but it still limits the author to the 
predefined HTML tags (and introduces yet another language syntax 
in the process).  With XML we finally have the capability to define
new tags, but now the author has to know yet another language (the DTD
language) and yet another set of syntax rules.

It just seems silly to me that an author has to learn multiple sets
of widely-different syntax rules just to adapt HTML to a particular
problem space.  In languages such as C++, Java, and Perl, I can use 
the existing built-in primitives of the language to build new 
constructs that better fit the problem space I'm working in.  And
this is not limited to programming languages--document description language 
such as TeX and even troff have the capability to add new constructs
to the language from existing primitives.  So somehow I find it very 
disappointing that the W3C and related organizations haven't
been able to do something similar for the web (although there are now
beginning to be some tentative efforts in this general direction).