Planet vs Universe

Due to a once fairly even split of opinions, we actually have two versions of Planet: Planet Gentoo and Gentoo Universe. The universe aggregates entire blogs (articles of any topic) into a single page, whereas the planet site only aggregates “on-topic” RSS feeds, where “on-topic” is loosely defined to be anything related to Gentoo/open-source/tech-stuff.

The Planet Gentoo on-topic thing doesn’t work too well. To pull it off, we require developers to provide us with a separate RSS feed for the Planet. Usually, people just create a new category called “Gentoo” and give us the category-specific feed URL. However, many people use their “Gentoo” category for Gentoo-only articles, then they publish other interesting open-source related articles outside of the Gentoo category, so they do not appear on Planet.

Also, under normal circumstances, there isn’t a lot of difference between Planet and Universe (that is perhaps slightly different today due to the site being spammed with old content).

So, I’m thinking of killing Planet and replacing it with what we have on Gentoo Universe – complete aggregations of everyones weblogs, where the only restriction is that content must be ‘sensible’ and written in English. How do people feel about this? I’m especially interested in the opinions from those who originally said they didn’t want to be reading what people ate for breakfast every day (the people who wanted an “on-topic-only” Planet to read).

This change will be discussed on the gentoo-dev mailing list before I take action, if I decide this is a good idea.

Looking for a co-maintainer

Maintaining the Planet and Universe sites really is quite straightforward, yet people will tell you how slow I am to add or remove weblogs. Most of the time, I’m too busy with other things, so planet requests end up sitting in my mailbox until a rainy afternoon.

For this reason, I’m looking for a co-maintainer who can be more on-the-ball than myself. I say co-maintainer, actually you’ll probably be doing most of the work, but I’ll definately still be hanging around and contributing the odd few minutes now and then, as well as continuing the more advanced maintenance tasks.

To be more specific, what the position really involves is:

• Adding new weblog feed URL’s to 2 config files (one for planet, one for universe) when developers see the light and set up their own weblog (or when new developers join the project). These files are kept in a subversion tree, just commit them and Planet picks up the changes within 1 hour.
• Removing those feed URL’s when developers leave.
• Uploading developer hackergotchi’s to the designated location on request, and referencing them in the config files.
• Creating developer weblogs on our b2evolution installation – this is provided for those who don’t want to host their own weblogs.
• Removing comment spam from our b2evolution install (we are working on better anti-spam measures).

That’s really not a lot – but it’s not particularly interesting either.

All you really need is experience with weblogs (even if that’s just installing and running one), basic knowledge of subversion, and a notion that you’ll be able to be more responsive than myself. PHP/Python skills are a bonus if you are interested in getting even more involved, but that’s not a requirement. This is open for both existing Gentoo developers and non-developers too. Anyone interested?

Republishing of old content

Occasionally, the Planet republishes all articles in someones weblog as if they were written on the current day.

This has been happening a lot lately, partly due to people mass upgrading to WordPress 2.0 (for some lame reason that ends up slightly changing the dates of all your previous articles), and partly due to some buggy weblogs which randomly change the article dates when they feel like it. We’re tracking the recent occurances in bug 128895.

The sites are powered by a small python application cunningly named Planet. One thing I miss from the old Planet version was that when this happened, I could simply open Planet’s cache files, modify the dates of the old articles to be too old to be aggregated (i.e. changed the publish date to be 1 year ago), then re-run Planet in offline mode. This was an acceptable workaround.

The newer versions of Planet do not use a human-readable cache format anymore, so I can’t do the date hacking thing. Is anyone interested in writing a python app which parses the cache files and allows me to modify dates? Planet is dead simple to set up and hack on, no webserver required – its just a script which outputs to a .html file in a certain directory. The source code can be found here.