Class identifier

From DocDataFlow
Jump to: navigation, search

Crawler is built using object-oriented programming techniques.

Many of the Crawler concepts map directly onto underlying object classes.

For example, every adapter type, every granule type, every granule ordering... is programmatically represented by an object class.

For granules and granule orderings, these underlying object class are assigned a unique alphanumerical identifier, called the 'class identifier'.

These class identifiers are 'reverse domain names': by taking an existing domain name (rorohiko.com), reversing it, and add a number of period-separated strings to it, we can create identifiers that can be nearly certainly unique. Other companies using Crawler can define class identifiers that are guaranteed to be different from any other company's class identifiers as long as they stick to this 'reverse domain name' rule, and take some care in the choice of the identifiers.

There are many granule types; some examples:

  • com.rorohiko.granule.indesign.color
  • com.rorohiko.granule.folder
  • com.rorohiko.granule.text.word
  • ...

Another example. These are two predefined granule ordering class identifiers:

  • com.rorohiko.ordering.frame.vertical
  • com.rorohiko.ordering.frame.horizontal

In Crawler's config files, it is often necessary to refer to specific object classes. That's where these class identifiers come in: they are a convenient way to refer to the use of a particular granule type or granule ordering in a text based INI file.

There is also a short-hand notation.

For granules, the prefix 'com.rorohiko.granule.' can be dropped, so in INI file entries where a granule type is expected, the entries

com.rorohiko.granule.text.word
text.word

are equivalent.

For granule sorts, the prefix 'com.rorohiko.ordering.' can be dropped, so in INI file entries where a granule ordering is expected, the entries

com.rorohiko.ordering.frame.vertical
frame.vertical

are equivalent.