These are chat archives for frictionlessdata/chat

7th
Sep 2017
Rafael Pereira
@rflprr
Sep 07 2017 03:54
@pwalsh, @rufuspollock — Two questions related to extending the specs:
1) At data.world, we see applications for table schemas going beyond CSV and tabular JSON resources. For example, we have a lot of Excel users on our site. For some formats, like Excel, we need to extend the spec to accommodate multi-table schemas (i.e. each worksheet in an Excel workbook is a different table). Are you aware of any prior discussions on this? Do you have anything to recommend?
2) With respect to spec extensions, is there a process for "registering" them, so to avoid duplicative and/or conflicting work?
Paul Walsh
@pwalsh
Sep 07 2017 05:09

@rflprr

Point 1 -

Sure, we already support excel, ODS, google sheets and other formats in our implementation libraries, and platforms that build on them like OpenSpending. Not sure about if/how we elevate this to be explicit in the specs (it is easy to do; not sure it will be a good decision though, need to think about it).

Also on multiple sheets in a book (Excel, ODS, Google Sheets, etc.). We work pretty closely with Eric at Metatab ( http://metatab.org/ - which is essentially a DP descriptor as CSV ) and basically, we are working towards directly integrating that as an official serialisationof the specs.

Point 2 -

We need to address how we maintain the registry going forward, as this is starting to become quite important for various things we have coming in the next 6 months (additional DP profiles, semantic concept profiles, etc.), but for now, there is a minimal process I described a while back here, which you can follow so we can start a review of any new profiles https://github.com/frictionlessdata/specs#adding-a-new-specification

Ethan Welty
@ezwelty
Sep 07 2017 17:18
Congrats on 1.0! Before I put any work into lining up https://github.com/ezwelty/dpkg with the new schema, I'd like to have a bit of guidance on the state of "official" R implementations. https://github.com/frictionlessdata/datapkg-r hasn't been touched for a year and is very barebones, while https://github.com/frictionlessdata/datapackage-r is currently just a placeholder. I gather you want the implementation to follow this: https://github.com/frictionlessdata/implementations and therefore datapackage-r and tableschema-r are the future ?
roll
@roll
Sep 07 2017 17:27
@ezwelty New iteration of R libraries work are happening here - https://github.com/okgreece/tableschema-r, https://github.com/okgreece/datapackage-r
abhiraaj
@abhiraaj
Sep 07 2017 17:33
hi