Re: [jakarta.ee-spec.committee] Understanding the Current Specification

[David Blevins <dblevins@xxxxxxxxxxxxx>]

I think that would be a wonderful way to set everyone up for success.  Even if we all here became instant TCK Process experts, the world at large is not and the "start the conversation in the middle" danger is going to happen the moment we push out any rules.  Lacking shared context of the past, they are highly unlikely to like or understand any rules we make.

I would like to make an honest effort get permission from the people who filed the specific challenges as they were submitted non-publicly due to TCK requirements at the time.  If I can't, we'll yank their names.

If the group here likes the idea of sharing it more widely, we could potentially change the permissions or move a copy of the folder into the "Public" folder of our Specification Committee team drive and I could send a post to our wider Jakarta EE community list.  Sharing from our team drive would be good in gaining "openness" points for this group, which can only help us.  If that folder became filled with documents, it wouldn't be a bad thing.


-- 
David Blevins
http://twitter.com/dblevins
http://www.tomitribe.com
310-633-3852

> On Dec 6, 2018, at 8:55 AM, Ed Bratt <ed.bratt@xxxxxxxxxx> wrote:
> 
> David,
> 
> This is great! Thank you for putting this together.
> 
> It would be useful if we could share this more broadly. Given there are many "at large" community members who have never seen these results, or know what distinguishes compatibility testing from more common unit or functional tests, getting this kind of material out more broadly can be nothing but helpful.
> 
> Thank you again!
> 
> -- Ed
> 
> On 12/5/2018 3:35 PM, David Blevins wrote:
>> The current status of basically only Bill and I being deeply familiar with the TCK process I think is a major hinderance.  I think if we were to venture forward in the current state, it would result in a conversation primarily between Bill and I and you all feeling like you had to pick between your favorite parent.  I like Bill far too much for that.  You all need your own experience or as close to it as you can get.
>> 
>> In that vein I've attempted to put together a crash course of actual TCK challenges, exclude lists, and file names from Apache's and my own experience with the certification process.
>> 
>>  - https://docs.google.com/document/d/1CAnAYCtafDuf_s4DQhUZEIPO0GBOaT48X74N8LUQh-w/edit#
>> 
>> I ask you all to spend time reviewing the material in advance of next week's call.  We can spend some time discussing it in more detail when we're together.
>> 
>> Once we all have at least some virtual experience with real details and real challenges, we can begin to form our own opinions and make true progress.  I think if we start defining a process now, we'd be starting "in the middle" and would not be able to resolve with confidence or in any reasonable time.
>> 
>> This is still just a portion of the story.  CDI and Bean Validation have the same concepts but implemented entirely differently.  They also have their own document on how they prefer their process to work.  MicroProfile also has it's own process, less defined than CDI, but still third generation in many ways.
>> 
>> So our and your homework will be:
>> 
>>  1. Understand the Sun/Oracle TCK process and materials
>>  2. Understand the CDI/Bean Validation TCK process and materials
>>  3. Understand the MicroProfile TCK process and materials
>> 
>> 
>>