User:Ilya/OpenWetWare/Todo

Open

• username conflict problem (Austin's solution seems to be the most comprehensive but also the most difficult to implement)
• stop sending bounces to return address provided by spammers (accept only outside mail for local delivery and only local mail for outside delivery?)
• filter out http 200 and 404 entries logwatch reports
• install openldap
• follow up openwetware hostname alias transfer to model (https://help.mit.edu/SelfService/)
• Add vhost directives to Apache logs to distinguish traffic among the sites
• Eventum
• RSS feeds in wiki extension
  • add feed extensions to User:Ilya/OpenWetWare/Feeds
  • feature requested: formatting of feeds (Magpie?)
• Embed PHPBB
• Semantic mediawiki - installed on test (see Austin's email for details)
• Move pheromone wiki to rackspace server:
  • user accounts - openid
  • delete spammers' user accounts
  • <Googlesearchheader>
• try Drupal, see CMS
• try Pligg - an open source social networking content management system (CMS). It combines social bookmarking, bloggin, syndication and a democratic editorial system that enables users to collaboratively submit and promote articles
• stop rotation of OWW access logs (make a script to archive the logs before they are rotated?)
• enable rsync over SSH for off-site backups
• graph of connections among pages on the wiki
• citation box - a form to fill out (publication name, first author, date, etc) to help users cite materials outside Pubmed; examples on Wikipedia?
• IRC backend (like #bioinformatics channel on IRC) or Jabber service (via GTalk or own server).
• Create SPF record

Chat

• one common room with many private ones vs room per page? (need IRC/jabber server plus a front end that integrates with wiki - makes logs accessible from within the wiki for easy searching)
• the front page could display the users currently in chat, which would make going there more enticing (TK)
• good help page for jabber service
• JWChat is a full featured, web-based Jabber™ client. Written using AJAX technology it relies on JavaScript and HTML at the client-side only.
  • Demo page
  • SourceForge page

DOI

• The DOI system was established to provide persistent identifiers for digital content because URLs were "never meant to be an identifier but only to designate the location of objects" (http://www.press.umich.edu/jep/04-02/davidson.html). In other words, DOIs were designed to be generic enough so that they could be used on top of any system, including but not limited to DNS/Internet. The DOI system provides a level of indirection between the identifier and the object it identifies. This is very useful for keeping URLs current - when location of an object changes, you just need to update the URL associated with the object's DOI.

• Besides resolution services, the DOI proxies/resolvers (which are operated by the Registration Agencies and are currently web-based) also maintain some structured metadata (like author, date, title, etc) for every object. This is useful for more efficient searching, for example.

• To obtain the real DOIs for OWW content, we'd need to register each DOI with a DOI Registration Agency. This apparently requires an yearly fee (just like the Internet domain registration fee) and possibly, a one-time fee for the DOI prefix (http://www.medra.org/en/terms.htm, http://www.crossref.org/02publishers/20pub_fees.html).

• To be confirmed. DNS is hierarchical: one needs only to register a domain name and then can have unlimited number of subdomains. DOI systems is flat, so each DOI (suffix) needs to be registered with the system, even if one owns a prefix.

Blogs

• Problem: new blog notifications are sent to you@example.com. For example:

New blog created by admin\n\nAddress: http://blog.openwetware.org/outsidethebox/\nName: Outside The Box

• New comment notification for post authors and others

Aggregation

• From Austin: an alternative to merging rss feeds on the wiki is to use Planet. Many sites out there use it, for example:
  • http://planet.gnome.org/
  • http://planet.mozilla.org/

Promotion

• Iimplement a viral system that would allow existing users to give accounts to others but with a limit - like GMail used to have. For example, if I'm a user, I have five pre-approved accounts to give out and so would each of those five, etc.