Present: Art Anger, and Bob Therriault 1) Bob and Art talked about using redirects to reorganize the wiki and whether there is an issue with having two pages that have similar information. If one was in an Archived folder the other could be in a live folder that might be updated. This would address the problem of unarchiving a page that had been archived by mistake. This would mean that the redirect would follow the live page and the Archived page would only be seen by a user by searching the wiki
2) Bob talked about approaches that other array languages have used for documentation, as it was a discussion on the most recent episode of ArrayCast. https://www.arraycast.com/episodes/episode38-array-language-profile BQN which has very good documentation done by Marshall Lochbaum was the result of hundreds of hours working on it. Marshall also talked about the importance of including examples that show the function of a primitive being used in unexpected ways. He also spoke about the importance of including simple examples first for those users who are looking for a refresher and then including more complicated examples to show the variety of applications. Richard Park from Dyalog talked about doing documentation of very simple things that a developer would need to know to get an application working. This avoids the problem of having a language which is used for code golfing or leetcode but rarely used for practical purposes, which is a problem that some people have suggested that J has. Conor Hoekstra spoke in the podcast about the functionality of Rust Up in Rust which will construct the infrastructure of an application and this may be a path to take for Add-ons and making them easier for people to contribute. 3) Bob and Art discussed the email response of Tom McGuire about how he had used Jupyter notebooks. Tom's User page https://code.jsoftware.com/wiki/User:Tom_McGuire/JJupyterNotebooks Tom's email: Just like to mention that I was working with the Jupyter notebook interface pioneered by Martin Saurer and I adapted my own Binder launch script based on Joe Bogner’s jkernel-docker for Binder. Both of the aforementioned have github sites that have the code: Martin: https://github.com/martin-saurer/jkernel <https://github.com/martin-saurer/jkernel> Joe: https://github.com/joebo/jkernel-docker <https://github.com/joebo/jkernel-docker> Jupyter notebook runs in a browser and needs a server to control the session. Using Martin’s original kernel I would run the server that comes with the Anaconda python distribution locally on my machine. With the Binder website you can now launch Jupyter notebook files directly out of a repository on GitHub. This allows anyone to share the file. For those who are unfamiliar with Jupyter it is similar to running J or python under an emacs editing session. All user entry is in text editing boxes that get executed either by J, python, or the underlying markup language. It is kind of an executable wiki. Even though code may be entered in separate text editing boxes they are all controlled by one J process so global nouns and function definitions are usable throughout the Jupyter page you are editing. Jupyter notebooks are extensively used in the Python community especially by the BigData/Data Analysis folks 4) Wrapped up with the plan to develop the yellow wiki as a prototype for people to explore by the end of October and then have the changes implemented through the month of November on the blue wiki. After that it will be more maintenance and developing a team so that there is a division of labour in this large project. For access to previous meeting reports https://code.jsoftware.com/wiki/Wiki_Development If you would like to participate in the development of the J wiki please contact us on the general forum and we will get you an invitation to the next J wiki meeting held on Thursdays at 23:00 (UTC) Next meeting is October 20th, 2022. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
