If you have worked with RSS feeds, you are probably aware that there are several versions of RSS formats and there is also the Atom format (see http://en.wikipedia.org/wiki/RSS_(file_format) ). While RSS 2.0 seems to be the most popular one, all of these formats are being actively used by some websites somewhere. There are some similarities in element naming among the RSS revisions (if you ignore the namespaces), but Atom and RSS are totally incompatible. These incompatibilities make it very difficult to support multiple formats, especially for programs that need to parse and display them. The diagram below illustrates this issue:
Even thought these formats are incompatible, it's obvious that all of them contain the same set of information. They include:
Since they are the same information that are just in different format, clearly we can just write in one unified model and convert them to the other formats. Yarfraw's core data classes offers a unified model that is big enough to cover all of these different formats. With that, you only need to work with one data model and at the same time have the benefits of supporting multiple formats. The diagram below illustrates the idea:
Yarfraw's core model is capable of holding information from all supported format. When the API reads a feed into the core model, all the data that are common in all formats will be mapped to a common field, the format specific elements and attributes are then mapped to the corresponding format specific fields.
For example, all formats have a <channel> (or <feed> in Atom format) element that contains meta data about the feed. This element will be mapped to a data class called 'ChannelFeed' in Yarfraw. As you would expect, <item> in RSS and <entry> in Atom are both mapped to a data class called 'ItemEntry'.
Sub-elements, such as <title> element of a <channel> or <feed>, are also mapped. In both 'ChannelFeed' and 'ItemEntry', there is a field call 'title' that holds this information. Some of the common elements, while contain the same information, are named differently in different formats. For example, there is <category> elements in both RSS 2.0 and Atom 1.0 format, but in RSS 1.0, this element is called <dc:subject>. Since they are the same data, they are mapped to a field under 'ChannelFeed' called 'CategorySubject'. The following table lists all the special mapping among formats:
Note the following table is not necessary complete, for more details, I encourage to take a look at the Javadoc . The model also supports extensibility elements, see Builder Example and other documentations on the site.
| Yarfraw Data Model | RSS 2.0 | RSS 1.0 | Atom 1.0 | Comments |
|---|---|---|---|---|
| ChannelFeed | <channe> | <channel> | <feed> | Note that in RSS 1.0 second level elements such as <item> elements under <rdf:RDF> are also mapped here |
| CategorySubject | <category> | <dc:subject> | <category> | <dc:subject> is an element in RSS 1.0's dc extension module |
| Cloud | <cloud> | ignored | ignored | This is only used by Rss 2.0. |
| Content | <encoded> | <encoded> | <content> | This is not officially supported in RSS 2.0 and RSS 1.0, but if there's a <content:encoded> element under <Item>, the content of the encoded element will be mapped to this class. The type will always be 'text' in this case. |
| Enclosure | <enclosure> | ignored | ignored | This is only used by Rss 2.0. |
| Generator | <generator> | ignored | <generator> | RSS 1.0 does not support this element |
| Id | <guid> | ignored | <id> | RSS 1.0 does not support this element |
| Image | <image> | <image> | <logo> and <icon> | |
| ItemEntry | <item> | <item> | <entry> | |
| Link | <link> | <link> | <link> | |
| Person | Use as a type for 'person' type info, such as <webMaster> <author> etc | Use as a type for 'person' type info, such as <dc:publisher> <dc:creator> etc | 'Person Construct' | Note that RSS 1.0 and 2.0, only 'EmailOrText' field is used, as the text content of the corresponding element |
| Source | <link> | ignored | ignored | This is only used by Rss 2.0. Note that this element is not the same as the <source> in Atom. |
| Text | string elements such as <title> | string elements such <title> | 'Text Construct' | Note that RSS 1.0 and 2.0, only 'Text' field is used, as the text content of the corresponding element |
| TextInput | <textInput> | <textinput> | ignored | <TextInput> element of Rss 1.0 and Rss 2.0. This is ignored by Atom 1.0 |
| Yarfraw Data Model | RSS 2.0 | RSS 1.0 | Atom 1.0 | Comments |
| ChannelFeed.title ItemEntry.title | <title> | <title> | <title> | a rare element that every format happens to use the same name |
| ChannelFeed.managingEditorOrAuthorOrPublisher | <managingEditor> | <dc:publisher> | <author> | This is a list, but for RSS 1.0 and RSS 2.0, only the first 'Person' in the list is used. |
| ChannelFeed.links ItemEntry.links | <link> | <link> | <link> | This is a list, but for RSS 1.0 and RSS 2.0, only the first 'Link' in the list is interpreted. |
| ChannelFeed.descriptionOrSubtitle | <description> | <description> | <subtitle> | |
| ChannelFeed.docs | <docs> | ignored | ignored | Rss 2.0 only |
| ChannelFeed.lang | <language> | ignored | 'lang' attribute | |
| ChannelFeed.webMasterOrCreator | <webMaster> | <dc:creator> | ignored | |
| ChannelFeed.pubDate | <pubDate> | <dc:date> | ignored | |
| ChannelFeed.uid | ignored | ignored | <id> | Javadoc |
| ChannelFeed.ttl | <guid> | <sy:updatePeriod> and <sy:updateFrequency> | ignored | Rss 1.0 - This value is parsed from both the <sy:updatePeriod> and <sy:updateFrequency> for example: updatePeriod:hourly and updateFrequency:2 = ttl: 30 minutes |
| ChannelFeed.cloud | <cloud> | ignored | ignored | RSS 2.0 only |
| ChannelFeed.rights | <copyright> | <dc:rights> | <rights> | |
| ChannelFeed.imageOrIcon | <image> | <image> | <icon> | |
| ChannelFeed.logo | ignored | ignored | <icon> | Atom 1.0 only |
| ChannelFeed.textInput | <textInput> | <textinput> | ignored | |
| ChannelFeed.skipHours | <skipHours> | ignored | ignored | RSS 2.0 only |
| ChannelFeed.skipDays | <skipDays> | ignored | ignored | RSS 2.0 only |
| Yarfraw Data Model | RSS 2.0 | RSS 1.0 | Atom 1.0 | Comments |
| ItemEntry.descriptionOrSummary | <description> | <description> | <summary> | Javadoc |
| ItemEntry.authorOrCreator | <author> | <dc:creator> | <author> | This is a list, but for RSS 1.0 and RSS 2.0, only the first 'Link' in the list is interpreted. |
| ChannelFeed.contributors and ItemEntry.contributors | ignored | <dc:contributor> | <contributor> | |
| ItemEntry.comments | <comments> | ignored | ignored | RSS 2.0 only |
| ItemEntry.enclosure | <enclosure> | ignored | ignored | RSS 2.0 only. To add an enclosure to Atom, use the link element. See Atom 1.0 specs |
| ItemEntry.uid | <guid> | ignored | <id> | Javadoc |
| ItemEntry.pubDate | <pubDate> | <dc:date> | <published> | |
| ItemEntry.updatedDate | ignored | ignored | <updated> | |
| ItemEntry.rights | ignored | <dc:rights> | <rights> | |
| ItemEntry.content | <encoded> | <encoded> | <content> | Javadoc |
Since release 0.9, Atom 0.3 are also supported. Most of the elements are mapped to the core model the same way as Atom 1.0, with a couple of exceptions. Some elements in Atom 0.3 are not mapped to the core models but will be available through an extension module call Atom03Extension (see extension support ).