[the-datatank] Tijd om TDT te gebruiken

[Jan Vansteenlandt]

Hi guys,

Indeed an interesting website (data.ofkn.org), their main focus however of
that website is creating/playing with a new data standard for tabular data.
Perhaps when this standard (datapackage/SFD) is really getting used by
people we should consider supporting it ourselves. As far as I can see, the
package mainly exists of meta-data in a json file, combined with the actual
data.

As for the pragmatic roadmap goes, yes testing needs to be implemented way
more, testing on formatters and headers requires some more work though than
the tests we currently have. We also need to use broken datasets, in order
to check how our code responds to incorrect data structures (e.g. a CSV
file with 4 columns, but has a defined column count of 5).

Kind regards,

Jan


On Sat, May 11, 2013 at 4:58 PM, Pieter Colpaert
<pieter.colpaert at okfn.org>wrote:

> Hi all,
>
> http://demo.thedatatank.org/datasets.about
>
> I'm adding datasets over here from http://data.okfn.org
>
> It's really amazing how many problems still pop up when using the
> visualizers, or when using spectql, or when just trying to use TDT as as
> a data proxy.
>
> We need to start listing up all the features that are or should be
> stable right now, list them on http://thedatatank.com/features and start
> working towards a release than can be used by everyone asap.
>
> What would be a pragmatic roadmap right now for the next *stable*
> release? Writing more tests first? Doing more user testing?
>
> What extra people do we have to engage in order to make The DataTank a
> platform that just works? Testers? Unittest writers?
>
> Kind regards,
>
> Pieter
> --
> OKFN Open Transport Working group
> OKFN Belgium vzw/asbl
> +32 (0) 486/747122
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/the-datatank/attachments/20130511/331864c9/attachment.html>