Re: Starting to modularize legacy

Ian - I will run for cover too - this is one of those discussions that can go on forever.

Here is my logic - re-factors make maintenance branches hard because unless you pull the trunk over to the maintenance branch - the modifications just don't merge. But when you pull too much from trunk you get stuff that is not bug fixes.

So my "October thought" is to get 2.2 out, in production at many sites, get 2.2.1 out and possibly 2.2.2 out and then land a bunch of refactors en-route to 2.3.

The re-factors are not usually problem for stuff working - they just make maintenance branches problematic.

In terms of Cambridge - my hope is that 2.2 is what you will be running with some bug fixes in 2.2.1 and 2.2.2 and then you can "stand pat" as we move from 2.2 to 2.3.


On Feb 23, 2006, at 8:13 PM, Ian Boston wrote:

I agree with nearly all of this, but I want to clarify the
"October is a good time" bit because Im not certain I understand and October is a very sensitive time for Cambridge.

We will are planning to have launched Sakai to the whole University and are also planning to have turned off our exising campus wide community services having migrated everyone to Sakai.

We are not a big institution by US standards and we can probably keep up, but more than likely would aim to take an older version rather than risk egg on our faces.

So its that, wait until October to do a big change and bring the rest of the code base along between October and December or aim to have the big change completed by October?

In the UK 'holiday break' often means July - August ?

I agree with everything else although I would like to add that I quite like the current set of Entity concepts, having just made wiki an Entity, which has saved me a lot of time. The only bit I dont like about that one, is its package name.


ning for cover :) )

This automatic notification message was sent by Sakai Collab
( from the WG: Framework site.
You can modify how you receive notifications at My Workspace > Preferences.