Daniel's Blog

Okay, I’ll try some textile and see if it actually comes out on your blog ;-)

I setup MoinMoin+Subversion for my group at work and a wiki/blog seems to be the most common misconception/requested feature. I promote wiki’s because they are dead simple to use, but in order to use a wiki well requires some learning and practice. And wiki’s are certainly not suited for encoding every type of information possible (blogs are a good example). However, I would go as far as saying that (with practice) a wiki is like an external memory source that can be immediately used by my brain. Everytime I learn something, I choose an appropriate page name and then I write what I learned in a page of that name.

I’ve re-written some groups of pages in my wiki up to 3 or 4 times and discovered a few things. One: it is important that pages be short. If a page is too long, then it should be split up. Two: page names need to be appropriate. If I want to know about a particular subject, then I expect to find a page named after than subject. This can include the violation of the StupidPageName convention. For example, I recently created a page titled “Architecture” which contains hyperlinks to pages about architecture. Sometimes, especially for weird or ambiguous page names, I prefix them with some sort of qualifier such as “Sound/sfront” or “Sound/MP4-SA”. MoinMoin handles this prefixing similarly to a hierarchial file system. The StupidPageName convention is a good fall back, but sometimes it it is best violated. Third: using a consistent style is important to for rapid assimilation of information in pages. At the very top of my pages I usually have a “see also” section to link to other wiki pages. Then I have a table of contents. Then I have a section (called “Links”) for hyperlinks to other websites. Then I add other sections as appropriate. Fourth: use redirects. If I find that to get to a subject and find that I usually type in the wrong keyword to get there, then I simply make a page that contains that keyword and I put a hyperlink inside of it to get me where I want to go.

A set of pages prefixed with “PackagingSoftware/” have been fairly useful to me. These contain notes about languages (C, Python), build systems (SCons), and package systems (ebuild). For example, if I forget how to encode run-time library search paths into a C program, then I can just look at “PackagingSoftware/C” and expect to find relevant information. And if that information doesn’t exist, then I do a google search and import my discoveries into the wiki. If I want to remember how the heck to achieve a DESTDIR effect with SCons, then I have a well defined place where I can keep track of such information.

But, I agree with you on the wiki/blog thing. Making a modern wiki do blog basically destroys the major benefit of wikis (in my opinion): simplicity/ease of use. I’ve been trying to come up with some non-sucky (generalized) scheme to interlink wikis, blogs, a calendar, and a papers database. But, I haven’t discovered the tao yet.

Okay, I’m being unusually philosophical now. Here’s a thought: wikis cover space whereas blogs cover time. Wiki’s are great for covering information that doesn’t have to be cast into a timeline, but blogs (or better: a mailing list) can let you reconstruct a series of events as they lead to a conclusion. Here’s something that neither a wiki nor blog would do: what if you want to encode a segment of history, say the Napoleanic wars and other events that occurred at the same time. A wiki doesn’t provide enough dimensions (namely, time) to effectively encode this information. Blogs and mailing lists are concerned with journaling events as they happen. So what would need to be done to encode the Napoleanic wars in a manner that could be represented in a hyperlinked medium with timelines/calendars, maps, and other good stuff? Or how about the evolution of U.S. Code? For instance, we might want to track the state of U.S.C. as it has ever existed, but we also might want to relate that to events that occurred in Congress and around the world. And then, we might also want to relate that with more examples from history (e.g. to check if we’re committing actions that have occurred repetitively in history). History is good because it is subject to interpretation, and thus can benefit from a wiki-ish / emailing list-ish sort of model where people’s various ideas are recorded.

Baird,

I’m going to email you about these topics separately. I seem to have three bugs in my blog:

1. I’m not getting emails about comments
2. Textile didn’t work for you, but it did work for me (in this message).
3. Previewing doesn’t show you what’s actually going to happen. I’m not sure what I can do about that.

Hmm. What to do?