This might sound a bit pessimistic...

Small analysis of the current situation: Current Maemo5 calendar lacks features. It has been made using GTK. Maemo6 is going to have QT apps. Therefore, it is unlikely that current maemo5 calendar will be enhanced. So if Nokia instead of stopgag wants to make good calendar for Maemo6, they have UI specialists and people who are able to use competing products to see what features a good calendar should have (I guess this we will know earliest next September - Nokia is rarely on schedule). If they won't still do it, somebody will do it commercially. On the side I read that KDE PIM port has been talked of too.

So either one waits, takes some existing calendar or programs a new. Good calendar needs for first a storage which is flexible and can store all kind of weird items. The current storage might be good, maybe it'll be used also with Maemo6. Second you need a user interface design with good usability. Finally "just" the functionality to make it all work is required. After all of this you need exporting, importing, desktop applet, alarms etc.

I would guess that those people who have made their own calendar have a lot of free time, strong itch to scratch and possibly no idea what they will be involved at. Basic calendar without repeating events is easy, but when there is repeating events the can of worms is open. For example, think how you would store biweekly repeating event with no ending date which has for half an year different location and hour? So that it is fast? So that the whole thing can be deleted at once?

There is currently available at least bundled calendar, GPE Calendar, mCalendar, Kimi. There used to be some more such as osmo and dates. Most of the calendars fail both with usability and repeating events. Why it is so?