Published on Thursday, July 20 2006
It seems that the River of News model for aggregators is getting new attention. Humanized Reader got great reviews from Steve Rubel and 37signals and they deserve all of it: their solution is particularly well designed (even if i feel it being too much blog like…) and the dynamic scroll is awesome.
This renforces my own experience: I’ve never been able to use an aggregator for more than a few hours before i tried the NewsRiver aggregator built into the OPML editor which has been my feed reader ever sicne i tried it for the first time (thanks Dave!):
The interaction with those modeled as a mail reader seemed unnatural to me, I shared much of the feelings that the Humanized crew has:
They had lots of modes. They made me stop and think about navigation every time I looked at the title of a post and every time I finished reading one.
An aggregator is a reading instrument for a very special form of literature which is different from email or books and as such should be built and designed for a different reading experience.
I’ve found that the river of news design is a natural way to read the flow of information coming in the form of feeds. Its interface stays out of the way when reading the news while it gives the correct support for organizing them. Showing the freshest on top and pushing down the (long) page the older it lets me use my prior experience of to find older items.
I’m so convinced of this that a couple of weeks ago i started to build an online aggregator following the river of news design. It’s called Tiber (like the big italian river I live nearby) Now I’m really looking forward to release it and see how it compares with the others.
Published on Friday, August 26 2005
I just found that in Ruby on Rails there is a “./script/console” script which lets you access all the application from a consol prompt, in this way you can manipulate all of your objects and do all sort of things and i find it really useful specially when you need to debug some nasty bugs in your model.
I had known this before I would have not spent most of the afternoon going after a bug in the Rails application I’m developing. After a long quest around my code, I extracted what seemed the offending piece of program from the model, built a test case and run it in the irb, and there it was a stupid initialization error.
Using console I would have just skipped all the plumbing and beoing able to interact with the single parts of the application I would have concentrated the attention on the model.
Published on Thursday, August 25 2005
While studying Ruby on Rails and looking for documentation I found this post by David Heinemeier Hansson
Addressing concerns with Active Record (Loud Thinking).
It explains some of the rationale behind the Active Record machinery and why it has been made so that you can do the easy thigs VERY easily while it is trivial to revert to sql for the most complex things. Qouting from the post:
Active Record shouldn’t give the expectation that you’ll never have to touch SQL again. That’s an illusion and many ORMs have been build and crumbled on that illusion.
This is something that I’ve come to understand while working on large java based projects using Oracle’s bc4j as the object relational mapping layer, when there is the need to present complex data the best option is to use the database tools which are built expressely to manage the data, and are optimized for these tasks: SQL, views and stored procedures. In such cases we would often revert to write custom queries into ViewObjects or let all the processing be done by some stored procedure (which was easly done also with bc4j!).