arteymix / datapedia Goto Github PK

This will keep the internal architecture clean and top-notch.

It is a requirement for any data that replaces the current data to be non-regressive with its predecessor. However, if by the time a data gets approved since it was non-regressive with current data of a certain time, it gets replaced by a another approving data that is regressive to the first data, we will have trouble.

We have to validate once more when the data is being approved.

Regressive approving data should eventually be cleaned up.

Write tests for all kind of accesses.

POST means you want to update the data. An existing data must exist.

PUT means you want to create a new data. No data should exist at the endpoint.

I have my reserves with the XML data format as it is not data-oriented but document-oriented. This definitely break the bijection with other data formats like JSON and will introduce problems.

There could be an implementation of plist which is a xml-encoded dictionary. http://en.wikipedia.org/wiki/Property_list

A data counter that counts the number of submissions, archives and current data and display these statistics in the footer.

Maybe having a page for usage statistics.

This is a core feature in Datapedia: the ability to support multiple data formats.

My thought is that we could only implement data formats that are sufficiently JSON-like.

There must always be a bijection between supported formats that passes through Python object. If we encode some data from a given language, we must be able to decode it in any of the supported languages and the reciprocate should always be possible.

I've programmed Datapedia in a way that data encoders and decoders are interchangeable. There is still some features to implement like the merging of multiple data (in different formats) and other stuff.

I know that if we implement another language, it will be an easy task to introduce other languages. Right now, I'm thinking of BSON and MessagePack.