log in


2008-02-29 16:40 UTC

Suggestions for TIME

And maybe other stuff

TIME home pages?

A "home page" for the user of what pages they've written, how many times they've updated it, etc. Good way for the user to keep track of their articles and find out who wrote their favorite.
TIME Regex for Bold and Italics.

Proposal: Bold * single star only tripped when paired, on each line. Block bold comments *** to the next ***, which must be on a different line.

For example:
 This is sample text that should not be matched.
And this is some* <h1>html</h1> that should NOT be touched.
This should not be *touched.
This *should* be touched.
This *should* be touched, as should *this*.
Will it catch this *?*
*Whole line, maybe*
***Shouldn't match this.
*Foo*, not *bar.
***This is a multi-line bold comment,
for when you are really angry about something
and definitely want it highlit.***

The Perl regex: s#(?<!*)*([^<>*\n]*?)*(?!*)#<b>$1<b>#g does this for us. The multi-line version is here: s#***([^<>*]*?)***#<b>$1<b>#g;, and matches three asterisks in a row, across lines. Let me know if there are problems and stuff, or if this is not useful.

Known errors: multi-line: If you remove the line *Foo*, not *bar. from the text above, it matches the start of the first one, but not the end of the last one. Hmmm.
_italics_ and *bolding*: italics and bolding Doesn't render properly.
Escape sequences.
TIME short term goal?

As a short term goal, maybe try and make TIME contain all those tech articles wikipedia DOESN'T have, or those that are useless on wikipedia. For example, documentation to a reference in C#, or how to use a given C function set. This might help stir interest into other areas, such as "how to learn C#", and regardless, the increased documentation would help immensely with learning as it is. Sometimes it's hardest finding information in a concise format about the topic you're looking for, like String.Format in C#.

Anyways, a 'coder wiki' would probably make life easier in the short term, and help lead to long term benefits.


Each article should carry information as to who authored it, as well as when, and when the last update was. This would give more statistics on who is contributing, limit the number of bad articles due to people not having their name on it - one is far less likely to do something badly when one's name is attached, and how often they update the information therein.
TIME internal linking

While the linking is pretty sweet as is, [[295:foo]], where the link shows up as "foo", it's not very intuitive. We can't all remember the index number of the article we mean to look to without jumping pages.

Adaptive Content <-- apparently this works.

Possible solutions:
  • Allow linking by title name; if spaces are an invalid character, use dashes or eliminate spaces e.g.
    - "teapot-on-learning-programming-language-basics", or
    - "teapotonlearningprogramminglanguagebasics".
    • Both of those are a lot to type though, so it would be somewhat less useful than if one had a search bar built in to the edit page to insert links (not the one at the top - one on the side of the edit box).
    • However, if linking were done by title name, creating articles "on-the-fly" could be implemented. This has some associated problems though:
      • Empty articles. How many will be created accidentally, or on purpose, and never filled? This might be more useful when TIME is larger, such as Wikipedia, and have links that are empty display in red so users know to fill them in if they are knowledgeable on the subject.
        • Users could have empty articles they created emailed to them so as to provide a reminder to either fill them in, or find someone to do so.
        • Empty articles could be a category searched by all users, in case they have the know-how to fill one in.
      • Typos; these can create empty articles accidentally, with no real easy way of deleting them: titles can change though, so typos can be fixed IF changing the title doesn't make an article non-empty.

Time Searches
  • Simple algorithm to determine relevance instead of last updated (which works while there is a small user-base).
    • Maybe a structured search, by tag/title/text precedence.