| ★ wanayoo — archive 1999 https://github.com/ipython/ipython/wiki/Roadmap:-IPython | Nouvelle recherche | Portail wanayoo |
This document describes the goals, vision and direction of the IPython project. IPython is a large project with many different areas of functionality. With development ramping up it is critically important that we focus our efforts in a disciplined manner. We simply can't do everything at once.
To provide direction and focus to the project, the core IPython developers have started to meet every six months, mostly in person, at UC Berkeley. These meetings are being funded through the Sloan Foundation. Along with these meetings, we plan to have a major release every six months (roughly July / December). The meetings will happen around each release, to provide focus and direction for the next six months of work. Our first meeting was in January 2013, to plan for the 1.0 release in July and this document is a direct result of that meeting. Each meeting will result in this document being updated with a roadmap for each release.
Below you will find our upcoming releases and a list of the tasks we will focus our development efforts on for that release. These tasks categorized by difficulty / amount of work (Easy, Medium, Hard), and priority (0-3, with 0 being most important). Where relevant, the developers leading the work are also noted. This is not to say they will be the only contributors, only that they will be responsible for moving the task forward. If you are interested in working on one of these tasks, please coordinate with these individuals.
For the most part, we would like developers to focus on the top priority tasks, before moving on to the lower priority ones. This is not to say that things will never be done out of order, but our priorities have been chosen carefully, taking into account dependencies between different tasks. If developers contribute new features outside the focus of the current release, it is likely that we will ignore their contributions until a later release. Work on bug fixes will always be given high priority.
Our work for 2013-2014 is being driven by our grant from the Sloan Foundation. More details of this grant can be found here. For 2013, we are going to focus on really solidifying the single-user notebook experience, which is principally in two major areas:
Following that, in 2014, we will begin the development of the multi-user notebook. This multi-user notebook will be similar to the single-user notebook, but the server will map directly onto Linux users on a single environment, so that you only need to run a single server for all users of your system (be it a local shared machine, EC2 image, etc.). This multi-user notebook server will be focus on the usage case of small-medium sized groups of trusted users. It will not have features that allow it to scale to large numbers of users on the open internet.
Development focus for IPython 1.0:
Development focus for IPython 2.0:
pyin, pyout)interact and interactive
.txt to .rst (Paul)pyout and pyin (Min)user_variables from the execute_request message? Make sure user_expressions returns a dict. We currently have no policy for error handling of these requests. (Min)complete_request message could likely be simplified (ex, does it need the text key?) (Min)--script
.ipynb to something else. We like the name .inb.ipy-cell-input. This should be done to create a "public" API for styling notebooks (Jon)ipython notebook mynotebook.ipynb that links to existing server if available, otherwise starts a new one.While we haven't really begun this work, there are some things we are thinking about that will affect users of the Notebook, in particular people who are already developing JavaScript widgets for the Notebook.
Development focus for IPython 3.0:
This marks the end of the Sloan Foundation funding, by which point we should have a solid multi-user IPython Notebook server with interactive widgets, and powerful utilities for interacting with other formats, such as generating Sphinx / HTML or LaTeX / PDF documents from IPython Notebooks, and running code in at least a few languages (Python, Ruby, R, etc.).
Last edited by Zachary Sailer,