reBlogger 4 – goals

Thanks to our customer and beta testers feedback, we’ve been able to identify the crucial goals for reBlogger 4. This post is perhaps coming a bit late, since the development of reBlogger 4 is under it’s way for quite some time, nevertheless I will present you with the features of our upcoming version.

We could divide our development efforts in following areas :

1. simplify customization of reBlogger
2. improve search/filter engine
3. improve and ease certaing administrative tasks
4. add social context to reBlogger

Although, these goals are long run goals, even in reBlogger 4 each of these aspects we’ll be improved.

In order to simplify customization of reBlogger following has been done :

  • reblogger has got a toolbar (visible only to logged in administrator, similar to wordpress)
  • the look and feel of reBlogger can be changed through themes panel. Themes can be downloaded from our official reblogger.com web site, or developed by our users
  • the layout of reBlogger can be changed through the new layout panel, which comes prepackaged with 1,2 and 3 column templates with optional header and footer
  • The all new custom content panel allows inserting any kind of html in any given part of reBlogger, therefore making it easy as copy&paste to set your own header, footer, third party gadgets (counters, flickr…) or ads.
  • reBlogger has been divided into controls (logical units of functionality), each exposing a vast set of properties (e.g. you can decide do you want to show the date by the post or not). All these properties can be set through your web browser without going into any code

the search/filter engine has been improved in following manner :

  • posts are now categorized in topics (previously known as keywords)
  • each post can be categorized by positive keyword (previously known as search term) and negative keyword (meaning that post must not contain that keyword in order to be placed in the particular topic)
  • the adding, deleting, editing and sorting of topics and keywords has been ajaxified making it much faster and more user friendly than before

improve and ease certain administrative tasks

  • added option to add feeds through OPML files
  • added option to add list of feeds
  • added auto suggest lookup of feeds
  • feeds are being added in the background, posts are being fetched in background to allow for smoother experience
  • improved and expanded statistical data on the state of reBlogger (size of db, number of posts, trends etc.)

add social context to the reBlogger

In the coming version of reBlogger, we are building digg-style voting for posts/stories. Voting will be done through a smooth atlas interface.

Going “stealth”

I would like to apologize to all the readers that were reading my 10 day development journal of reBlogger 4 engine. We have decided to move this product to "stalth" mode for the time being. We are doing some very exciting stuff at the moment and don't want to spoil it before it's done. I'll keep writing on this topic and then probably we'll publish it at same point in the future. Thank you for understanding.

Ivan

Paging with new ROW_NUMBER() function of MS SQL Server 2005

Ok, after tackling temp tables, nested select statements and what not for paging, finally there is something really nice in Sql Server for paging.

Thanks to new ROW_NUMBER() function paging is piece of cake. Here is a MSDN link with a nice example. 

Search vs. Explore

As I have been thinking of reblogger and it’s uniqueness it appeared to me that we need to make a clear cut between the terms search and explore. Reblogger is an exploring tool, everyone else is doing the searching. So, to start this out, here are the definitions from wikipedia.

Exploration is the attempt to develop an initial, rough understanding of some phenomenon.

So, to be able to search we need to know “who” or “what” are we looking for, regardless of does it exist or not. With exploring, we are going to find what that “who” or “what” may be.

Let’s examing following sites and what are they offering :

To do this, we’ll use a simple case scenario, we need to write a essay on “inflation”. The crucial point here is that, we are not looking for a particular information (we are not SEARCHING), but we are trying to get “rough understanding” as wikipedia puts it. If this “rough understanding” is up-to-date at the same time, then we are on our way!

Google

Obviously, I type “inflation” at google, and among 5-6 inflation calculators, I’ve got a wikipedia link, some statistics, few Federal Reserves links and a link to nice essay about “inflation for beginners”. Now I’ve just caught my breath thinking this will be helpful. Well, it almost was, until I found a paragraph describing how atrology can be used to guess the inflation rate?!? So what’s missing here? Relevancy! Google is looking for keywords, which is fine if I was looking for let’s say inflation rate in Botswana (http://www.google.com/search?hl=en&lr=&safe=off&q=inflation+rate+in+Botswana&btnG=Search), but the truth is I had no idea what was I looking really for. I was exploring.

Technorati

Let’s try with technorati. Typing the “inflation” in basic search return bunch of posts on “inflation”, which was cool. Then I’ve choosed to look only in blogs about “economy”, since I’m interested in scientific perspective of inflation (for the moment being!!!). So, that worked out well… with few more refinments (remove posts with Alan Greenspan… etc).

So, the problems I had with Technorati are following :

1. I’d had to go through way too many things to find the posts I’m looking for.
2. I’d like much more specified tagging structure… if inflation is a keyword (rb terms), I’d like to see following search terms (rb terms) : monetary theory, neo-keynesian theory, phillips curve, hyperinflation, price controls, deflation… etc. That would be a nice starting point.
3. I really didn’t want to read all posts… my ideal picture here would be : ok, I have posts on “inflation”, I see the search terms and look them up at wikipedia, then I dig deeper looking for “phillips curve” for example.
4. No history! Once when I find a blog that is informative as well as relevant I want to keep up with it. In Europe, at the end of college every student has to write a final paper which needs to be usually 200-300 pages long. It takes about a year to write that, so I want to learn about inflation on a more permanent basis. I also want to keep up with some other things like sport, movies etc. and I DEFINETLY have no desire to add 300 bookmarks to firefox and be bombed with irrelvant information.
5. it would be really nice to see what blogs is reading and how is filtering posts economics professor from sweeden that wrote 3 books on inflation!!! At technorati, I see the aggregation of masses, and I presume most of them are not really the inflation experts.

del.icio.us

With del.icio.us I had the very same problems as with the technorati with the addition that I could not have look for inflation among the economics blogs… so I’ve got a lot of posts with “inflated” fame, self-assesment, buildings…. etc. Then I figured out that del.icio.us has no posts and that it’s actually a bookmark thing (now you all say duuuh!), so that pretty much ruled it out as a exploration tool, because I knew way too little about any of the posts to decide on their relevancy…

digg

It’s a technology orientated site so no tests could be done, but in general there is way too many people deciding on one thing, so it is useless unless one is trying to figure out “what’s hot” at the moment. Some good concepts though.

wikipedia

The good in bad

Ok, we had some delays with the patching up reBlogger 3.0.5 and releasing 3.0.6, so I could not get right into the coding stuff, and what a blessing was that!!!

My initial idea was to go and redo the database structure to this multiplayer mode, which would basically result in a two separate directions of development. Needless to say we would have reoccuring bugs, lot’s of recoding and in general mess. It has just occured to me that the fastest and by far the best way to build reBlogger.com would be to write an extension for reBlogger 3.0.6.

Basically what we would have is 2-3 additional tables that would represent some kind of joins between users and keywords/negative keywords. Since we don’t want to add same feed more than once and also we don’t want to add same post more than once, it is only logical to leave all that data in the same database.

Anyhow, reading through Mark’s blog I’ve got a rather solid understanding of what we are actually building. Then going through this article I realized how hot reBlogger is, and how far ahead we are with our software.

What I think is that we need to make an API (and that we’ll need to develop reblogger.com so it’ll be a good testing point) and start writing extensions, mainly focused on the front-end.