[dokuwiki] Re: The database question: a new spin

  • From: "Joe Lapp" <joe.lapp@xxxxxxxxx>
  • To: dokuwiki@xxxxxxxxxxxxx
  • Date: Wed, 14 Sep 2005 18:00:42 -0500 (CDT)

Hi Seun:  I'm not sure we're on the same page.  I'll just reply briefly...

> > The main benefit is that you improve server availability
> I'm not so sure that this is likely for common usage scenarios.

Right, server availability is not an issue in common usage scenarios.  By 
definition, availability defines the peak load that the server can handle, and 
if peaking is common you need a new server.

> > you will probably also get some, mostly negligible, increase in page load 
> > times.
> I disagree.  Nothing is faster than reading one file to write a HTML page.

But DokuWiki is often accessing multiple files in the process of returning one 
HTML page.  Seeing this in the code is what prompted my suggestion.  But making 
it 'fast' isn't my primary objective.

> > However, the improvement to RSS feed generation would be non-negligible
> I agree.  but the RSS feed can be easily cached in a file.

I do think we need to cache the feed, and for most sites this caching will 
probably do the trick, but some very few sites may be edited with such 
frequency that caching doesn't help.   Besides the cached RSS page periodically 
expires, and each time somebody will be waiting.

> > Link color determination is another area that would benefit.
> I have always felt that this is not such a necessary feature
> > And full-text searches could also benefit, depending on how they're 
> > implemented.
> Practically speaking, Google Sitesearch is a suitable alternative solution.

I'm using Google sitesearch on my site right now and am not happy with it.  
Google hasn't quite noticed my site yet, and when it does, I expect the search 
results to be a bit stale.  Besides, there's nothing better than owning the 
search result page.

> I think I like the way dokuwiki doesn't require a database.  I really
> don't think there's any reason why performance should suffer.

But I'm suggesting that the database be optional.  Performance isn't the main 
problem I mean to address.  I'm just being ambitious and envisioning DokuWiki 
sites that can take a serious beating, such as during those few brief moments 
of fame each of us hopes to one day get.

I realize that we passed a lot of emails and it's hard to catch up.  Hopefully 
this helps.  Thanks for putting some energy into keeping DokuWiki on track.

DokuWiki mailing list - more info at

Other related posts: