Date for TW 2.1 release?

[Udo Borkowski]

Is there any idea when TiddlyWiki 2.1 is released?

I am asking because I am working on a plugin that could take advantage
of the MetaData feature scheduled for 2.1. Alternatively I could use the
DataTiddler for that purpose but I would prefer to use the core MetaData
feature. But if the release of 2.1 is say in more than one month I
probably will release the plugin with the DataTiddler.

Looking at all tickets assigned to 2.1 in Trac it seems to take quite
some time to complete them all. Could we postpone some of them for a
later release?

A more general question:

What about the following release process approach: with every milestone
(like 2.1) also a planned release date is given. The goal is to release
the product at the given date, with all "major" features/tickets
implemented. When making a feature "major" one needs to estimate if it
can be implemented in the remaining time to the milestone. If not it
cannot be made "major" (or some other "major" feature is down-graded).

If there is some development time remaining until the release date lower
priority tickets may be implemented. Typically the plan should be that
there is enough "remaining time" a) to have some buffer for unexpected
delays in the major tasks and b) to ensure that also "nice to have" but
not that important features are added to some release. E.g. Ticket #5
("Remove phantom <style> tag that's confusing UltraEdit") just requires
one extra space char in the code to be closed, but is rated as
"trivial". If only priority would count this would be the last task to
do, so it may be pushed to a later release if the schedule is too tight.

The idea behind all this is that the milestones are not "defined by
features" but by a timestamp (and a prioritized set of features). From
what I saw in other projects those projects that are using predefined
release dates have a "better drive" than those solely refering to a
feature list. Also having a concrete release date makes it easier to
plan (e.g. for third party/plugin developers).

(Of cause, details like how testing, beta phases, working with multiple
contributers, "vision" etc fit into this approach need to be discussed.)

Udo

[Martin Budden]

On the more general question, I agree with Udo's view, namely that the
release should be time based, with features cut to make the release
date. As Udo says, this gives the project more drive (for example, it
would give me motivation to finish something before the release date).
It also makes it easier to plan. In my experience the "train leaves the
station" release model results in more successful projects.

And if a feature misses a train, there will be another one along
soon...

Martin