Re: Popularity of Property Graphs

Hi Kelvin:
Our first order of business is to convince the W3C to initiate a WG
to standardize a Property Graph model.   In my mind this is an abstract model
which may be represented in several ways.  Whether one of those is an extension
of RDF is a subsequent discussion, perhaps best deferred to the future WG.
All the best, Ashok

On 1/13/2014 12:29 PM, Kelvin Lawrence wrote:
> I feel I need to state that, for me at least, if we start to morph this work into a "how can we dovetail Property Graphs into existing RDF technologies" as a primary deliverable that we will lose most if not all of the rapidly growing property graph ecosystem folks interest in anything we produce. I seem to find myself continually having to state that a goal of this group must be to build upon existing work in the PG space. There are graph engines, query languages and serializationsbeing used and  deployed in customer scenarios today that do not use any RDF technology. I am fine with there being mappings to and from RDF but a primary reason I am interested in Property Graphs is that RDF has not been a good fit for the work I am doing and I think PGs are a good fit. We have some internal projects that found RDF so difficult to work with for what they are doing that they moved to PGs and are finding things much simpler.
>
> Therefore, speaking for myself and the groups at IBM that I am representing on this CG, I have no interest in a primary deliverable that basically says we morph PGs into RDF graphs and use SPARQL to do the queries. While I feel there is definitely a need for a declarative query language I do not think reverting to RDF is the way to go here. I really hope that whatever we do in this CG is complimentary to all of the work going on in open source around PGs today in the industry.
>
> I recall at the first meeting of this CG, Ivan reminded us that if we become RDF focused we will alienate the very people we are trying to embrace. I fear I see the start of that trend in this thread of discussion but I hope I am wrong.
>
> Cheers
> Kelvin
>
> Kelvin R. Lawrence
> Distinguished Engineer & CTO, Software Standards
> Member of the IBM Academy of Technology (http://www.ibm.com/ibm/academy)
> IBM Software Group
>
>
> -----Michael Petychakis <mpetyx@epu.ntua.gr> wrote: -----
> To: "<public-propertygraphs@w3.org>" <public-propertygraphs@w3.org>
> From: Michael Petychakis <mpetyx@epu.ntua.gr>
> Date: 01/13/2014 09:06AM
> Subject: Re: Popularity of Property Graphs
>
> Dear all,
>
> Unfortunately tomorrow i will not be able to attend the call due to a last minute flight.
>
> Personally, I would like to express my consideration (following the discussion from last week) on the fact that more and more companies are utilising property graphs while on the same time draw away from RDF technologies.
> The first deliverable of the recommendation (http://www.w3.org/community/propertygraphs/wiki/Recommendation) suggests that we describe the pg data model, which i am totally in favour of.
> This way, we could create the corresponding RDF serialisation and have SPARQL as the main query language instead of creating a new one. We could explore queries in cypher, gremlin and similar pg query languages and find a match of them in SPARQL.
>
> To quickly sum up, I totally agree with the starting point being a description of the data model for property graphs and later we could explore the correlation with the Semantic Web Cake( RDF, SPARQL, etc).
>
> Regards,
> Michael Petychakis.
>
>
>
> On Jan 8, 2014, at 1:02 AM, Ashok Malhotra <ashok.malhotra@oracle.com <mailto:ashok.malhotra@oracle.com>> wrote:
>
>> You are correct JSON-LD does have datatype support.
>> I guess I was thinking of JSON which has only string and number
>> All the best, Ashok
>>
>> On 1/7/2014 5:56 PM, Patrick Durusau wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> Ashok,
>>>
>>> On 01/07/2014 05:24 PM, Ashok Malhotra wrote:
>>>> Hi Patrick: I agree that we are staring to see more and more data
>>>> whose structure fits Property Graphs.
>>>>
>>>> I also agree that the models used by vendors are often
>>>> idiosyncratically different. So, yes, standardization of strings to
>>>> UTF-8  among other things would be very useful.
>>>>
>>>> But if we add support for datatypes we may leave JSON-LD behind
>>>> and there may be some objection to this.
>>> ???
>>>
>>> I'm not sure why adding property types would leave JSON-LD "behind?"
>>>
>>> http://www.w3.org/TR/json-ld/ reports at 6.3 that JSON-LD has
>>> non-normative support for datatypes and uses an XML Schema Part 2
>>> datatype as an example.
>>>
>>> Yes?
>>>
>>> Hope you are having a great day!
>>>
>>> Patrick
>>>
>>>> All the best, Ashok
>>>>
>>>> On 1/7/2014 4:30 PM, Patrick Durusau wrote: Greetings!
>>>>
>>>> There seems little doubt that property graphs are gaining ground
>>>> but I am less certain that the interchange of property graphs is
>>>> quite as well settled.
>>>>
>>>> For example, one of the non-participants thus far in the
>>>> propertygraphs discussion supports label and property name
>>>> characters outside the ASCII character set if and only if each
>>>> character is preceded by a back tick. Creating the first family of
>>>> "back tick" languages I suppose.
>>>>
>>>> Could that be transformed? Perhaps. I don't know how you would
>>>> apply the various Unicode algorithms since those are dependent upon
>>>> order i the original string. (Thinking of zero-spacing
>>>> characters.)
>>>>
>>>> Declarations about serializations that mandates the use of UTF-8
>>>> would be a step in the right direction. Possibly enumerate the
>>>> other aspects of a property graph in the abstract that may or may
>>>> not be supported.
>>>>
>>>> Allow multiple properties on nodes? Edges? Multiple nodes to one
>>>> edge? etc.
>>>>
>>>> Not mandating any serialization but giving serializations an
>>>> outline of what they may or may not choose to support.
>>>>
>>>> I can't imagine serious players using labels and property names
>>>> that don't support Unicode. (Remember name characters and start
>>>> name characters? Why repeat a mistake?)
>>>>
>>>> Hope everyone is having a great week!
>>>>
>>>> Patrick
>>>>
>>>> -- Patrick Durusau patrick@durusau.net <mailto:patrick@durusau.net> Technical Advisory Board,
>>>> OASIS (TAB) Co-Chair, OpenDocument Format TC (OASIS) Editor,
>>>> OpenDocument Format TC, Project Editor ISO/IEC 26300 Former Chair,
>>>> V1 - US TAG to JTC 1/SC 34 Convener, JTC 1/SC 34/WG 3 (Topic Maps)
>>>> Co-Editor, ISO 13250-5 (Topic Maps)
>>>>
>>>> Another Word For It (blog): http://tm.durusau.net Homepage:
>>>> http://www.durusau.net Twitter: patrickDurusau
>>>>
>>> - -- Patrick Durusau
>>> patrick@durusau.net <mailto:patrick@durusau.net>
>>> Technical Advisory Board, OASIS (TAB)
>>> Co-Chair, OpenDocument Format TC (OASIS)
>>> Editor, OpenDocument Format TC, Project Editor ISO/IEC 26300
>>> Former Chair, V1 - US TAG to JTC 1/SC 34
>>> Convener, JTC 1/SC 34/WG 3 (Topic Maps)
>>> Co-Editor, ISO 13250-5 (Topic Maps)
>>>
>>> Another Word For It (blog): http://tm.durusau.net
>>> Homepage: http://www.durusau.net
>>> Twitter: patrickDurusau
>>> -----BEGIN PGP SIGNATURE-----
>>> Version: GnuPG v1.4.11 (GNU/Linux)
>>> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
>>>
>>> iQIcBAEBAgAGBQJSzIYXAAoJEAudyeI2QFGoBRgQAPU6+loPYoGh6lgKVmFENrKl
>>> 0bqVTir+kE1uHmHJYy8XLR1Vvh8rUrfjq421fUNdNTMWdnnBYAC+M96YgZukB3l0
>>> Wlwnv6Kvy4l/WlD7Bd0RKK1aXyPA93/CGqkYmlc3OirQkAMusFa9udt2E+y20Jw4
>>> u1CpZCC+Eg++FvODIdo4AXAB2+69MLxFLAAKRAfv21Bda3vBTHFEVvJA1vXwZ1tC
>>> QbMzIO3oXhqwZxO3OYhbhHoTUevnj5odfWdgs01PCLb7mTzJheJnhiWDEMyq0b6g
>>> 1D6q/hx6/U2MbtWPqzptqzA9+E5dHLmwXKTLYdO/99D4onqp0MPFgx4Crgx4He7r
>>> nF1wPZ5IFNRXl00leYY3P39Uka36RwJ2mq9FO7ClsaUtv5LmV/SgL9AGW6LnWOao
>>> CKikNonn9qlMsg+PDOfu4R4unnzMEPGO3HeoklMcVA3k9nocjc8Or1X7m9F/gEI4
>>> GUXXEXBm0C4YhVLdWob/8+mU5lyT0SmZrqvPt/dpznmqSlTfJZ8/P3fAApgPXaO6
>>> p8JqSWCdTdF1fIosL8eXstAguLyJFN6AaR4vwfgnZkiDa50TvoQ1ZnLb8JD+/xnS
>>> 47/HLNIAGDqong6GlPE712pprFISoX6NcBXIUDW/m1JOtELlvvTQ5woJ2z7Zq9g5
>>> jFzuhnbRPu54oDzkqPJK
>>> =27yv
>>> -----END PGP SIGNATURE-----
>>>
>>
>>
>