Home - acdh-oeaw/wugsy GitHub Wiki

Welcome to the wugsy wiki!

Anyone can put anything on this page. Right now, before the tool is actually in use, I think the wiki is useful to collect random thoughts, feelings, feedback, etc.

Some thoughts on generalisability

I think an important brainstorming point is the generalisability of the system and its code.

From recent meeting, we talked a lot of about processing of images, tags and captions, with a heavy NLP component. Realistically, much of the backend processing should be doable through wugsy methods.

I also think, in general, people are not aware of the extent to which good code is modular, and perhaps we need to find ways to make them aware. For example, wugsy is a collection of more or less independent apps, linked only by federated login and a homepage. In our cases, the reasons for having a single site with many apps greatly outweighs having a site for each app.

The other side of modularity: wugsy is not just a web app, but should also be an importable library that processes data, something like:

from wugsy import corpus
c = corpus('my_file.txt')
# and so on.

Dicta

My other idea is for contributors to come up with some wise words to live and code by. Right now, mine is:

Always code with multimodality and multilinguality in mind. When prototyping, using English is fine, but it always needs to be treated as a default selection, rather than hard coded in.