Re: RDF's relative IRI resolution is ambiguous

> On Sep 5, 2015, at 2:33 PM, Eric Prud'hommeaux <eric@w3.org> wrote:
> 
> * Gregg Kellogg <gregg@greggkellogg.net> [2015-09-05 11:38-0700]
>>> On Sep 5, 2015, at 11:18 AM, Andy Seaborne <andy@apache.org> wrote:
>>> 
>>> On 04/09/15 12:44, William Waites wrote:
>>>> I agree that long term curation and maintenance of test suites is a
>>>> good idea. I wonder if it is wise to rely in the long term on Github
>>>> -- who knows how long it will live, it's a private company, etc. It
>>>> might be better for the source code repositories to live at the W3C.
>>>> 
>>>> -w
>>>> 
>>>> --
>>>> William Waites <wwaites@tardis.ed.ac.uk>  |  School of Informatics
>>>>   http://tardis.ed.ac.uk/~wwaites/       | University of Edinburgh
>>>>         https://hubs.net.uk/             |      HUBS AS60241
>>>> 
>>>> The University of Edinburgh is a charitable body, registered in
>>>> Scotland, with registration number SC005336.
>>>> 
>>>> 
>>> 
>>> [Tried the CG mail list but got "550 Unrouteable address”]
>> 
>> I was able to send a message to the list (cc’d again on this message). Next week we can drum up more support with targeted messages to RDF and SPARQL lists and make a plan.
>> 
>>> Valid concern.
>>> 
>>> We can start now on github and make sure we have a plan.
>> 
>> Eric suggested this, where we use URLs on w3.org (presumably redirects from the existing test suite locations) and redirect to GitHub. That gives us control if GitHub goes away, or a better alternative comes. Obviously, we’ll need some support from W3C staff for this to work. We do this now for the CSV on the Web Test suite (http://www.w3.org/2013/csvw/tests/), and it works pretty well.
>> 
>> One probably with putting tests on GitHub is the inability to set HTTP headers; this is used by the JSON-LD tests, but those are hosted on a different server, with push-receive updates from GitHub. If someone were able to host such a service for tests at large, this might be another intermediate.
> 
> This is exactly what I was worried about. It sucks not controlling /etc/apache2/magic . I figured that this meant that GET's from github wouldn't give nice media types like text/turtle. (I'm hoping that mod_proxy will allow me to override the media type.) Any idea if we're in for more pain than that?

It looks like it gets Turtle and JSON-LD:

curl -I http://w3c.github.io/csvw/tests/manifest.ttl => text/turtle
http://w3c.github.io/csvw/tests/manifest.jsonld => application/ld+json

But, there’s no control for other types, or for additional headers. A more intelligent proxy, that could honor a .htaccess file in that directory, would be more useful. Or, a post-receive hook to just mirror the content of the GitHub (or whatever) site, and allow apache config to do what it’s supposed to. I think Ivan had some concerns about this.

Gregg

 
>>> A simple one would be to have CG reports which are releases of github
>>> work (or even just a dated copy).
>> 
>> Yes, i think using gh-pages for the released version of test suites, with branches for things which are in development, allows for reasonable curation. A release can then coincide with a report.
>> 
>> Gregg
>> 
>>>  Andy
>>> 
>>> 
>> 
>> 
> 
> -- 
> -ericP
> 
> office: +1.617.599.3509
> mobile: +33.6.80.80.35.59
> 
> (eric@w3.org)
> Feel free to forward this message to any list for any purpose other than
> email address distribution.
> 
> There are subtle nuances encoded in font variation and clever layout
> which can only be seen by printing this message on high-clay paper.

Received on Saturday, 5 September 2015 21:59:17 UTC