GLEP 30: Planet Gentoo web log aggregator

Author	Daniel Drake <[email protected]>
Type	Standards Track
Status	Final
Version	1
Created	2004-10-24
Last modified	2019-11-07
Posting history	2004-10-25, 2004-11-10, 2005-03-11
GLEP source	glep-0030.rst

Contents

Status
Credits
Abstract
Motivation
Specification
Rationale
Backwards Compatibility
Reference Implementation
References
Copyright

Status

The new Planet Gentoo came online 2005-03-10, so this GLEP is now Final.

Credits

Stuart Herbert for the original idea
Alexander Plank who also has put effort into promoting this

This GLEP proposes the creation of "Planet Gentoo", a new section of the gentoo.org website. It would aggregate weblogs (commonly known as "blogs") written by contributing Gentoo developers onto this single page, available to the public. I will refer to this new website section as "the planet" in this GLEP. We would also provide an installation of a weblog engine for developers that do not currently have their own weblog.

Motivation

I'm trying to reduce the gap between the Gentoo user and development communities.

Many large open source projects and software distributors have their own Planet where contributors weblogs are aggregated. See the Planetplanet homepage for a complete listing. These other aggregations appear to be successful and bring their relevant communities together.

These aggregations are often quite interesting to read, since they contain a variation of topics, not all of which are related to the project.

The planet would add another method of user interaction with developers, as most weblogs allow readers to post comments.

Readers would get more interested in the Gentoo project, and would get a feel for the personalities of the contributing developers. Although I am not suggesting this should be used as an announcement tool, developers could also use this to get general messages over to the user community.

This would also benefit our development, as developers would also be able to keep track of what other developers are working on and promote more collaboration amongst each other.

Specification

The proposed design is an installation of Planetplanet at a part of the gentoo.org infrastructure, e.g. http://planet.gentoo.org

We would also provide an installation of Wordpress or a similar weblogging engine, for developers that do not have their own weblog or would wish to move their log to an official Gentoo hosted website. This could be provided at (e.g.) http://weblogs.gentoo.org

Rationale

Planetplanet is simply a weblog aggregator written in python. It is executed as a cronjob and fetches content from all the weblogs it has been asked to, postprocesses and aggregates them into a single html file (based on a template), and outputs that html content to an area provided by a webserver. This should allow for ease of integration with any existing infrastructure.

Planetplanet is configurable through a single configuration file, which lists the log feed URL, real name and user name for each contributing developer.

Wordpress is a weblogging engine written in PHP. It relies on MySQL for the data store. After the initial setup, all configuration is done through a web-based interface.

A group of people would be assigned the responsibility/CVS access to maintain these two services. I would suggest the existing infrastructure team to have this responsibility. If required, I will assist with the initial configuration.

The planetplanet installation would fetch and aggregate developers weblogs only, we would not accept non-developer logs.

Developers do not have to host their weblogs with us, practically all blogging packages provide an XML feed (typically RSS) which planetplanet will happily fetch and process. Other feed formats are also accepted, examine the planetplanet documentation for more info. In the event of a contributing Gentoo developer leaving the project, their log would be removed from the aggregation. If their log was hosted by ourselves, it would be closed. For this reason, some developers may choose to host their log elsewhere - this is not a problem and is left to the decision of the individual developers.

It is true that the addition of the planet to our website collection would add yet another source of Gentoo information to our collection; our users and developers are already overwhelmed with IRC, mailing lists, forums, and the central website. In contrast, a lot of information is currently replicated over those mediums, but the planet would provide a taste of something new. Developers can choose their own topics and are free to write about things that perhaps would not fit into our existing communication mediums. Additionally, the planet will attempt to bring some existing content together by linking to recent commits, bug lists, and dev.gentoo.org webspace for each contributing developer.

It has been suggested that the activity of the planet may be an issue: although a lot of activity would make the planet very successful, it may look bad on us as a whole if the planet is inactive.

Alexander Plank started a gentoo-dev discussion back in August regarding this exact idea. Alexander set up a Planet Gentoo mini-survey to see which developers have weblogs or would be interested in obtaining one to be aggregated on the planet. At the time of writing, the survey yielded a list of 21 interested developers.

Looking at the other planets available, a contributor count like this would be more than enough to keep the planet active and interesting. Through my experiments of finding developers existing weblogs through google and aggregating them on my local Planetplanet installation, I have come to the conclusion that just a small quantity of active weblogs is enough to keep the planet going strongly.

I also note that the list of 21 interested developers was built up over a short space of time, and it was only mentioned once in a rather large thread on the gentoo-dev mailing list. I predict that given more publicity amongst the developers, this idea would be even more popular. I also predict that if such a planet were to go live, other developers would be motivated to join in, and new developers joining the project would be keen on contributing.

The success of the planet will be evaluated by examining the number of hits to the planet. 3 months after launch, one weeks worth of logs will be recorded, and hits will be counted. If the hit count for that week is below 1000, the planet will be deemed as not having met its target, and appropriate action can be taken if the planet appears to be harming our image.