-
Notifications
You must be signed in to change notification settings - Fork 4
Final Thoughts, Conclusions and Outlook
The DASISH Final conference was held in Gothenburg on Nov 27-28, 2014 and now - as we are quickly approaching the new year 2015, the project is nearing its end.
On this page, some final reflections, conclusions and an outlook section will try to summarize what is most important as regards the development of the DWAN Wired-Marker-based client, but will also try to give some hints concerning the question where to go from here.
The original Wired-Marker code (development code name: ScrapParty) was developed in Japan as part of the Integrated Database Project sponsored by the Ministry of Education, Culture, Sports, Science and Technology for supporting the construction of databases. This turned out to be problematic in the course of our project, because we had to deal with both several local and a - new - remote database connected to the DASISH annotation REST service. Furthermore, the most important database functions needed to get tweaked in order to meet our needs. The code itself proved to be "heavy", not only were there comments in Japanese, which did not make life easier for us, but also was the number of source lines of code (SLOC) more than comprehensive (objectmng.js alone - a file with central functionality for the extension, e.g., counts 5420 lines of code).
The Wired-Marker add-on (http://www.wired-marker.org/en/) supports Firefox versions 2.0 - 10.*. However, we live in a fast-changing world and the current Firefox version is 34.0. This also caused more or less regularly reoccuring JavaScript errors for a lot of the new Mozilla Firefox version updates that were released while our implementation work was ongoing. The warnings and errors were mainly due to changes and updates of the Web API Interfaces (https://developer.mozilla.org/en-US/docs/Web/API), e.g. regarding the number of parameters for certain methods and the like, and made it necessary for us to come up with suitable bug fixes more often than we would have liked, because it slowed down the actual implementation progress for new functionality of the client. Last but not least, it is important to be aware of the fact that - naturally - the Wired-Marker add-on is an old-school, traditional, overlay-based Mozilla Firefox extension, with a jar-structured chrome.manifest file and other details. At this point in time, there are newer techniques which might prove to be a better choice if a group of developers would like to start developing another annotation client from scratch. These techniques are add-on SDK extensions and restartless extensions (https://developer.mozilla.org/en-US/Add-ons). Another lesson learned is that client portability is one important feature that the research community asks for and makes it possible to use the client with different "common" web browsers. Possible future approaches might be web annotation tools for the semantic web that can be installed, for instance, in the user's digital library or as a bookmarklet. One current example for this type of approach is the Pundit web annotation client (http://thepund.it).
The objective of DASISH task 5.6 was, however, not primarily, the construction of a client with all-embracing functionality, but rather the design and implementation of an annotation framework, i.e. the DASISH RESTful web service (https://corpus1.mpi.nl/ds/webannotator/). The client was needed mainly to support testing activities by interested pilot users and partners.
So, finally, and to conclude, where do we go from here? Working on the annotation framework task within the scope of DASISH has - as described above and described elsewhere in the respective deliverable (http://dasish.eu/deliverables/) - resulted both in final client and backend software products that can be used by the SSH communities for collaboratively (or on one's own) annotating web-based information, but has also led to valuable experience, that at least some of the partners might be able to make use of and build upon in connection with future project activities. The excellent science pillar of the H2020 e-Infra call is only one such example.