Here’s what each article looks like in my reader. Here a list of the sections as shown on the left of the image.
Title & Link – Article title and the link clicks through a tracking page that logs the followed links.
Actions – More info on this below.
Action Display – This area is normally hidden unless an action is clicked that needs additional input.
Text – Body of the post.
Metadata – Additional data in this order now: Weight, Diggs and Techmeme flag.
Favicon – This is the favicon pulled from the blog.
Favorite – Marks it for long time storage. Used for posts like how-to’s or as a reference.
Mark Read – Marks the article read and longs time for usage reporting.
Review – Flags the article for later. Used when I don’t have time to fully read so I can get back to it. Main difference between this and favorite is that these are just flagged until I get a chance to really read the article.
Email – Right now hooked up to gmail so I can send an article to a friend.
del.icio.us – This brings up the del.icio.us interface with link and title prepopulated.
Blog it – This displays a small textbox for me to enter in 120 characters of text which posts to twitter, tumlbr and jaiku. It’s 120 characters because I reserve 20 for the tinyurl automatically inserted into post. Examples: http://twitter.com/smilbandit
Info – I haven’t figured out what I want to do with this yet. I know I want to be able to link to a source of additional data for this post but I’m not sure what that is yet.
Tags – Pulls up the tags associated with this article.
Long time since my last update. Since then I’ve added Techmeme into my information sources. I was temped to include the firehose feed but thought better of it. Instead I use the information from Techmeme and flag any article that is on Techmeme. This flag is used to pump up the weight of an article and enhance the weight of the tags associated to that article.
This makes it so that I’m not contributing to my RSS fatigue but helps notify of that something that normally will read has some importance.
This post mimics my thoughts on why I wanted to include Techmeme but not something like the firehose feed. Why I Read Techmeme via A VC
I’ve added digg stats to the unread articles. The thought is to crowd source the articles that are weighted low on my personal schema but dug alot. Eventually I don’t want to feel that I need to read every article. By using digg stats I’ll know that the more talked about posts will be brought to my attention and should help with discovering content that is outside what I might normally read.
Right now I have the query engine for digg stats down to about one hour or so for 200 articles, most come back with no diggs. It will run every three hours right now but probably with my reading habits it will only need to run twice aday since it’s function is ancillary to the main function.
I’ll probably add additional sources like reddit, del.icio.us or maybe techmeme if I figure how to plug into it.
First time I read about fav.or.it. Felt good that I was going down a similar track as someone else but also a little bad because my idea wasn’t the first to hit the thoughts of everyone. I saw a few screenshots and noticed that they weren’t open for registrations yet so I decided that I’d try to not look at it anymore and develop my path.
From what I saw it looked pretty good, much better design then mine. For me a design is something I have to come to through use. I don’t seem to have that ability to take a blank screen and know how things should be laid out.
Having comments as part of the reader design is definitely different and probably not something I’m real interested in doing. I don’t comment on many posts although I did when .mobi was getting setup, just didn’t think it was a good idea plus I liked m. better. I also wonder how blogs themselves will like this if it becomes popular since their only source of ad space will be in the feed. I hope it doesn’t mean that my feeds will be stuffed with pop-up like ads.
Added a search feature. I don’t think I’ll use it much but I figured I may need it one day and I was at a lull.
It’s pretty basic you can search the posts title or body. What makes it a little bit cool is that I can search for a specific tag and there is a check box to only return posts marked as favorites.
Eventually I’ll add an advanced search so I search by feed or date range among other standards search options.
After I added weights to the feed I modified the feed list. I first removed feeds with less then 5 posts or weight > 1 as I can pick these up in the latest list and keep them from cluttering my feed list. I tried ordering the feeds by weight but that made finding a specific one difficult. Now they are ordered by name but I added colors based on the weight so I could scan easily for specific ones.
I now wanted to make certain tags bubble up even though their weren’t many posts on it. I started by giving a weight to the tags and then tallying that rather then the number of posts associated. I fine tuned the sizes of the tags in the display based on the maximum weight of the unread posts, rather then hard coded ranges. I didn’t want to make the weights limited 1 to 10. I wanted to see where the right limit was organically, right now the largest weight I applied for a tag is 40 so that [newsflash] from fark. I find the [newsflash] from fark a better gauge of importance for me then anything brittney related in CNN’s top stories feed. I also gave lego 20 since not many posts are made about lego and patent got10.
Here’s the logic as the code works to figure out the tags size. (d = maxweight / 4)
Tag Weight
Style
Font-Size
= 1
smallest
.4em
> 1 and < d
small
.4em
>=d and < (d*2)
medium
1em
>=(d*2) and < maxweight
large
1.5em
=maxweight
largest
2em, and all caps
This created a number of really small tags that weren’t very important and was making the tag cloud to large for my tastes. So I had the code stop writing out any tag that where the stye was small or smallest. I’ll post a video soon so you can see that this made the tag cloud create a dynamic effect that I didn’t really plan but has worked out great. As you read the posts associated with the highest weighted tag and remove it from consideration, the maxweight is redefined and the cloud is redisplayed. This worked out great until about weight < 10 where there were just to many tags being set to medium and above. I now have hardcoded lows, maxweight !< 10 and d !< 5. At this point I know that most of these post aren’t really important to me but I can pick and choose those that interest me and possibly weight them higher for later.
After awhile I missed knowing if a tag had a lot of posts or not, so I implemented a second level indicator based on color of the tag. This lets me pick and choose the smaller tags that have a lot of posts and when I’m presented with a large tag cloud.
Posts
Style
Color
< 3
low
faint blue
>= 3 and < 7
medium
faint green
>= 7 and < 10
big
blue
>=10 and < 20
high
purple
> 20
huge
red
This all happen in about two weeks of development usually at about 1-2 hour intervals that I had time to work from home.
Feed Weights
When I clicked on a tag to show the posts I have them listed in descending chronological order, which is my preferred reading style. I noticed that that was making the me too posts bubble up to the top of the list. I tend to like to read posts from certain feeds over others so I implemented weights to feeds also. I included the feed weight into the tag cloud calculation as a multiplicative of the tag weight. These I’m trying to keep between 1 and 5. I also now have the posts ordered by weight and then descending chronological order. I’ve been toying with the idea of reversing the chronological order but just haven’t got to it yet. I’ve got Techcrunch weighted at 5 and if they make a post with a patent category the weight in the tag cloud will be 50, which will make the patent tag bubble up to the top.
I’ve had other ideas like placing the certain feeds favicon under the tag showing that a specific blog posted about it but I don’t think it will have much worth and might overly clutter the cloud.
I started adding actions so I could do something other then read the posts.
Favorite: marks the post as a favorite so when I write an script to clean up the database I can keep posts that are interesting. Generally I see using this with technical blogs I’m subscribed to that have code snippets or something.
Mark Read: I already have a "mark all read" button but this will immediately mark a post read and hide it from view.
Email: what you’d expect but I used a gmailthis bookmarklet I found and have it linked to gmail to send.
del.icio.us: same as email but will share the link out to del.icio.us.
Blog About: this is a little icon I made myself that will let me click and create a blog post about the article.
