RE: Remaining Options for SOSA-SSN Integration, Was: SOSA/SSN integration architecture

Hi Rob,
Can you put the votes back  on the reduced wiki page please? I don’t think anyone who took the trouble to express an opinion there misunderstood what they were doing.  If you want to remove this sentence from my comment, that is fine by me: “Btw I don't think the "A potential technical implementation..." works because there is no OWL mimetype -- only RDF-XML and tutrle which does not help.”

Having said that, there is indeed a move to develop content negotiation  around “profiles” or “formats” or “data models” through IETF – our own Lars is working on it.

-Kerry
From: Rob Atkinson [mailto:rob@metalinkage.com.au]
Sent: Tuesday, 21 March 2017 12:14 PM
To: Rob Atkinson <rob@metalinkage.com.au>; Maxime Lefrançois <maxime.lefrancois@emse.fr>; public-sdw-wg@w3.org
Subject: Re: Remaining Options for SOSA-SSN Integration, Was: SOSA/SSN integration architecture

We can also look at the specific recorded votes from absent commenters (after our discussion to agree on the accuracy of the option descriptions) and see if they raise any valid issues we havent properly considered.

Rob

On Tue, 21 Mar 2017 at 12:10 Rob Atkinson <rob@metalinkage.com.au<mailto:rob@metalinkage.com.au>> wrote:
Armin - our emails crossed, so I haven't wilfully ignored your request re votes.

My opinion is still the same though - the votes relate to a wider set of options, and now technical clarifications have been made to better explain the exact difference between options then these votes are now out of context, but preserved in the original context and not lost.

Happy to iterate to try to improve the option descriptions so  we are all more likely to be on the same page in interpretation of the options.

Rob

On Tue, 21 Mar 2017 at 12:07 Rob Atkinson <rob@metalinkage.com.au<mailto:rob@metalinkage.com.au>> wrote:

Hi

I've take the liberty of correcting the key technical details per option, and removing the copy of the old votes (which are still on the original page).

In particular I've tried to clarify the confusion caused by loss of the mime-type discussion of Option 5, and its subsequent unwarranted transference of implication to option 8 - and restore the correct distinction:

Option 1 - cannot go from SOSA to OWL axioms
Option 5 - relies on content negotiation to find OWL axioms bundled with SSN extensions
Option 8 - use owl:imports to find OWL axioms and keep SOSA axioms separate from SSN narrowed semantics

Again, I reiterate my concern that the use of the phrase "SOSA imports SSN" to characterise option 8 is incorrect unless you explicitly interpret SSN to be nothing more that the axiomitisation of SOSA, and I dont believe anywhere we have agreed on that exact definition.

I think that we can revisit the three options now with greater focus, and without reference to other options we have taken off the table.

Rob

On Tue, 21 Mar 2017 at 10:13 Rob Atkinson <rob@metalinkage.com.au<mailto:rob@metalinkage.com.au>> wrote:
Thanks Maxime, very helpful

I would strongly suggest we remove existing votes and rationales, and review the pros and cons first and make sure we agree,

For example,

1) for Option 1 the con Armin pointed out was there was no means to discover the stronger axiomitisation of a SOSA term

2) Option 8 is characterised as requireing content-negotiation to discover SOSA+OWL - that is actually the CON for Option 5, Option 8 is a solution which explicitly avoids the con

3) the "Con" reported for Option 8 is that OWL editors may automatically follow owl:imports when the user may not want them to. (Personally that seems like a strange corner case and the user's problem in choice of how to use tool





On Tue, 21 Mar 2017 at 00:52 Maxime Lefrançois <maxime.lefrancois@emse.fr<mailto:maxime.lefrancois@emse.fr>> wrote:
Dear all,

This is related to  ISSUE-139 and ISSUE-146

Following the decisions made today at the F2F meeting, I created a wiki page with the three remaining options:

https://www.w3.org/2015/spatial/wiki/Remaining_Options_for_SOSA-SSN_Integration


This page contains the description of the three options and describes the main pros and cons of each of them.

Best,
Maxime