Wednesday, September 20, 2006

Developer patterns - a different type of pattern

A copy of an message sent to John Allsop in relation to his study on Web Patterns.


Probably a bit late for this years conference but I had another idea for your "source code web crawler". It would be interesting to see stats on another type of common pattern that we web designers seem to follow.

I would like to see what directory names we use for site structure. I.e. What's the consensus on using a folder named 'css', 'styles', 'c' (News Ltd use this) or something else (ie. regex on the directory parent of files ending in '.css')? A similar study could be done on the images ('img', 'images', 'media', 'i') and scripts ('js', 'scripts', 'j') folders.

You could take this even further to see what the most common files names are. I estimate that there are a heap of 'style.css' and 'logo.gif' but I haven't picked-up a trend on the "default" JavaScript filename.

Let me know what you think,

I'll let you know the outcome.

Tuesday, September 19, 2006

Comments on "Comments on community"

A while ago Jeremy Keith made the comment:

Comments should be disabled 90% of the time.

Although I agree with this statement I do not agree with the solution he is trialing to solve it. Especially in his case where he guesses the majority of the readers are geeks (I would agree being a "geek"). He is lucky to have a tech-savvy audience like this as geeks have blogs (and if they don't, they should if they want to keep their geek status).

To cut down on the whimsical comments that follow nearly all online personal opinion and to not burden the author with having to moderate each "I was here" virtual tag. I would use the power of TrackBacks. By forcing users to comment on their own blog they [should] be "forced" to put more of an effort into their reply as there is a greater onus on there own online reputation. Which is what Jeremy originally suggested. Also by using your own blog for response makes it much easier to collate and backup your own time and wisdom and, if your inclined, restrict it's use (password protect or copyright licence), it is almost impossible to do this when commenting on a system out of your control.

Unlike common comment replys being viewable to read at the base of an article, track backs are often liked off to another site which is an extra effort that the reader may not go to. Pulling in a "sample" of a reply using some AJAX, like Adactio is currently doing is a great way to keep the users reading flow. You could then take this idea a bit further to increase usability by using some logic to display the TrackBacks in a user selectable order (other than chronological). Such sorting could be done on a Google like weighting (ie. check, using using some Technorati APIs, to see which pinged posts have the most subsequent pinged posts. Or use other API's (like or Digg) to rank the responses based on popularity. You could also use the Author's own blogroll to highlight peer replies as the author already regards their ideas and feedback above others.

Using such logic I think trackbacks could make a "comeback" and raise the collective commenting intelligence.