W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2012

Re: error reporting

From: Steve Harris <steve.harris@garlik.com>
Date: Thu, 27 Sep 2012 13:33:04 +0100
Cc: Andy Seaborne <andy.seaborne@epimorphics.com>, public-rdf-dawg@w3.org
Message-Id: <7942F122-6D49-4D27-AA0D-D6D9BC512391@garlik.com>
To: Lee Feigenbaum <lee@thefigtrees.net>
On 2012-09-27, at 13:07, Lee Feigenbaum wrote:

> On 9/27/2012 7:40 AM, Andy Seaborne wrote:
>> On 26/09/12 01:13, Arthur Keen wrote:
>>> Having a standard way to report errors makes sense in general and would
>>> help newcomers (like Algebraix) in the market, by increasing
>>> interoperability with existing client libraries and applications.
>>>  However we agree that it would be risky to add it to the spec right at
>>> the end of the process and we  do not wish to hold up the
>>> standardization process for SPARQL 1.1, however we ask that error
>>> messages be considered in subsequent standardization activity beyond
>>> SPARQL 1.1
>> The issue I see is that SPARQL exists in a wider context of standards and code.
>> HTTP is widely (!) deployed and well understood.  Many client-side toolkits exist to work with it.
>> Reuse is good. Adding additional mechanisms is good to make the system better for one purpose (SPARQLing) but comes with a cost of making general web frameworks toolkits less suitable for app building and adds to web developer overload.  This cost is something we all (self included) tend to underplay when we are in the midst of a technical design.
> I don't think this particular use case is well-understood in HTTP, is it? Is there a standard understanding for how to report (detailed) errors over HTTP with respect to media types?

Not as far as I can tell.

Web servers tend to return HTML. e.g.
$ curl -I -H 'Accept: image/jpeg' 'http://www.google.com/images/404.jpg'
This is ignoring the Accept header, and file extension (as you'd expect maybe).

SOAP has some kind of fault message, but I don't understand it well enough to comment.

REST APIs don't seem to have much commonality - some return plain text bodies, some custom XML/JSON/whatever payloads with error messages in, some use HTTP headers - not seen any that use custom "reason phrases", or HTML yet, but it's a big world

e.g. https://www.developergarden.com/index.php?id=377&L=0&vURL=ch03s05s03.html
or $ curl -I 'https://graph.facebook.com/search?q=alice&type=foo'

- Steve

Steve Harris, CTO
Garlik, a part of Experian
+44 7854 417 874  http://www.garlik.com/
Registered in England and Wales 653331 VAT # 887 1335 93
80 Victoria Street, London, SW1E 5JL
Received on Thursday, 27 September 2012 12:33:38 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:07 UTC