More transparency would indeed be nice. But there is that balance between documenting all the time and getting stuff done...
As to an executive summary or spreadsheet of CDM issues, we have some similar stuff in place and options for others. In the general case, projects form an intertwingly web, not a nice neat spreadsheet. A "central list" presumes there is a center that works for everybody. * #473<https://informatics.gpcnetwork.org/trac/Project/ticket/473> is the task of getting GPC sites to CDM prep-to-research readiness, with milestone/data-domains3<https://informatics.gpcnetwork.org/trac/Project/milestone/data-domains3> serving as a scheduling mechanism. To the extent that Russ, Jim and I are the center of GPC dev (leaning on MCRF for CDM), that's the basis of the central list. Our pattern for these "all sites ..." tickets is to start with one ticket where the reporter and/or owner are responsible to nudge all the sites along. Then, once a critical mass of sites are done, we close that ticket and make sure there are separate tickets open for any remaining stragglers. * The depgraph for #473<https://informatics.gpcnetwork.org/trac/Project/depgraph/ticket/473> is one summary view. * The table of tickets on data-domains3<https://informatics.gpcnetwork.org/trac/Project/query?group=status&milestone=data-domains3> is another view; it's quite spreadsheet-like. * At KUMC, we typically create a ticket to represent our part of the "all sites ..." business pretty early in the process. In this case, That's #486<https://informatics.gpcnetwork.org/trac/Project/ticket/486> "adapt i2p-transform for HERON ETL results for SAS CDM v3 Data Characterization". You're more than welcome to open one or more tickets for issues at WISC. * #486 includes various i2p-transform issues by reference (mostly on the SCILHS repo, but a few on the kumc-bmi fork). Both trac and github let you subscribe to issues, so you can get email on exactly those issues you're interested in. Github lets you watch a whole repository too. Those are primarily mechanisms to keep track of who has the ball to do what and to track changes as they happen. They provide some documentation, though it's more like a transaction log (delta) and what readers typically want is more the current state of the database than the transaction log (sigma). The gpc-dev archive<http://listserv.kumc.edu/pipermail/gpc-dev/> is also more of a delta than a sigma. (See also The Bees and the Ants<https://www.w3.org/2002/11/bees-and-ants/paper.html>.) Options for providing more of a summary include: * Trac supports ad-hoc keyword that span milestones etc.; we have several tickets with a CDM keyword<https://informatics.gpcnetwork.org/trac/Project/query?keywords=%7ECDM&col=id&col=summary&col=keywords&col=status&col=owner&col=type&col=priority&order=priority> and some with v3<https://informatics.gpcnetwork.org/trac/Project/query?keywords=%7Ev3&col=id&col=summary&col=keywords&col=status&col=owner&col=type&col=priority&order=priority>. I haven't been very active in curating those keywords (as opposed to data-quality, obesity-cohort, and a few others), but you're welcome to do more of that sort of thing. * In the internal KUMC trac, I made a CDMv3Readiness wiki page; I haven't feel a need for that in the GPC context nor i2p-transform, but if you do, you're welcome to start one and see if it picks up steam. * I do maintain a NAACCR_ETL<https://informatics.gpcnetwork.org/trac/Project/wiki/NAACCR_ETL> page, which is perhaps in the same genre You're welcome to do something with google docs to organize things from your perspective, but for CDM and i2p-transform, we have more than enough lists already, it seems to me. As to a central ontology, that would be DataRepositoryManagement#gpc-ont<https://informatics.gpcnetwork.org/trac/Project/wiki/DataRepositoryManagement#gpc-ont> (available on babel, on a good day). We're more agreed to some parts than others, but that's the nature of the beast, as I see it. p.s. I presume it's OK to share your last message with the group, Keith. -- Dan ________________________________ From: Wanta Keith M [[email protected]] Sent: Tuesday, June 07, 2016 1:06 PM To: Mish Thomas Cc: Dan Connolly Subject: RE: CDM v3 Appendix B fixes - pencils down? Many of the tickets on the wiki are related to other contexts besides CDM. So I guess this list would be specific issues for CDM. Some of the issues are in GitHub comments (like Nathan Graham mentioned), and some are in actual tickets, and some are not posted. I guess it would be nice to have more transparency on what each site is fixing. It won’t be perfect ever, but it would be good to have some type of executive summary. I’d rather not make a GPC ticket for everything I find and correct because I’d be documenting all the time. But if there’s just a document, I’d be more diligent about it. From: Thomas Mish [mailto:[email protected]] Sent: Tuesday, June 07, 2016 12:50 PM To: Wanta Keith M Cc: Dan Connolly Subject: RE: CDM v3 Appendix B fixes - pencils down? Keith, The issue list should pull from the issues discussed as a part of the GPC-DEV list. Speaking for Dan, this is why he asks us to bring these issues up as we find them to the list - so that others can see/comment on the issues as we go along. TM Thomas Mish School of Medicine and Public Health UW-Madison Sent by Outlook<redir.aspx?REF=LSARLZalEnejrd0R0RGcSG7kKipRSKvMajz5I2FUa4ZUfy_Hxo_TCAFodHRwOi8vdGFwcy5pby9vdXRsb29rbW9iaWxl> for Android From: Wanta Keith M Sent: Tuesday, June 7, 12:36 PM Subject: RE: CDM v3 Appendix B fixes - pencils down? To: 'Dan Connolly', [email protected]<redir.aspx?REF=aDkAnxUxG9Y7BP3So911bDkP2MfHhtSIypozRdMrq59Ufy_Hxo_TCAFtYWlsdG86Z3BjLWRldkBsaXN0c2Vydi5rdW1jLmVkdQ..> GPC - It’s great that we have multiple code repositories that you are willing to share for our CDM scripts. I will reference the one that KUMC manages for sure. I’m not so concerned with fixing specific issues. All of our organizations have great data analysts to figure the issues out. I’m not concerned about fixing specific issues. What I’m asking for is if we can have a central list of issues we have fixed in our CDRN, so that if someone joins the group or someone or someone has a question, or we begin to organize an ontology (with full documentation) for our CDRN, we can simply look at that list and see who had the same issue with their i2b2 and CDM scripts, and what the resolution was, in plain English, along with a reference to the code repository. Some software companies would refer to this as an Executive Summary. Then we can start to establish major and minor versions for CDM within GPC. It might seem like lot of work to maintain such a list, but I’m sure everyone would benefit. And if we used something similar to Google Docs (or some tool) for such a list, I’m sure it would be useful for our group. Eventually (I’m hoping), we get a central ontology for our GPC CDRN, that we can all agree to as a group. If we can’t organize a simple spreadsheet of CDM issues we’ve run into as a group, I think we’re asking too much to manage an ontology as a group. Let me know if someone disagrees. Keith Wanta (WISC) From: Dan Connolly [mailto:[email protected]<redir.aspx?REF=LFIQP-zn6kY8zblbNjUBN_9KCanGjC3cYzTIVwC4sEJUfy_Hxo_TCAFtYWlsdG86ZGNvbm5vbGx5QGt1bWMuZWR1>] Sent: Monday, June 06, 2016 3:41 PM To: Wanta Keith M; [email protected]<redir.aspx?REF=XaKA-VRC8YdkOYslq8UuhcKIorFxgo7If8pWP9e-WNFUfy_Hxo_TCAFtYWlsdG86R1BDLURFVkBMSVNUU0VSVi5LVU1DLkVEVQ..> Subject: RE: CDM v3 Appendix B fixes - pencils down? If you're asking whether you should track your i2p-transform and shilhs-ontology changes using git and github, then yes, I highly recommend it. You're more than welcome to fork https<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..>://github.com/<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..>kumc<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..>-<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..>bmi<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..>/i2p-transform<redir.aspx?REF=xjjb3xP7yvHa3CAL9cJ_nTBBI7yXUCIR6_W_BlRPVexUfy_Hxo_TCAFodHRwczovL2dpdGh1Yi5jb20va3VtYy1ibWkvaTJwLXRyYW5zZm9ybQ..> . Newer versions are typically better than older versions. In particular, for kumc-bmi/i2p-transform, we do new development on branches and only merge reviewed code to master. This is mostly just platitudes; If you can give an example of which CDM v3 fix you have in mind, I might be able to give a more useful answer. -- Dan From: gpc<redir.aspx?REF=XRF-BEnDCOVYvuR7RI5ehg4-4KX478CTUeQ9DDIP1X9Ufy_Hxo_TCAFtYWlsdG86Z3BjLWRldi1ib3VuY2VzQGxpc3RzZXJ2Lmt1bWMuZWR1>[email protected]<redir.aspx?REF=Nwt2DamOQ9n4UDjZEmXplEnzGuaYeBKXwDq50oe1jVC14DHHxo_TCAFtYWlsdG86Z3BjLWRldi1ib3VuY2VzQGxpc3RzZXJ2Lmt1bWMuZWR1> [[email protected]] on behalf of Wanta Keith M [[email protected]] Sent: Monday, June 06, 2016 3:02 PM To: [email protected]<redir.aspx?REF=xNsQvqqDbpYlV_qIdJQYYpy6DhniXbi002QQZAjB3Z614DHHxo_TCAFtYWlsdG86R1BDLURFVkBMSVNUU0VSVi5LVU1DLkVEVQ..> Subject: CDM v3 Appendix B fixes - pencils down? GPC – Have we make a decision on whether we should be making our CDM v3 Appendix B fixes based on some past revision of the i2p-transform and scilhs-ontology repositories, or…. Should we be applying code changes overtime from GitHub for CDM v3 Appendix B fixes? (putting aside ontology mappings we do internally) Thanks, Keith Wanta (WISC)
_______________________________________________ Gpc-dev mailing list [email protected] http://listserv.kumc.edu/mailman/listinfo/gpc-dev
