Trackback

"Getting the Blogging story right in Plone" is a good post, getting to lots of the problems that we see and have seen in blog products in Plone. I just want to respond to a few of the specific points as Tom Lazar and I have been working on many of them for the upcoming Quills 1.6 and "2.0" (that version number is not set in stone yet, I'd say) releases.

So what’s wrong with blogging in Plone right now? Mainly I think it’s commenting, at least this is the reason why I switched my COM.lounge TV blog to Wordpress. With Quills I had several problems back then. Some might be due to the version I used (it simply seemed to be bugs) but the most important was trackback spam actually.

Yup. Commenting sucks in Plone. This is something that Plone needs to fix, not Quills. Happily, people do seem to be working on this area. See the new release of easycommenting (which I haven't had a chance to try out yet).

Trackback is something that the Quills project should handle. For the 1.6 release, we have recognised that our implementation is not nearly as good as it should be, and so ship without it. That's better than shipping with something unusable (due to spam) as was the case in Quills 1.5, but obviously it's not optimal to ship without this important functionality. My view is that trackback should share lots of code with commenting, so careful work on the latter has the potential to push forward the former as well.

enable every content or folder to be turned into a blog. This was an idea I had a while back and it seemed that Tom had it, too :-) It probably is simply about adding some marker interface to a folder to enable various blogging features.

This is what is slated to be be Quills 2.0. There is code that does this in the collective under the QuillsEnabled project. It worked a couple of months ago, but is currently broken (I expect) as the code it shares with Quills 1.6 was refactored somewhat, without these changes being reflected in QuillsEnabled... yet. I believe Tom plans to sprint on this after the Plone Conference!

If a folder is marked as a blog it might enable a blogging API for using a blog editor

See QuillsEnabled :).

If a folder is marked as a blog it might have technorati tags as a viewlet available

This may be something that PloneBookmarklets does or could address.

enable blogging APIs (the more we can provide the better) to also edit non-blog content, like normal pages or news items (what’s the difference anyway?)

See QuillsEnabled :). We currently have a MetaWeblogAPI implementation that will work with standard plone content as soon as QuillsEnabled is updated. I know Tom is keen on AtomPub. Ian Bicking appears to be working in this area as well, albeit not explicitly for Plone.

Enable simple statistics on your blog. It would be nice to directly be able to see a rough click count per entry a la Wordpress, the same would be nice for your RSS subscribers.

I agree this would be nice, but has to be something the Plone and/or its documentation handles, I think.

Enable a "more" section on posts which doesn’t show up in the overview listing or the RSS feed.

Quills basically has this already. The weblog view can be told to only display entry summaries (i.e. the standard AT description field), instead of the full text. This is not wired up to the syndication feeds yet, I believe. But that should be easy enough for an only slightly enterprising hacker :).

Enable a feed per category (should be easily doable with collections but making it even easier for end-users or even automatic might be nice)

Quills has had this since 1.5. Append /@@atom.xml to any of the 'topics' URLs, and you get an atom feed for that topic. Likewise for the other feed types we offer. e.g. http://quills.sitefusion.co.uk/blog/topics/Core/@@atom.xml.

So, Quills already offers lots. And where it doesn't quite meet expectations, we have code in the works to address them. Things are getting better :).

My first step today was to fix the plonetrackback bug for deleting received trackback pings (issue 90).

After deleting a trackback ping, it gets uncataloged.  However, the way that the pings are stored in the catalog is rather convoluted and hackish.  As a result, the standard unindexObject call didn't work as expected, and instead also uncataloged the actual WeblogEntry instance to which the ping was attached.  I resolved this with a change to uncatalog_object(trackback.getPath()).