[eml-dev] EML versions

[Matthew Jones]

Hi Margaret,

Its a great thing to get out a new release. I fully support your push. A 
decision on a release should go out to all of the EML developers for 
approval, so I suggest that you prepare a release candidate, tag it in 
CVS, and ask for feedback.  Then we can put the release to an approval 
vote as is the procedure for the EML project -- if its a well-conceived 
release there should be universal support.

There are a bunch changes already checked into the head of CVS for EML 
that are (almost?) ready to go. Like the access control changes that are 
described in one of the bugs.  Testing needs to be done on many of these 
things.  Make sure you run the JUnit tests, and that everything passes, 
including the build of the website and documentation and specification, 
etc. Testing takes a while, and I really think you need one of me, 
Chris, or Chad to help because we know all of the release gotchas (or at 
least we used to).

There are several 'unit' bugs in addition to 3051.  Several of these 
already have fixes in CVS.  They all need to be reviewed I think.

I'm very busy until after the Ocean Sciences meeting, so I wont have a 
chance to help until mid to late March, but catch me on IRC in #eml if 
you want to discuss particular things.

Matt

Margaret O'Brien wrote:
> Hi Matt -
> Last week, Inigo San Gil (LNO) and I talked about EML versioning and 
> release, and he is eager to help out. In fact, we set a target date of 
> March 31, and I'll coordinate. We came up with these bugs that we though 
> should be addressed early -- probably in the first 2.1 release:
> 
> 2054, 2703
> both of these create validation issues, and make it difficult for new 
> EML-adopters to use some parsers. As you said, 2054 (additional 
> metadata) is already done. Inigo has been investigating 2703, but it 
> might need some discussion. We added a note to it recently.
> 
> 2076
> my own personal favorites concern the literature module. There really 
> should have been multiple entries here- blame my inexperience. To the 
> best of my knowledge, only 2 (cardinalities of volume and pageRange) are 
> essential to making EML usable for publications. The others need 
> discussion.
> 
> 3051
> 2 units were left off the standardUnits restriction list - or should 
> this be added to 2.0.2?
> 
> I'm free to talk this week as needed -
> Margaret
> 
> 
> Matthew Jones wrote:
>> Hi Margaret,
>>
>> If we were to get the BDI grant, I think coordinating another EML 
>> release and integrating it with our OBOE approaches would be a good 
>> thing.  If not, we'll need to continue to limp along on volunteer 
>> efforts. Hopefully we'll find out soon -- BDI usually notifies in 
>> early January.
>>
>> Most of the work for 2.0.2 has been done now. Only 2 bugs remain in 
>> bugzilla for 2.0.2 (2083 and 1662).  Some bugs targeted at 2.1 are 
>> done too, such as:
>>   http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2054
>> This bug deals with the validation issues, and I have already checked 
>> in fixes to CVS as detailed in the bug.  This change, however, could 
>> cause validation issues for some people, so it should not be released 
>> as a minro point release, and instead should be released as a partial 
>> API change as 2.1.0.  This doesn't mean that we need to do more work, 
>> or even finish all of the 2.1 features currently listed.  Rather, it 
>> means we should label the release properly for the scope fo changes 
>> we've made.
>>
>> So, if you'd be willing to coordinate the next release, probably 
>> labeled 2.1, and simply make sure that all of the bugs are properly 
>> assigned for the right target milestone, and test to be sure they are 
>> indeed fixed, then I think we could release a new version in short 
>> order.  Maybe you and I could discuss this sometime this week?
>>
>> Matt
>>
>> Margaret O'Brien wrote:
>>> Hi Matt
>>> We haven't talked awhile about EML versions. Do you have some time to 
>>> talk this week? I've been talking to Inigo San Gil (LNO) lately, and 
>>> he is very concerned that EML loses credibility the longer we work 
>>> with a version that has validation issues.
>>>
>>> As we left it (last summer during a dev-lunch), the feeling was that 
>>> EML2.0.2 might not be needed at all, and we should consider what to 
>>> include in 2.1 instead. You had asked me to consider shepherding this 
>>> project. I'm willing to, but SBC's PIs are not willing to give up 
>>> that much of my time, and LTER doesn't have supplemental funds 
>>> available for IM. So for me, it's been on a back burner. If it were a 
>>> smaller effort (ie, just a few bugs targeted for 2.0.2), I might be 
>>> able to swing it. Or maybe 2.0.2 is essentially done and sitting in 
>>> someone's local copy.
>>>
>>> BTW, any word on the BDI proposal? since that has major EML work in 
>>> it, it seemed reasonable to work on the next version along with that 
>>> project.
>>>
>>> I've attached the summary of bugs and comments/notes that we used 
>>> during dev-lunch, and marked 3 of them as 2.0.2 targets.
>>>
>>> regards,
>>> Margaret
>>>
>>
> 

-- 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Matthew B. Jones
Director of Informatics Research and Development
National Center for Ecological Analysis and Synthesis (NCEAS)
UC Santa Barbara
jones at nceas.ucsb.edu                       Ph: 1-907-523-1960
http://www.nceas.ucsb.edu/ecoinfo
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~