Developers Meeting
11 April 2011
Meeting Notes
Definitions Moderator: ttwetmore resigned
“My vision for Better GEDCOM is at odds with some others.
The issue is whether BG should be a modification to GEDCOM that allows the current generation of software products to better share their data,
or whether BG should be a model for the next generation of products, supporting both today's features, but also going to the next step of supporting record and evidence based applications, and applications that will interface with large data providers.
BG was apparently formed to solve the former problem.
I and some others have been pushing to go further and solve the latter.
I don't believe that solving the first problem will have any impact on the genealogical industry, and that, if BG is to have a lasting impact, it must address the second.”
Andy - Tends to agree with Tom. Fix what’s wrong with current GEDCOM, but expansion for future evidence-conclusion won’t appeal to developers. The developers need an out on the other end. Wants to see both problems handled today. Take the whole package.
Tamura - Tom is participating in OpenGen, and they are receptive to the evidence-conclusion model.
Sandy – Part of the issue for getting BG adopted. If it was easy to go forward with more evidence-based model, it would have been done already. Repository existing docuements. Ability of vendors to move forward and to see a need to depart from their current defensive mind-set. And the only way to (possibly) get them to change – would be to support an evidence-conclusion model. If you look at the software out there – different source templates. None currently support evidence-based model. Vendors from major tree software must participate.
Roger – Ultimately an evidence-conclusion is far away – but ultimately the [goal]. Vendors add evidence basis tags and change in user interface as well as line data structure. .bettergedcom is a good start and evidence conclusion is the BetterGEDCOM future.
Michael – is a developer, working on a program. What he wants in BetterGEDCOM is something where everyone agrees. We may get more developers if there is something at BetterGEDCOM. Data model. There is a lot of talk – theoretical stuff. But without a model there is no possibility. [He responded to Tom’s data model specifications.] If we want to interest developers. FamilyPursuit.com - he sees his product as conclusion-centric.
Roger – People say “I bought Mac, importing GEDCOM into Reunions and now some of my data is lost.”
Geir – talking about creating a model. A complete model of BG standard will cover more than evidence-conclusion. Should we give priority to evidence-conclusion? Not saying we should. How do we know that this evidence-conclusion stuff is what vendors want to discuss?
GeneJ – Smart science – (per Adrian’s post to the wiki message/view/Research+Process%2C+Evidence+%26+GPS/37470240 ). Feels if we go too far, we go back to a GEDCOM 1.0.
Geir – One high level consideration – if we do not get the vendors on board, why are we here at all? Either we get them on board, or reassign our short term goal. Map the requirements. One possibility. Count how many people who are currently working on this – five. We aren’t going to create the standard. If only five? What will be useful in the future?
Geir – the other concern over the last few weeks. Being pulled in various directions by three people to work on other parts of the wiki.
Tamura – reality check:
Tom seems both right and wrong to me.
Having a good model is important, but current problems are not unimportant.
I think he would agree with that, but just does not agree with the
priorities other gives to some issues.
Fact is, different people have different perspectives.
Different perspectives is good, it brings a rich variety of ideas to the table.
It is important for BetterGEDCOM to have some *focus*.
BetterGEDCOM activity all over the place; every addresses what they
see as a problem
You are addressing 15 years of neglect, so there is no lack of issues to address
BetterGEDCOM aims to solve the issues
You cannot solve everything at once
You will have to make multiple versions
You need to make choices, and not everyone is going to agree with the choices
To really move forward, you need to make choices what to do first.
Michael – If we don’t get the vendors on board. The most useful part of the wiki is the Requirements Page. He can solve the problems now that he knows what they are. Then the data model.
Roger – Ran Tamura’s torture tests using Reunions. Level 2 or Level 3 tag… recommend to vendors changes that can work. Moving toward Louis Kessler and Tom Whitmore’s suggestions.
Geir – Re: Mike’s comments on Requirement’s Page, if it is the priority. It may at some stage cover all the issues, so that will not be a way to prioritize. Fixing the existing problems – they should be done in the light of the future.
Sandy – No reason why approach cannot be split into multiple levels. Form smaller group – immediate issue today (some are deliberate.)
1 – Immediate issues – define fixes to communicate.
2 – Evidence Conclusion Model
GeneJ – Tamura mentioned 15 years of neglect. Roger – testing on existing software. We could test until the cows come home. Enough testing to ID main issues, well documented. Another group works on solutions.
Tamura – Some vendors will immediately fix, some ignore. Get some other blogger to do a positive review. Reviewers who give only positive reviews are also part of the problem. More testing (real reviews) is good. Real tests – real opinions.
Michael – Reading the wiki page “What is BetterGEDCOM about?” Page described problems only, he felt he would find a data model. [Not interested in remaking old – looks to future.]
Geir – Possible priorities. Also consider [calendaring] priorities so every few weeks we hit on a topic of interest to some wiki participants.
· Requirements
· Data model
· Record structure
(Comment on the minutes from Geir : I don't remember why I mentioned these 3, but did not mean that we should work eg. 2 weeks on Reqs, then 2 weeks om the model, and so 2 on records.)
First step on evidence-conclusions – call the new page BetterGEDCOM Evidence-Conclusion model (not DeadEnds in the name.) Create common starting point for that discussion. (Comment from Geir: I said we need to create a table of content to previos discussin on E&C as a starting point.)
GeneJ – We need concrete examples when discussing.
Geir – If we’re going to go forward with this. A few days ago Tom posted that FamilySearch is doing something according to GenTech. I feel we have to go forward with Evidence-Conclusion with an open mind.
NOT DISCUSSED:
Comments
Too much running around in different directions. We all have different ideas. And nothing will get done if there's no real concrete measureable goal. I'm not talking about the ones that says "BG should be ...". I'm talking about things like "Implement a Level 0 Evidence Structure".
With just a few people, I think BG should work on just one problem at a time, and come up with a concrete extension to GEDCOM that is a solution to that problem. That will be progress and will stimulate the next improvement.
But trying to fix everything at once is a recipe for spinning wheels.
p.s. I'm not avoiding. I'm just very busy.
Louis