Thursday, May 24, 2007

A few thoughts about Ruby on Rails

1. Starting is fun and easy. A few configurations, sqls and commands and "almost" everything you need is done.
2. First road block comes when you create views which query multiple tables in one go.
3. Life becomes easy again when you have to modify designs and templates. Page templates resides in different folder so html headache is minimum
4. You really have to bang your head in the wall if you try to integrate your app in a portal with an already-present-cookie-based login system!
5. Never ever try to do #4 with a thrid party app.
6. Rails somehow seems very immature with the way DHH changes stuff. Apps created in rails 1.1 doesn't work with rails 1.2. Even ruby versions and "optional" gems create problem and to add to the misery these errors are either cryptic or so insanely-damn-fucking stupid that you won't ever realize that it's a version clash. (A "wrong number of argument "... error surfaced because I had a older RedCloth gem and it was a discovery almost by accident when I out of sheer frustration uninstalled everything and installed everything again)
7. Rails has weird gotcha and it stems from the fact that it is cryptic. An innocent looking column name in a table would give you so much pain that you feel like banging your head against your monitor. Don't do that becuase first it can be dangerous and second it's not your fault. Check out these links
8. Performance of a "default" app sucks big time. A lot of tweaking is necessary to make it fast enough to be barely usable.
More on rails bitching later

Thursday, May 03, 2007

Business Idea #1

Okie here goes my first one:
1. General Knowledge on the move
Wikipedia is a free source of knowledge but this knowledge base is restricted to web. What if there be a tool which summarizes those articles (possibly divding it into multiple pieces and shorten it) and makes it available over mobile phones. For instance you want to look up "Lake saimaa" (It is the largest lake in finland and is very beautiful), it picks up the relevant article from wiki and sends back the user a sumamry.

In this case since the article size itself is small (character count 1025) so summarizing it (making it 320 character long) won't be a big issue as such. Some of the popular "sms" shortcut transforms can be used. But what if somebody searches for say "India". Then the summarizing code has to be more intelligent to take the article and ascertain which are the important parts of this particular article (in this case geography, political system, economics, population, military, culture etc). Then it should return say about 200 character long article on India and list it's population, GDP, Average expectancy of life etc and say "make more relevant search by saying ex. India GDP" etc.
This way people are intrigued to send more smses to search server for more content which would in turn result in more revenue.
Now what if somebody sends a query about some recent happening or some obscure topic which wikipedia has no knowledge about. Then we land in a soup. Wikipedia alone cannot cater to all the knowledge "needs". We will need more content, news, articles and such.

Okay then so much for mobile content.