[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ecf-dev] ECF Release Train Participation
- From: Scott Lewis <slewis@xxxxxxxxxxxxx>
- Date: Wed, 15 Dec 2010 19:52:27 -0800
- Delivered-to: email@example.com
- User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:126.96.36.199) Gecko/20101207 Lightning/1.0b2 Thunderbird/3.1.7
On 12/15/2010 6:06 PM, Jeff McAffer wrote:
I completely understand the resource problem but have an honest question. What is the real incremental cost of being on the train?
Unfortunately, in my view it's quite a lot. The must dos for Indigo
continue to increase...and most of the existing (as opposed to new) ones
also have a significant yearly incremental cost. There is also the plan
creation, the review materials creation, the reviews themselves, and the
inevitable time cost of the IP process (witness the yearly last minute
fire drills around license consistency, etc), and the incremental releng
costs (which never seem to actually go to 0...although all of us would
like it if they did).
You don't have to do a new or extra release for the train. You could contribute the current released version of ECF to Indigo with no additional work other than the must dos.
As for the train requirements, yeah, that is a burden. However, I'd argue that they are in place, for the most part, to help our consumers.
In my view they a) place too much of a burden on the project committers
(i.e. the committers are the ones that do the actual work of meeting the
requirements); b) are overly focused on helping certain consumers
They are likely good things for your consumers too.
Perhaps for some consumers...but at the necessary cost of other
consumers. For example, every bit of time that Markus, Wim, or I spend
upon releng or IP for the simultaneous release (for example), is time
*not* addressing other consumer's desires for new discovery features
There are sure to be some that are less relevant to RT-ish projects. We should seek to change this or get exemptions where it makes sense. From what I can see though the real work of being on the train is doing a release. So is the question "should ECF do releases" or "should ECF be on the train"?
No, I don't think this is right. After doing at least 7 releases of ECF
and 5 simultaneous releases, I can confidently say there *is* a
significant incremental cost of being on the train...over doing a
regular ECF release (which is always a fair amount of work). And thanks
to the additional must dos, this incremental cost consistently
increases. It has become akin to a set unfunded mandates IMHO.
There's little doubt in my mind that we/ECF will continue to do regular
releases (e.g. we just had a 3.4 release in November and have 3.5
release tentatively planned for Feb/Mar 2011). In fact, given our new
build infrastructure and move to git, our release frequency might
actually increase. This is also of benefit to many consumers.
In my opinion, however, the value of the simultaneous release to communities like ECF's has gone down...and the costs of participation have gone up over the past few years (i.e. my previous note). The simultaneous release itself is no longer very 'new' to lots of folks (users/consumers of Eclipse RT technology and/or Eclipse)...and so doesn't really result in much improved exposure/distribution for participating projects (like ECF).
Not being in the Indigo repo would significantly dampen exposure and consumption in my opinion. The Helios repo is a very convenient and trusted location to get Eclipse components.
Perhaps...but no resources is no resources. How is that to be
addressed? If additional resources can be found and/or contributed then
the choice to continue with Indigo is obvious/easy. If additional
resources can't be found, then I don't feel I can commit to something
that I knowingly cannot do.