We have are numerous ideas for what we want to put on the site (and some thoughts on suitable locations too, for that matter). You might too, and together we can weave the Pike site of tomorrow into existence. Here are some of them; your suggestions will be incorporated here too, as time permits. Those of you who already have the rights to edit site content are of course invited to add your ideas here right away on your own. Common to many of our ideas and plans is the dependency on a login system for the site, referred to as (requires site auth system) below. The rationale for this is that, in order to implement the feature, we first need to have a means for giving you an identity for the site, and a way of proving that you are that person when you want to perform some actions (writing a presentation of yourself for others to see your involvement with Pike, changing the checkin message of something you committed, and so on). Adding a nice overview of what the site holds like the one Janne made for the previous generation Roxen.com website would be nice. While we so far have settled for using the old Roxen Community Crunch bugzilla descendant for keeping track of Pike bugs and feature suggestions, we would like having it here on this site, and preferrably a better system while at it. Included in the deal we signed with Roxen IS is also the successor to Crunch that Daniel Lublin designed as his Master's Thesis for Roxen some time ago. Unfortunately it was never deployed, but doing so is in the pipe. Stay tuned. Install issue tracker on pelix and configure it for Pike, the Pike site and possibly additional hosted projects (requires site auth system). Create conversion procedure to import old tickets. Remodel it to fit the Pike site layout. Mail/LysKOM notification to export new tickets (as well as changes to existing tickets) to LysKOM conferences and a new bugs mailinglist. Allow CC:ing anyone who wants updates on a ticket. The world should link the Pike site, and the Pike site should link the world. In short, we want all sites covering GL, MySQL, SDL, programming languages, system administration tools, bioengineering aids, and anything remotely Pike related or with a problem domain that fit Pike to link us. In return (and to help us keeping track of them all as well as to provide good links for us Pike people to what others do with Pike), we should link back to them too. The pike list won't stay at roxen.com forever, and there isn't much around as list archives go, at present. Of course, we have this neat place on the site reserved for a convenient web list archive, but it has not yet found its way there, somehow. :-) The /documentation/ section of the site needs both structural, markup and content upgrades. Among the details that would benefit from fixing: Reference documentation formatting could be improved. Full text search of the reference documentation. Article framework should get in place soon. New articles wanted! Old article content needs to be formatted and uploaded. Java and Pike 3D programming in Pike The page navigation needs redoing (vertical instead). Reinstate navigation arrows. Printable PDF, Postscript, RTF or DOC files wanted! Template upgrades are still needed; there is a lot of dead code rotting away, and the FrameMaker PDF generation system either should be scrapped entirely or brought back to the semi-living again. There are plans on scrapping the current wiki and replacing the quick proof-of-concept RXML hack in use today with a real implementation, either UseModWiki or if we can stand the wait, Leif's master's thesis project wiki. Create /download/archive/ Make a nicely formatted MPL license on the site. Update binary dist statement (to what?) Update CVS statement Port wishlist functionality (of some sort) from pike-community. Integrate with issue tracker? Port Hacks section from pike-community. Perhaps every hack should be given a free webpage (edited with an insite editor) and a cvs repository hosted with cvs browser et al. Only migrate hacks that work in Pike 7.3 (or update the code so that they do). The default template could use some generalization to allow parametrization of page layouting and whether to use the tent construction linebreakage scheme or not. Further into the future, dropping the XSLT usage altogether, substituting it for XQuery or something similar would be nice. Writing a space sensitive RXML type for the CVS browser would also be a nice improvement. While not strictly a (web)site issue, implementing anonymous cvs access on our solaris sparc server is an issue we have not yet addressed - running cvsd on non-intel non-linux systems is a hassle, Leif reports. Your help is welcome as always, if you too would like to get there sooner. From our cvs browser labs, we still have quite a few ideas percolating at the moment, and some (perhaps even all of them) will certainly happen sooner or later. Here are two lists of issues we would want to fix and ideas to implement for the user interface and indexer components respectively: Drop the line breakage - we too aren't very happy about how the line breaking scheme employed mostly everywhere on the site affects commit messages that do their own linebreaks for neat reading; this is high on the TODO. Annotate mode Linked titles like the diff popups in log, browse, diff and view revision views. Performance tracking and improvement of the view revision view would probably prove worth pursuing. Date boundaries are not relative to your timezone everywhere. Download revert script link in the log view to revert an entire change package. Dates in log view should link back to the checkins view. Plugin framework to allow for checkin reformatting filters for SiteBuilder and Wiki checkins. List all files belonging to a checkin in the log view just as in the recent checkins view. Report an error when a diff or show file failed (for instance when the file has been \rm:ed) Query functionality to search for specific checkins. File revision graph visualization/browsing (framed?). Log message editing as a web interface to cvs admin (requires site auth system). Add signal handler to print backtraces on receiving SIGUSR1 or similar Weed out possible MySQL locking bug that seems to occur at least at Lysator when indexing incrementally. Generate repository-wide version numbers when indexing incrementally too. Use Parser.RCS for local repositories, instead of cvs. Fix --sync-with-db mode for sweeping away \rm:ed files from the Code Librarian database; presently, there are a few minor bugs left to weed out there. Clean up Parser.RCS to be a pmod, discussion in kom thread 8615205.