Normally, mailing list discussions contain a lot of useful information and thoughts, which is then lost in the archives. This is why, we prefer blog posts for proposals and contributions. As a rule of thumb consider this: For quick, short questions about several minor topics the mailing list might be better. For the proposal of new ontologies, use cases and technologies blog posts are better, as they will stay visible for a longer time period.
There are several ways, how you can leave your footprint on this site and contribute to the project. Standard contributions are blog posts to the following topics/categories: Request for comments (RFC), Implementations, NLP Ontologies, Tutorial Challenges, Tutorial Solutions . You are expected to use the first paragraph (two or three sentences) to give a short introduction about your connection to the topic of your blog post (your motivation, use case or designated use). Of course, you can also decently link to your work/project/company/coming events pages.
To submit something just send a short abstract/draft either to the mailing list nlp2rdf (at) lists.informatik.uni-leipzig.de or to Sebastian (hellmann (at) informatik.uni-leipzig.de) and you will then get an account to post on this site. We will also list your name and your contribution on our Involved People page.
Additionally, all data entered in the tables below the blog entries is made available as Linked Data via the Semantic Wiki WordPress Plugin, so you can crawl it and export it again. There will be a SPARQL endpoint soon.
The Contribution Formats
Request for Comments
NIF-1.0 is a stable version, but it is far from being complete. It is currently focused on Part of Speech tags and OLiA, Entity Linking and the SCMS vocabulary and other simple annotations. There is still a lot of room to include your ideas and ontologies for NIF-2.0. The basic process is to write a blog post about a Use Case that is not covered currently by NIF-1.0, or propose extensions and other ontologies that could be incorporated, different URI designs, etc, etc… These ideas can then be discussed and refined and finally included into NIF-2.0 .
Implementations
Blog posts about the implementation of NIF-1.0 should contain a short description of the tool and what annotations it uses. Furthermore, the following would be nice: where to download it, licence, status and pointers to the source code, if available. If possible, we will also try to integrate your application into the LOD2 Stack.
NLP Ontologies
Blog posts about ontologies that can be exploited for NLP. These are not per se Linguistic Ontologies as they are normally modelled for Linguistics. The posted ontologies should have real practical value and it should be possible to solve an NLP task with them.
Tutorial Challenges
If you have an interesting and/or challenging NLP task, you can write a blog post here so others can try and solve your task. It is difficult to exactly define what challenging means, but here is a try:
- Challenging for traditional NLP frameworks in the sense that either the integration of components would be too time-consuming or the task involves reasoning or ontologies. Constantly breaking pipelines or missing representation of tagsets/linguistic knowledge.
- Challenging for RDF/OWL in terms of performance and also tasks that can potentially show the limits of NIF and the representations used.
- Challenging for newcomers to NLP (the solutions are supposed to finally become tutorials.)
- Challenging in some other way…
We except the task to adhere to these restrictions:
- The time to achieve the task stays within a certain limit (maybe a week or two maximum).
- It should be related to NIF or NLP2RDF in some way. Minimally, it would need a Semantic technology to solve the problem.
Have a look on the Tutorials page to see current challenges.
Tutorial Solutions
A tutorial solution is an answer to a tutorial challenge. You can answer to tutorial challenge, if:
- You are the first one and there has not been a solution before.
- You know a better solution (less code, better performance, easier to implement).
- Your solution reproduces another solution in a different language.
- Your solution has extra features and/or is a significant extension of another solution.
- Your solution should ideally use NIF or at least RDF and Ontologies in some way.
A tutorial solution blog post should describe the used process, so that it can be reproduced easily by others. Although the solution can build on some prerequisites, it should also describe the basic steps or point to a web site that explains this part. Code snippets for copy & paste are definitely welcome and the source code should be available. The code and the application (e.g. web services can be hosted by nlp2rdf.org and LOD2).
