greghochsprung.com

As you may (or may not) know, I am a really big fan of Dan Benjamin and his 5by5 podcasts. I was listening to this week’s Build and Analyze episode (with Marco Arment) and they were talking about Brent Simmons’ Plea for Baked Weblogs and Marco’s ideas of how to implement this. That really got me thinking. How cool would it be to write my own simple blogging engine?! And as I was thinking about it, it seemed like it could be done pretty easily, so I think I am going to try to do it myself (even if it is just for my own education and never gets used). This post will be a basic outline of all I think it would take to do it.

Since I know Python best, I’ll use that for most of the programming and found a Python version of Markdown. Other than that, I’ll use Dropbox to store all the posts so I can write and edit posts from my iPhone or iPad via the PlainText app (my current favorite text editor).

My basic idea is to write two shell (or Python?) scripts, called rebuild and rebuild_all. Rebuild will be a cron job that runs every minute. If there are any new text files ready to be posted, it will change them from Markdown to HTML, rebuild the home page, and rebuild the archive page (only the files that need to be rebuilt to accommodate a new post). Rebuild_all will be a cron job that runs once a day to go through and rebuild all HTML files on the entire website (which will allow it to catch any edits I make). Since all Markdown files will be stored in my Dropbox, I can edit any of them at any time from any device I want. I think this would be a very simple, but elegant solution.

Here is a listing of the basic tasks each script will do:

Script 1 - rebuild (every minute)

• Check /drafts folder in Dropbox for any file beginning with “zz” (my way of denoting a file ready to be published)
• Put that file in /posts/year/month/day? folder in Dropbox and remove “zz” from the name
• Add “/posts/year/month/day?/title” to a text file which will be the home page log file
• Run Python script 1 that uses Markdown and a template HTML file to build the static HTML file for that post and put it in the www folder
• Run Python script 2 that uses Markdown and a template HTML file to build the home page from the home page log file
• Run Python script 3 that updates the archive HTML file

Script 2 - rebuild_all (every day)

• Traverse all of /posts folder in Dropbox, and for each Markdown (txt) file run Python script 1
• Run Python script 2

That’s it. It seems like it should be simple enough - now I just need to find the time to try to do it!

I recently downloaded a new app to my iPhone called into_now. It is basically a social app that allows you to check in to what you are watching on TV. Like all good social apps, it allows you to see if any of your Facebook or Twitter friends had joined so you could “friend” them to get started. I am so accustomed to the “follow” version of “friend” (like Twitter) that it didn’t even occur to me that I would actually be sending friend requests to everyone I tried to “friend” (like Facebook). Few of my friends are as geeky as me, so none had joined yet, but several people I follow on Twitter had. So, since I figured into_now would have a more Twitter-like (public timeline) model, I tried to friend several people who I have never met without even thinking about it. When none of those people showed up in my timeline or list of friends right away, I realized that I must have sent them an actual friend request! Oh well, hopefully none of them mind.

Through this experience, I realized that I have become very accustomed to Twitter-like “follow” model of social networks. Very few new services have the Facebook-like “friend” model. One that comes to mind is Path, but that is for a very specific purpose. I don’t really understand why into_now chose this model, but that is their decision. I am writing this post to ask developers who choose to implement this model to include some sort of warning or confirmation before actually sending a friend request. When I tap somebody’s name to friend them, a dialog asking “Are you sure you want to send a friend request to this person?” that gives me the option to confirm or cancel would be very helpful. I know that LinkedIn requires an extra step before requesting a connection and am pretty sure Facebook does. That would make people like me feel a lot less embarrassed when we try to friend everyone we follow on Twitter, but have never met.