On 08/21/2013 09:20 AM, Jaroslav Reznik wrote: > Is it just related to our instance (because of bloat we have) or it's > just impossible to have working search at all? I can see some wikis > that are not as bad as our is but usually it's fail to be honest... AIUI, MediaWiki search is a bit touchy where it comes to some things, such as page names. This is why [[Natural page naming]] is preferred over [[Naming/pages/with/fake/nesting]]. In the end, our UX friends remind us that search should be a last resort if the information architecture fails. What challenges a wiki is the information architecture - getting pages in categories, putting categories as subs to other categories, having only a handful of master/top-level categories, and having pages that are obvious starting spots for certain types of content. Take team-draft-work for example - it's hard to get it out of the stream of search results, but good categorization and maybe a "team-related material that probably isn't what you are looking for" header blocks might help. That is, put out a template for team drafts so that it makes it clear from the top of the page that it's a team draft or team-related documentation that isn't necessarily there to help users. BTW, isn't there a wiki@ mailing list? I haven't seen anything there in a while, but it's a natural group to work with, and a good place to draw discussions in to, if we want. - Karsten -- Karsten 'quaid' Wade http://TheOpenSourceWay.org .^\ http://community.redhat.com @quaid (identi.ca/twitter/IRC) \v' gpg: AD0E0C41
Attachment:
signature.asc
Description: OpenPGP digital signature
-- websites mailing list websites@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/websites
