Re: Default RDF serialization

On 28/08/12 12:46, Kingsley Idehen wrote:
> On 8/28/12 7:31 AM, Olivier Berger wrote:
>> Hi.
>>
>> David Booth <david@dbooth.org> writes:
>>
>>> FWIW, if the LD profile is going to recommend one RDF serialization as
>>> the default for RDF, I would argue strongly that it should be Turtle
>>> instead of RDF/XML, because:
>>>
>>>   (a) Turtle is far more human friendly to read;
>>>   (b) RDF/XML is not XML Schema friendly;
>>>   (c) RDF/XML has XML-based restrictions (such as prohibiting local
>>> names
>>> that start with a digit) that make certain RDF difficult to represent;
>>>   (d) RDF/XML has had a history of misleading developers who are
>>> familiar
>>> with XML (but not RDF) into thinking that RDF is just a kind of XML.
>>>
>> Let me add my 2 cents :
>>
>> - IMHO, Turtle (or another human readable format) should be used for
>>    writing the specifications (i.e. illustrating the /model/ part if
>>    there's some separation between model and implementations), and
>>    providing examples, in order to be as much appealing to non-RDF
>>    experts as possible.
>>
>> - I'm not sure Turtle support should be a MUST when RDF/XML would only
>>    have a lower requirement for compliance to the specs of any LDP
>>    service. At the moment, but this may change, I'm not sure the level of
>>    support of Turtle in open source libraries in different languages is
>>    as good as RDF/XML's. For instance, in PHP, if we want to offer a
>>    migration path for lots of REST PHP apps that may become LDP aware
>>    some day. The good points of RDF/XML is that you may do some parsing
>>    with an XML (DOM) parser with some level of success on the client side
>>    even though you may miss some RDF constructs (like reification, etc.),
>>    whereas with Turtle, one may be just limited by no supporting library
>>    available on the hosting server side.

If you mean PHP developer working directly on the XML, then, in 
practice, that isnt going to work out.  An app will need a library to do 
the XML->RDF step.  The grammar [1] over the infoset is not trivial 
although there is a RelaxNG (informal) grammar.  Just rdf:ID vs 
rdf:about is confusing enough.  The variability of the syntax makes it 
more than just a case of missing a few features.

Can we turn this into a concrete question?

    Which libraries don't support Turtle?

There are several for PHP that do (RAP, ARC2, and Redland with PHP 
bindings).

	Andy

(And a system must have an N-triples parser to pass the RDF-2004 test 
suite ... because the results are recorded in N-triples)

[1] http://www.w3.org/TR/REC-rdf-syntax/#section-Infoset-Grammar

>>
>> Of course, depending on the ETA of LDP's deliverables, the last remark
>> may be obsoleted as Turtle gains wider support.
>>
>> I'd like to make sure we think about non-Java adopters, for instance,
>> and consider the level of support of the different RDF serializations in
>> various languages before me make definitive choices.
>>
>> Just my 2 cents,
>>
>> Best regards.
> To make anything bootstrap to point of critical mass on the Web you need
> the following critical characteristics associated with Web documents:
>
> 1. cut and paste pattern friendliness -- when editing content
> 2. human readability -- when reading content
> 3. save to storage (disk, folder etc..) -- when saving new or updated
> content.
>
> In addition to the above, an document author shouldn't be impeded or
> distracted by:
>
> 1. denotation mechanism choices -- for documents of document subjects
> and topics
> 2. domain name ownership and DNS server admin privileges
> 3. name/address disambiguation matters
> 4. web server ownership and admin privileges -- for re-write rules,
> indirection, and content negotiation in general.
>
> A user should craft a document, save it, then share its hyperlink with
> the world or a group.
>
> With the above in mind, you ultimately end up with Turtle, its the
> format that passes all the tests above. All of the other formats are
> really aimed as specific profiles which I break down as follows:
>
> 1. Turtle -- everyone

+1
Including any web developer who needs to work without a library of some 
sort.

> 2. HTML5+MicroData -- HTML application developers
>
> 3. (X)HTML+RDFa -- Well-Formed HTML application developers with an
> (X)HTML background and preference
>
> 4. JSON-LD -- Javascript developers that prefer JSON based data
> representation over XML based data representation
>
> 5. RDF/XML -- early community of developers and academics associated
> with the Semantic Web Project.

There were, I believe (I wasn't there), 2 reasons for RDF/XML

1/ internationalization and charsets.

2/ Integration into large XML documents.

Neither of these apply nowadays.

> There is ample history in place re. the effects of formats on both
> Linked Data and the broader Semantic Web vision. There are no reasons
> for repeating disastrous missteps from the past. None whatsoever.
>
> The Web is not about programmers, its for everyone. That means a
> plethora of user profiles.
>
> Links:
>
> 1. http://bit.ly/RJzd9S -- Why Turtle Matters .
>

Received on Tuesday, 28 August 2012 13:27:40 UTC