Decide what to do with decisions

[peterdesmet]

As noticed in #153, we need to find a place for decisions.html: https://github.com/tdwg/dwc/blob/64311240eeb15c0016c858846f9adbfb91bd515d/terms/decisions.html

• What needs to be kept?
• In what form
• Is it part of the standard? @baskaufs

[baskaufs]

It is not clear to me whether this should be part of the standard or not. I think that we decided that is should not be considered as normative for DwC, since it's really an administrative record. I don't know whether it should be archived somewhere in this repo and contain only decisions related to DwC, or whether there should be a more generic place where decisions are recorded for all vocabularies. In the latter case, it should be housed in some repo maintained by the Executive, since they make the descisions. But I suspect it would not be productive to hold up the cleanup of the DwC repo waiting on an answer from them. For now, I would stash it somewhere with non-normative content in the DwC repo.

As far as format is concerned, I'd save it as a CSV and then generate HTML and RDF from that CSV as we do with other terms. If you want, I can take on the task of making the conversion to CSV.

[peterdesmet]

Or... we still have it referenced here in this issue. But that only works if we're not planning to update this.

[baskaufs]

@peterdesmet @tdwg/tag-team

OK, with some investment of time, I've "fixed" this. To answer Peter's question, after some thought I don't believe that the decision history is part of the standard. It's a part of the historical record of TDWG, but not really a part of the DwC Standard itself.

Here's the "fix":

1. Went through the Decisions web page and created a CSV file that encapsulates the information presented in that page. While I did that, I also changed the decision URIs from complicated hash URIs to simpler ones of the form http://rs.tdwg.org/decisions/decision-2009-12-07_1 that use slash URIs as with all of our other URIs. The previous URIs didn't dereference, so I didn't "break" anything by doing that. I also changed the local name to lower camel case to conform to our normal pattern for non-class URIs.

2. Went through the dwchistory.csv file that @tucotuco sent me earlier and munged it into a two-column table that relates terms to the decisions that affected them. There were some missing terms because the dwchistory.csv file did not contain information about the Dublin Core terms. I added them along with several others that were missing in the dwchistory.csv file.

3. I edited all the configuration files necessary to turn the CSV files created in 1 and 2 into machine-readable form. I dumped them into the Vanderbilt triplestore along with all the other TDWG metadata that was already there. So one can now query via the SPARQL endpoint for terms affected by decisions using this query:

prefix tdwgutility: <http://rs.tdwg.org/dwc/terms/attributes/>
Select distinct ?decision ?term where {
  graph <http://rs.tdwg.org/> {?term tdwgutility:decision ?decision}
        }
order by ?decision

Note: In the process of working on this, I minted a new utility term: tdwgutility:DecisionClass (http://rs.tdwg.org/dwc/terms/attributes/DecisionClass) to serve as a type for TDWG decisions. At the DwC hackathon, the nascent Darwin Core Maintenance Group was faced with sorting out what parts of the giant mass of Darwin Core documents and data were normative, non-normative but within the standard, and outside of the standard. We concluded that the tdwgutility: namespace terms (formerly dwcattributes:) did not belong within DwC, since its terms could apply TDWG-wide and also because as a list of utility terms, it needed to be easily modifiable to meet the needs of TDWG without the need to go through the vocabulary maintenance process. In the spirit of that decision, I minted the new term - we needed a class for decisions, so I made one. I used the local name "DecisionClass" to avoid confusion with the already extant "decision" property. I also changed the label of ```tdwgutility:decision`` from "Decision" to "Has Decision" to indicate its linking function, and applied the label "A Decision" to the new class term.

So if we can get the TDWG metadata repo migrated over from my personal Github site to the TDWG site, it seems to me that the two CSV tables in 1 and 2 above should serve as sufficient documentation of the decision history, meaning that we could ditch the web page that is the subject of this issue.

[peterdesmet]

@baskaufs awesome! Seems like we'll be able to close this issue, once #170 is closed.

[peterdesmet]

@baskaufs can this be closed?

[baskaufs]

From the standpoint of Darwin Core, I think we can close this. The decision history is available in raw CSV form at https://github.com/tdwg/rs.tdwg.org/blob/master/decisions/decisions.csv and that should be good enough for anybody who cares. There are other issues like rendering it in a more viewable way and whether all decisions (e.g. Audubon core term change decisions) should also go there, but that's not DwC's problem. @peterdesmet Do we need to have a link to the table at http://rs.tdwg.org/dwc/ or don't we care?

[peterdesmet]

I think it is fine like this.