AW: Question on RDF standardisation and large language models

Thanks a lot for the detailed feedback, Adrian.
It is good to see that the “living standard” option is being discussed. I personally would be all for it.

Thanks also for the other pointers. In
https://github.com/w3c/sparql-dev/issues/40

I like the reference to XQuery full text search. Re-using such definitions indeed can safe a lot of time.
In the vector search space I am not aware of other languages to look at, something to look up.

As you said, it boils down to the community being serious about the topic. Statements of interest are very welcome 😊

Best,

Felix

Von: Adrian Gschwend <adrian.gschwend@zazuko.com>
Datum: Mittwoch, 13. Dezember 2023 um 05:21
An: public-rdf-star-wg@w3.org <public-rdf-star-wg@w3.org>
Betreff: Re: Question on RDF standardisation and large language models
[Sie erhalten nicht häufig E-Mails von adrian.gschwend@zazuko.com. Weitere Informationen, warum dies wichtig ist, finden Sie unter https://aka.ms/LearnAboutSenderIdentification ]

On 13.12.23 02:40, Sasaki, Felix wrote:

Hi Felix,

> My question to the co-chairs and to Pierre-Antoine is: would taking up
> such a topic require
>
>  1. No changes to the charter
>  2. Re-charting
>  3. A new working group, to start after we are finished, i.e. after
>     August 2024 the earliest.

One of the ideas is to move to a living standard at some point, see:

https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.w3.org%2FArchives%2FPublic%2Fpublic-rdf-star-wg%2F2023May%2F0066.html&data=05%7C02%7Cfelix.sasaki%40sap.com%7C2844226208f94e3b9b8608dbfb92f842%7C42f7676cf455423c82f6dc2d99791af7%7C0%7C0%7C638380380903375336%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=BkZzcmtn%2BdfX0%2BrYxJXRWpmNyRQNUdnWtmG8VwedUrA%3D&reserved=0<https://lists.w3.org/Archives/Public/public-rdf-star-wg/2023May/0066.html>

We would be more agile that way. But to be clear, that's an idea, not a
decision.

For extending SPARQL, there is activity in the SPARQL-Dev group:

https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fsparql-dev&data=05%7C02%7Cfelix.sasaki%40sap.com%7C2844226208f94e3b9b8608dbfb92f842%7C42f7676cf455423c82f6dc2d99791af7%7C0%7C0%7C638380380903531631%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=9lHUvysuRXYPDkb%2Fqgyzh1jsTtqo78zMXqgq5txcxSk%3D&reserved=0<https://github.com/w3c/sparql-dev>

A good start would be to create an issue there and link to the existing
work, like AllegroGraph and maybe some PG vendors you mention.

I personally started using OpenSearch as vector database and at least in
Fuseki, there was support for Elastic at some point. So my idea is to
check if this can be resurrected and extended so that vector search is
possible as well. I think it would be close to what most vendors do with
Lucene based search, which is by the way also not standardized, see for
example this issue:

https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fsparql-dev%2Fissues%2F40&data=05%7C02%7Cfelix.sasaki%40sap.com%7C2844226208f94e3b9b8608dbfb92f842%7C42f7676cf455423c82f6dc2d99791af7%7C0%7C0%7C638380380903531631%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=qUJ0q6iJAftDup%2ByGQYgXuPH%2FkFPwD9YNtvCXDDXR5U%3D&reserved=0<https://github.com/w3c/sparql-dev/issues/40>

If one or more contributors are serious about an issue, they can move to
SEP:

https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fsparql-dev%2Ftree%2Fmain%2FSEP&data=05%7C02%7Cfelix.sasaki%40sap.com%7C2844226208f94e3b9b8608dbfb92f842%7C42f7676cf455423c82f6dc2d99791af7%7C0%7C0%7C638380380903531631%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=gW7%2Faikz%2FyY1RL8SODPp5lKC4Zt76wCd9K3YmxcFynA%3D&reserved=0<https://github.com/w3c/sparql-dev/tree/main/SEP>

Some of them are implemented in some stores. Andy Seaborne and Thomas
Tanon are active for example, among others.

But as usual it depends on how active and serious the community is about
an issue, without at least one person taking care of it, nothing will
happen.

regards

Adrian

--
Adrian Gschwend
CEO Zazuko GmbH, Biel, Switzerland

Phone +41 32 510 60 31
Email adrian.gschwend@zazuko.com