Managing Google Groups headaches

Chris Angelico rosuav at gmail.com
Fri Dec 6 09:51:13 EST 2013


On Sat, Dec 7, 2013 at 1:11 AM, rusi <rustompmody at gmail.com> wrote:
> Aha! There you are! Its 'page editor' here and not the html which
> 'display source' (control-u) which a browser would show. And wikimedia
> is the software that mediates.
>
> The usual direction (seen by users of wikipedia) is that wikimedia
> takes this text, along with the other unrelated (metadata?) seen
> around -- sidebar, tabs etc, css settings and munges it all into html
>
> The other direction (seen by editors of wikipedia) is that you edit a
> page and that page and history etc will show the changes,
> reflecting the fact that the SQL content has changed.

MediaWiki is fundamentally very similar to a structure that I'm trying
to deploy for a community web site that I host, approximately thus:

* A git repository stores a bunch of RST files
* A script auto-generates index files based on the presence of certain
file names, and renders via rst2html
* The HTML pages are served as static content

MediaWiki is like this:

* Each page has a history, represented by a series of state snapshots
of wikitext
* On display, the wikitext is converted to HTML and served.

The main difference is that MediaWiki is optimized for rapid and
constant editing, where what I'm pushing for is optimized for less
common edits that might span multiple files. (MW has no facility for
atomically changing multiple pages, and atomically reverting those
changes, and so on. Each page stands alone.) They're still broadly
doing the same thing: storing marked-up text and rendering HTML. The
fact that one uses an SQL database and the other uses a git repository
is actually quite insignificant - it's as significant as the choice of
whether to store your data on a hard disk or an SSD. The system is no
different.

>> MediaWiki uses an SQL database to store that lump of text, but
>> ultimately the relationship is between wikitext and HTML, no SQL
>> involvement.
>
> Dunno what you mean. Every time someone browses wikipedia, things are
> getting pulled out of the SQL and munged into the html (s)he sees.

Yes, but that's just mechanics. The fact that the PHP scripts to
operate Wikipedia are being pulled off a file system doesn't mean that
MediaWiki is an ext3-to-HTML renderer. It's a wikitext-to-HTML
renderer.

Anyway. As I said, your point is still mostly there, as long as you
use wikitext rather than SQL.

ChrisA



More information about the Python-list mailing list