User:Jan Hidders/HTML-free mark-up

See also: m:Wikitax

---

For reasons I won't go into here but feel very strongly about I think that the Wikipedia mark-up language should be completely HTML-free and not even contain HTML-like mark-up. Looking at Wiki's with a similar approach like MoinMoin and PhpWiki I tried to come up with a suggestion to replace the current HTML-like tags.

  < i > , < b > , < h1 >, ..., < h5 >, < hr >
  ... standard WikiWiki mark-up ... 

  < br >
  \\

  < center > In the middle < /center >
  >> In the Middle <<

  < font color = "red" > Red font < /font >
  ... depreciated ...

  < tt > typewriter font < /tt >
  {{ typewriter font }}

  
 preformatted text < /pre >
  ... start line with a space ...

  < strike > strike out < /strike >
  --strike out--  (Note: -- has to be followed/preceded by a non-whitespace char.)

  < ul > underline < /ul >
  __underline__

  < sub > subscript < /sub >
  x_subscript  (subscript continues until next whitespace, _ or ^)

  < sup > superscript < /sup >
  x^superscript (continues as with subscript)

  < table > .... < /table >
  [| This is a table with a single cell on the first line    |]
  [|   two cells           | on this line                    |]
  [|   two cells on        | this line too                   |]

See also the discussion at [[m:Wiki markup tables]].

  < var > x < /var >
  $$x$$
  Remark: perhaps sub/superscript should be limited to this,
          maybe even use LaTeX syntax?

-- [[user:Jan Hidders|Jan Hidders]]

Some comments:
* < br > are usually used where they are not really needed.
* subscripts and superscripts are usually used in math context.
* Colors are very rarely needed.
* Tags of form < TAG > and < TAG > < / TAG > are not problem for grammar. Tags with complex arguments are. There's no need to remove them and certainly no need to replace them with even weirdier markup.
* This table syntax is bad. It doesn't allow things like lists and multiple paragraphs in table and it would be really hard to write big tables in this style.
--[[Taw]]

: You are certainly right about the table syntax. I have now given an example of the PikiePikie syntax which is more expressive. About the sub and superscripts; I wrote some mathematics here and it irritates me endlessly to write < sub > and < super > (esp. since I'm used to the TeX notation) and it makes formulas harder to read. You are also right that < tag > is not really a problem for writing an easy syntax or even a big parsing problem, but I'd stil rather only use them if there is really not good Wiki alternative. -- [[User:Jan Hidders|Jan Hidders]] 07:11 Jul 26, 2002 (PDT)

Personally, I think using slashes surrounding anything (ex: /foo/ ) for markup is a Big Bad Idea, particularly in articles on computing (think UNIX pathnames) and linguistics (phonemic transcriptions are enclosed in slashes). I've spent enough time cleaning up after phonemes that were munged by the subpage conversion, I don't want to add to that. Other than that, could be interesting. --[[User:Brion VIBBER|Brion VIBBER]]

: You are of course completely right. I've changed that now. -- [[User:Jan Hidders|Jan Hidders]] 07:11 Jul 26, 2002 (PDT)

----

Boy, it's a lot easier to remember <code><strike></code> than <code>---</code>!
Removing HTML is supposed to be user ''friendly''???
— [[User:Toby Bartels|Toby]] 06:11 Jul 27, 2002 (PDT)

: I agree, but < strike > isn't really used that often. About a dozen pages or so, the last time I did a check on the database, and then mostly for Planck's constant. [[User:Jan Hidders|Jan Hidders]] 02:05 Jul 30, 2002 (PDT)

Ugh — in reaction to having to use <code><strike></code> to get hbar.
Anyway, my objection is to inventing a bunch of new arcane codes to get of rid of HTML that is at least as easy to understand.
<code>'''</code> is great, because it's easier to type even than <code><b></code> and is used so often.
But I'm against making up more and more markup symbols for rarely used items just to avoid HTML on principle.

----

Don't forget <code>$$x$$</code> → <code><var>x</var></code>.
Use lynx to compare math pages that I've edited with ones that I haven't to see why <code><var></code> is correct and <code><i></code> is wrong.
— [[User:Toby Bartels|Toby]] 06:11 Jul 27, 2002 (PDT)

: Interesting suggestion, I've added it. In fact I had that mark-up in mind in case we decide to integrate LaTeX as Axel Boldt suggested. <BR> I'm not sure if I completely agree that it is better to use < var > than < i >. I know that lynx usually underlines to indicates italics and leaves the < var > alone, but for example on my Lynx italics are indicated by a Magenta font. So I prefer variables italicized because then I see a difference in Lynx between normal letters and variable names. Also don't forget that this is what is done mostly at the moment, so that creates a certain expectation pattern that you are now breaking. [[User:Jan Hidders|Jan Hidders]] 02:05 Jul 30, 2002 (PDT)

Unfortunately, the <code>$$</code> idea is only suggestive of (La)TeX, not directly compatible with it.
However, it would make the math pages that I work on much easier to edit.

As for <code><i></code>:
Aside from the fact that <code>''</code> conflicts with primes occasionally, making an explicit <code><i></code> (or whatever) necessary, I definitely think that <code><var></code> is the appropriate HTML item to use.
If you wish, you can tell lynx to render <code><var></code> and <code><i></code> the same way, but I can't tell it to render <code><i></code> differently in some situations than in others.
Yes, this makes some math pages different from others, which just means that I need to edit more math pages — having done it wrong in the past is no reason to do it wrong in the future.

Ultimately, the only good solution here is to get <code>$$</code> (which I was thinking of before I ever saw this page) implemented.
To my mind, this is a very different situation from proposing <code>---</code> for <code><strike></code> — dollar signs aren't used for any other wiki code, the legend "var" isn't very clear to the uninitiated, and at least one person ^_^ would use it a lot.
In these respects, it is more like <code>'''</code>, which I've always approved of.
— [[User:Toby Bartels|Toby]] 05:50 Jul 30, 2002 (PDT)

--------

* A line begininning with whitespace does PRE
* "\\" for BR was a patch for UsemodWiki -- although BR shouldn't really be necessary: use a paragraph break in body text and bullet points for lists
* Font colours shouldn't be used either: use "==" etc for headings. Very few writers here use headings -- they make a single line bold with ' ' ' . Making proper headings is better for CSS.
* The sub- & super- tags are a nice idea. Maths expressions are horrible in raw format, but I'm not sure if those tags would substantially help though.

Over at UnrealWiki we're running a new Perl OO markup parser, and we've introduced some new markup. In particular support for nested lists of different types and tables: http://wiki.beyondunreal.com/wiki/Wiki_Markup -- [[User:Tarquin|Tarquin]] 08:07 Jul 27, 2002 (PDT)

:That syntax for mixed nested lists is actually already implemented, but I don't think it's documented. I'll go add that... --[[User:Brion VIBBER|Brion VIBBER]]