[ckan-dev] Resource types

Mon May 28 12:10:51 UTC 2012

On 28 May 2012 10:43, Toby Dacre <toby.okfn at gmail.com> wrote:
>
>
> On 25 May 2012 18:43, David Read <david.read at hackneyworkshop.com> wrote:
>>
>> Toby,
>>
>> Rufus and Richard Cyganiak worked hard on this and set out the current
>> datahub guidelines to be mimetype in preference to the well-known file
>> extension. This discussion would benefit from recalling what the
>> reasons were.
>>
>> I won't repeat my recent posts about lowercase vs mixed case, so do
>> check the logs.
>>
>> You could also look at the various things we did in DGU with the
>> indexing, form autocomplete, migrations etc. Having a list of 'good'
>> formats to suggest, and you could consider producing a list of
>> standard migrations, like 'test/csv' -> 'CSV' and 'Excel' -> 'XLS' for
>> example.
>>
>> I can't see the logic of creating a new field to store the 'unified'
>> format alongside the 'raw' format as the user typed it. Just change
>> the original field value - it's pretty obvious that csv, CSV and
>> text/csv are all the same, so why keep the mixture?
>>
>
> I learnt over the years that just changing peoples data is generally a bad
> thing.

Yeah, that was the established view for many years. And then they
invented the wiki and most people realised that the benefits of
collaboration outweighed the problems of people making bad
'corrections'. If someone do make a mistake then the history is there
to refer to. This is one of CKAN's core principles.

Dave

>  Maybe later we want to split out 2 things that we had merged
> initially,  Also it is extremely unlikely but some users may consider csv
> and CSV to be different things and have explicitly coded them up as such.
> To be honest I haven't gone as far as looking how this data is added yet.  I
> was just aware that there have been some conversations.
>
> Thanks for the feedback
>
> Toby
>>
>> Dave
>>
>> On 24 May 2012 14:04, Toby Dacre <toby.okfn at gmail.com> wrote:
>> > Currently resource types are a little confused.
>> >
>> > csv, CSV, text/csv should all be treated as the same type as far as
>> > search
>> > faceting is concerned for example
>> >
>> > It seems to me that we should really have this in the model but I'm not
>> > sure
>> > if we should have a new unified_format field or change the existing
>> > formats.
>> >
>> > I know there was some discussions on this subject in the past but not
>> > sure
>> > of the conclusion
>> >
>> > we need this for the demo-theme branch and I'd be happy to do the work
>> > if we
>> > can agree on how we want to do this.  It seems like the simple approach
>> > is
>> > to make all formats lowercase and ignore the first part of text/csv
>> > application/zip type formats we might also need to have some custom
>> > mappings
>> > for specific formats
>> >
>> > as ever what do people think/ objections etc
>> >
>> > Toby
>> >
>> > _______________________________________________
>> > ckan-dev mailing list
>> > ckan-dev at lists.okfn.org
>> > http://lists.okfn.org/mailman/listinfo/ckan-dev
>> >
>>
>> _______________________________________________
>> ckan-dev mailing list
>> ckan-dev at lists.okfn.org
>> http://lists.okfn.org/mailman/listinfo/ckan-dev
>
>
>
> _______________________________________________
> ckan-dev mailing list
> ckan-dev at lists.okfn.org
> http://lists.okfn.org/mailman/listinfo/ckan-dev
>