16

The OpenData SE site is for:

Open Data Stack Exchange is a Q&A site for developers, researchers, and anyone else interested in open data.

and based on its definition of what is on-topic there seems to be an overlap with the proposal for Web Data (Semantic Web)

for people wanting to read, write and query data linked on the Web, find or develop ontologies, write apps that use technology built on the specs developed at the World Wide Web Consortium (including XML, JSON, JSON-LD, RDF, SPARQL, LDP, OWL, RIF,...)

What does Web Data (Semantic Web) cover, which community does it try to attract that Open Data doesn't cover and/or exclude? Shouldn't it be considered to address that scope with Open Data first?

  • 1
    thanks for the pointer. Perhaps that would be a good merger... I'll go over it. – Henry Story Dec 9 '18 at 11:19
  • I suppose this will be a good opportunity for people from the semantic web community to check if there is anything they would like to change to how the open data community is structured. – Henry Story Dec 9 '18 at 15:21
  • 1
    From opendata.stackexchange.com/tags?page=1&tab=popular, I would guess that the OD community/forum is mostly about datasets, less about data science/the underlying technology: The first technology-specific tag (json) occurs with frequency 37, whereas the most popular tag (data-request) occurs 2489 times. So, while OD does cover most aspects of the description, it simply doesn't seem to be the first place to ask a technological question. – Chiarcos Dec 10 '18 at 8:33
  • 1
    @HenryStory I suggest you post a meta question on the per-site meta of opendata.se to have the community there chime in, instead of us making all kind of assumptions about their scope. – rene Dec 10 '18 at 10:38
  • 1
    Note that it would be helpful to ask on opendata.meta.stackexchange.com what is out of limit there if anything really is that this community would be interested in. – Henry Story Dec 10 '18 at 17:47
  • 1
    @HenryStory I've dropped a message in their somewhat inactive chatroom – rene Dec 10 '18 at 20:51
  • @Chiarcos the data-request tag (2493) is huge outlier when it comes to question frequency. Most "popular" tags are used no more than 100 times. data.stackexchange.com/opendata/query/945972/… – philshem Dec 13 '18 at 20:19
  • @philshem: Yes, it is a huge outlier. But not because it is atypical (check the following tags, many of these are just resource-specific), but because it is apparently the very tag that people most typically think of in that forum ;) – Chiarcos Dec 14 '18 at 14:18
  • @Chiarcos wrt forums – rene Dec 14 '18 at 14:52
8

The "Web Data" proposal was not originally called like this. At first, it was named "Semantic Web". There have been discussions on the Semantic Web mailing list regarding the lack of a dedicated Q&A forum on the topics of the Semantic Web. Thus, Henry Story made this proposal in Area 51. But after that, the scope of the proposal was discussed further and some people thought that "Semantic Web" is too narrow to build a large enough community to make the proposal fly as a self-standing Stack Exchange site.

I personnally think that the change of name and scope is a mistake and the questionable overlap with Open Data SE shows that the name at least is not ideal, perhaps misleading.

On the one hand, the Semantic Web, as a vision, as a project, as a set of technologies, or as a research field, includes much more than data on the Web. On the other hand, Open Data includes things that have little to do—if anything at all apart from being on the Web—with the Semantic Web.

  • Why would not starting to ask the questions that would interest semantic web folks on the open data SE slowly change SE to being better for those interests? I don't think that anything stops anyone on StackOverlow adding questions about new programming languages? I remember that When Scala-JS was in beta there were allready a large number of questions there... These communities are not closed as far as I can tell... Perhaps their name could be widened... I asked them over there opendata.meta.stackexchange.com/questions/413/… – Henry Story Dec 12 '18 at 16:10
  • I am going to ask a few questions on the opendata forum to see how it works. One can only ask a question every 40 minutes, so it will be slow. I noticed that there are a few tags that are missing and one needs over 140 points to be able to create new tags. But I should be able to get there quite soon. Here is my first question: "How was SPARQL influenced by erotetics?" opendata.stackexchange.com/questions/13625/… – Henry Story Dec 13 '18 at 10:18
  • 2
    Oh, I just found a post on opendata.meta that is two years old "semantic web technology questions should go to a new stackexchange community" opendata.meta.stackexchange.com/questions/367/… – Henry Story Dec 13 '18 at 10:26
3

"Open Data" seems to be too narrow in that e.g. one interested in manipulating her closed RDF database (e.g. data that is licensed as "non-commercial") would not intuitively try to ask at "Open Data". Also, if you want to adress people who used answers.semanticweb.com naming "open data" to "web data" seems to be more fitting.

"Web" just adds a new quality to the name of the area, so that:

  1. "Data" is the "what"
  2. "Web" is the "where" implying the "how", i.e. a special kind of "technology"
  • 1
    btw, data are algebraic structures, the how - pragmatics, protocols - are coalgebraic. And in a way one needs a community that have questions that allow both sides to interact. – Henry Story Dec 10 '18 at 12:42
  • it wouldn't be intuitive to ask at OD, but a question about closed RDF database would be on topic. RDF is a technology of open data, but doesn't need to necessarily be "open" – philshem Dec 13 '18 at 20:21
  • True, could be a nice question. And because of this and because it feels counterintuitive I vote to rename. – dr0i Dec 14 '18 at 14:52
3

Open data is neither a subset nor a superset of linked data or the Web of Data but the topics clearly overlap.

The vision of the semantic Web and linked data is to create an infrastructure of connectable and exchangeable information sources in the web.

This means that data endpoints are interoperable, even if they never knew about each other, because the technology is independent of a concrete serialization format, query language and structure of the concrete information.

The semantic data technology stack deals with

  • mapping different data formats and structure
  • querying over a lot of independent data sources
  • providing logic (semantic) to infer new knowledge
  • developing domain specific vocabularies, meta vocabularies (also referred as ontologies) and concepts
  • deal with data validation, data expiration and versioning of volatile data
  • making data independent of the tools, that produce or visualize them
  • dealing with live data, sensor data, knowledge, archived data,... in a standardized way

this is not a complete list but there have already been various posts that try to line out the coverage of semantic web technologies.

The following graphics show this:

LOD Visualization

semantic web (taken from the-power-of-semantic-technologies-to-explore-linked-open-data)

open vs. private data

In fact linked data can be private and should be private in a lot of areas (think of the semantic desktop, private information management or cooperate data, classified information, user data,...). Ideally private, partially private and open data can be merged or treated in the same way. If the vision of the semantic web comes into reality sharing of arbitrary information and knowledge, would be as easy as exchanging and sharing files today using Versioning systems or online collaboration software, where users form groups of people and people can share information among those groups and single individuals, but also with virtual actors like organizations, services, data pools,... It is still a long way until all of it reaches a more user friendly way that spreads further from the research labs, enthusiasts and cooperate labs to the public. But until now serious afford has been done by the research community to solve a lot of issues and sub problems that arise from that vision. Standardization is one of the processes, that helps to keep a stable base, where the volatile semantic technology can emerge.

why people should care about the web of data

One important part is to help people to understand how to properly prepare their domain specific data structures, interfaces and algorithms to easily integrate it into the web of data universe. The benefit for all would be, that using and mixing of specific services, data sources and tools would be much easier and certainly would not need programmers knowledge or knowledge about vocabulary and ontologies if properly done.

Often it is much easier than a lot of users may think, if they hear all those complicated sounding terms out of the technology stack. In fact you don't need to know much, if at all about reasoning, higher order logic, triples and open world concepts, just to make a former JSON structure or REST service produce JSON-LD or semantically describe the resources it provides.

That is exactly the reason why I would like to see a stable community around all those tools that already exist. The open data-community would be the entry point for all those enthusiasts and experts the seek for an easy to use solution to get their domain specific data or where to put there own if they want to share it. The Web data-community would be the place where people ask about more technically related questions, concerning the whole technology stack, certain database systems, about ontology and vocabulary engineering tools, query languages, data structures, serialization formats, ...

Open Data - high level user perspective

Web Data - implementation, engineering low level perspective

A comparsion to a diffrent topic:

  • 3D world and game development

    • The process of creating and obtaining 3D scenes, models, textures, animations,... would be like the open Data community
  • 3D computer vision and graphics

    • The process of developing the necessary technologies to render, edit, draw the 3D computer graphics would be like the Web of Data community.
2

Here is a mail by Antoine Zimmermann to the semantic-web mailing list that could be seen as an answer to this question. In any case it is very informative. I copy it here for easier reading.

WIW, in Oct. 2009, a certain Andrew Matthews created a web site for Q&A on semantic web technologies, called Semantic Overflow (semanticoverflow.com). The software used on this site was Stack Exchange 1.0. It was good, it was fresh, it was lively.

A little later, Stack Exchange launched Area 51. I seem to recall that someone tried to get a Stack Exchange instance on semantic web topics via Area 51 in circa 2010, but it never flew up.

In the meantime, in 2011, Semantic Overflow was acquired by a company called semanticweb.com (now Dataversity). semanticoverflow.com ceased to work and redirected to answers.semanticweb.com. This new site was not using the Stack Exchange engine. The software in place there really sucked in comparison to SE. Slowly but surely, users of stackoverflow.com went away from the site, using Stack Overflow more and more instead. Quality on answers.semanticweb.com went down the floor and the site finally shut down.

When answers.semanticeb.com was in terminal stage (yet not clinically dead) in ~2015, someone proposed a new semantic web Stack Exchange site on Area 51. This one neither got the critical mass required to launch the site for real. It was deleted.

I find that a Q&A site on semweb tech is missing but I'm not optimistic that it will happen on Stack Exchange anytime soon.

Of course, you can use Stack Overflow for some technical questions on semweb tech, you can use Open Data Stack Exchange for some Linked Open Data questions, you can use Data Science Stack Exchange for some data analysis questions, you can use Artificial Intelligence SE for questions on KR & reasoning, etc.

But you cannot use Stack Overflow for all legitimate semantic web questions. You cannot use Open Data SE for all semantic web questions.

Consider for example the question "What are the success stories of the Semantic Web/Linked Data?" asked on 22nd August 2010 on Semantic Overflow. I'd say it is a legitimate semantic web question, but according to Stack Overflow's policy, it would probably get closed there. And it does not relate to open data, so it should be closed too on Open Data SE.

I extracted a few questions from the late answers.semanticweb.com, that most likely would not last on SO if they were asked there:

I would like to see a place where those questions can be asked and answered and stay.

  • 2
    Just FYI: the example question "What are the success stories of the Semantic Web/Linked Data?" would be considered opinion based, if not too broad and if Q/A communities govern their content correctly that question would be closed on any SE site. The three linked questions seem reasonable btw. So those might fly on web-data.se, if it goes to beta. – rene Dec 10 '18 at 17:05
  • 1
    yes, that is probably where it is best to redirect people to other communities such as Quora quora.com/topic/Semantic-Web for more open ended questions. One should compile a list of different places that work best for different questions and before closing point people to better venues. – Henry Story Dec 10 '18 at 17:10
  • I made the title of the three questions visible as those were hidden behind a link. They seem very reasonable to me. – Henry Story Dec 10 '18 at 17:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .