Available Wiki Systems
[edit this web-page]
This file purports to contain information about 'wiki' systems which
are available on the Internet. A wiki is a system which allows all
visitors to a web-site to edit the pages, and create new pages, for
that site. The editing and creation process is normally done through a web
HTML form based interface
Also included here is a section on 'plain text markup languages'. These are
sytems which allow the writer to use some very basic structures in order to
dictate how the text file will appear or be interpreted when it is
transformed by a particular piece of software. The section on plain text
markup languages is included here because almost all 'wikis' use some kind
of plain text markup language in order to allow the user to edit the HTML
pages. This is designed to achieve two things: the web-visitor does not
need to know HTML, and the web-visitor does not need to install any piece
of software in order to edit the web-pages (such as, for example, a wysiwyg
HTML editor) since the normal HTML textarea form component can be used by
the web-visitor to make changes to the existing web-page. This assumes, of
course, that there is a piece of software running on the WebServer which is
capable of transforming the plain text which the web-visitor creates or
changes into an HTML format.
This is my idea of a better plain-text structuring. It contains as little
actual markup as possible, and the markup which it does contain is unobtrusive.
Allow the user to down load the source
file in various formats: docbook, text, html. Have a 'view source' link to
allow the user to see the plaintext behind and html doc.
Have an [FAQ] tag which allows faq sections to be formatted properly.
Have some symbols after the page title to determine things like, is a
table of contents generated, numbering of headings, etc
The Web collaboration review (Formats: html | text)
This file contains some information about various kinds of web collaboration
software, usually where the collaboration takes place through HTML forms, rather
than a software component written in some other languae.
The MoinMoin Wiki(http://twistedmatrix.com/users/jh.twistd/moin/moin.cgi/MoinMoin)
This is the wiki used by thehttp://www.docbook.org/wiki/ site. It is written
in python and looks reasonably good. There seems to be a slight problem using
this is the lynx browser because lynx may not be able to insert newline
character in textareas (although it does seem odd), which makes it very hard
to write text which will be formatted nicely.
The tikiwiki
Another system
http://www.twiki.org
This is a very widely used wiki system, (eghttp://www.java.net) which is
written in perl and runs as a cgi script. It has good 'differences',
templates, access control and is generally fairly shmicko
However it does suffer from the usual problem that the actual wiki plain
text looks like a dogs breakfast because of all the strange little
markup codes in it (although its still better than, say, html)
http;//tavi.sourceforge.net
Another system probably not as mature as above.
The philosophy behind a 'plain text markup language' is to allow a writer to
write documents in a very basic text editor (such as for example, Microsoft Notepad,
or an HTML textarea form component) and then convert the text into other display and
storage, and exchange formats using some kind of stream editor or a program. The
transformation program recognises a set of simple 'structures' within the plain text,
such as, for example, URLs beginning with a URL schema like http:// or nntp:// and
terminated by a space character.
Another feature of these 'minimal markup languages' is that the writer should be able
to learn the basics of the format in a very short period of time, say, a few minutes.
The term plain text markup language is possibly incorrect since mark-up seems to
imply the existance of 'tags' of some sort, which are not necessarily present in
a minimal text format.
My idea is that a plain text language should be a compromise between the need to
render or transform the text into a different format (such as for example,
docbook XML or HTML) and the need for the original text to be readable in its
untransformed state. This is where, I believe, many of the 'wiki' languages fail.
For example, the wiki idea of using mixed case words such as ThisIsAWikiLink to
create links between pages, in my opinion, violates the idea that the original
text should be readable in itself.
Another idea is that the plain text structure should encourage the editor to
think about the document in a structured way, and to edit the document
accordingly. For example, the writer should make an effort to decide in what
section in a particular file or web-page her changes should be made or added.
It is very common, when dealing with wiki systems, that the information becomes
extremely disorganised.
I do not agree with the idea (used in the 'apt' system) of using indenting to
control section headings.
The apt structure uses <text> and <<text>> to render text as italic and bold
The moin moin language uses backticks (`) to render text in a fixed pitch font
or as a 'code listing' or something similar. I like this because it is readable
and it also harks back to the unix idea of back-ticks being a 'system exec' call,
which seems appropriate since this format is often used for
mjbishop plain text markup-language (/utils/)
The web-site athttp://www.ella-associates.org/alexis-info/docs/ has been created
using scripts which recognise basic plain text structures.
Features:
attempts to preserve white-space in the HTML, by using nbsp elements and
<br> tags. Tries to guess what is a hyperlink. Uses all upper-case text lines as
section headings. Contains structures for links to various versions of the same document
in the situation, say, where there is a text, html, and pdf version of the same
document.
The 'apt' (almost plain text markup language). This is to be found at
http://www.xmlmind.com This minimal text language uses indenting to achieve formatting
Features:
Uses indenting to determine 'section headings'
The markup structures used by the seq faq conversion script. Somebody called
'Eric Pement' maintains an FAQ about the 'sed' stream editor, and old and venerable
Unix program. He generates the HTML of the faq using a Perl script which recognises
certain structures in the plain text of the FAQ. The output uses CSS and is reasonably
pleasant to look at, in my opinion.
The moin moin wiki language.http://twistedmatrix.com/users/jh.twistd/moin/moin.cgi/MoinMoin|
This language is a variant of the normal wiki language
In my opinion this system tends to suffer from 'link proliferation',
since it is too simple to create a new link, which eventually interfers with the
readability of the document
Here is the markup summary
Emphasis: ''italics''; '''bold'''; '''''bold italics'''''; ''mixed
'''bold''' and italics''; ---- horizontal rule.
Headings: = Title 1 =; == Title 2 ==; === Title 3 ===; ==== Title 4 ====;
===== Title 5 =====.
Lists: space and one of * bullets; 1., a., A., i., I. numbered items; 1.#n
start numbering at n; space alone indents.
Links: JoinCapitalizedWords; ["brackets and double quotes"]; url; [url];
[url label].
Tables: || cell text |||| cell text spanning two columns ||; no trailing
white space allowed after tables or titles
With comments by me:
Emphasis:
''italics''; '''bold'''; '''''bold italics'''''; ''mixed
'''bold''' and italics'';
Looks fairly horrible in the text file. I dont really think it is
necessary to give the user control over italics and bold and such
purely formatting things. If you give the user this sort of control
then they start to think in a 'visual' or 'presentation' oriented way
rather than thinking in an 'informational way. I believe it is
important that the user remains focussed on the informational content
of what they are writing.
---- horizontal rule.
Doesn't look that great in the text file, nor in the resulting HTML.
Is this really a good way to divide information?
Headings:
Title 1 =;Title 2 ==; === Title 3 ===; ==== Title 4 ====; ===== Title 5 =====.
This could be altered to just the leading equals sign. I think the
leading and trailing = signs are ugly and unnecessary.
Lists:
space and one of * bullets;
1., a., A., i., I. numbered items;
1.#n start numbering at n; space alone indents.
Links:
JoinCapitalizedWords;
I think this is very unreadable in the text file as well as in the
HTML. Big mistake. How about something like [*] linktext
["brackets and double quotes"];
url;
[url];
[url label].
Tables:
|| cell text
|||| cell text spanning two columns ||;
no trailing white space allowed after tables or titles
This is not a bad idea, but I think all those vertical bars in the
text file looks very bad in the text file. How about just start the
line with || but then use fixed width text columns to indicate the
table. This would at least look better but might present some
parsing problems.
Unformatted Blocks:
{and} to surround unformatted sections of text
OK but ugly. I prefer my -->> and --<<
` (backticks) for inline unformatted sections
This I like. It is simple and readable, but it will cause problems
when displaying unix scriplets which contain these characters.