The ParlRulesData Project aims to make available the formal rules of procedure for various parliaments over time. We provide these rules in their original language (or one of them).
The datasets for each parliament allow users to do two main things. First, users can view all of a parliament’s formal rules in place during a given period. Second, they can view how any specific rule has evolved over time. This is possible because the data is organised by consolidated versions and root numbers.
For each parliament, the data is organised into a series of consolidated versions. These represent the complete text of a parliament’s formal rules after each date on which they were changed. We focus only on permanent changes, excluding rules which were adopted on a temporary or 'sessional' basis. We also create consolidated versions for dates when the rules were re-numbered, even if the content of the rules did not change.
These dates were identified by consulting lists of reform dates in published versions of the rules, by comparing successive published versions of the rules, and by consulting records of parliamentary proceedings. (NB: these dates are those when rule changes were passed, which is not always the same as when they came into effect).
These consolidated versions allow users to view a parliament’s rules as they stood in any specific period covered by our data.
We further divide each consolidated version into sub-articles – the smallest unit of separately labelled or numbered text. By comparing successive consolidated versions, we have assigned sub-articles to 'root numbers' that allow individual rules to be traced across time.
This was done as follows, for each parliament. First, we took the earliest consolidated version in the data. Each sub-article in that version was assigned a unique root number. Second, a human coder compared this with the next consolidated version. Sub-articles which were still present in the second version in an identical or amended form retained the same root number. Sub-articles in the second version which couldn’t be matched to any in the previous version received a new root number. This process was repeated for each consolidated version, comparing it to its predecessor.
This system of root numbers makes it possible to trace when each sub-article of a parliament’s rules was first created, any time when it changed, and – if relevant – when it was removed.
On this website, we report how many sub-articles were added, deleted, and removed in each new consolidated version of a parliament’s rules. These have been calculated by comparing each consolidated version to its predecessor at the level of sub-articles. Additions and deletions are defined by the structure of our datasets, but users of the data can calculate changes in a number of ways.
Additions are where a root number appears for the first time in a new consolidated version. This represents a sub-article that can’t be matched to a sub-article in an earlier version of the rules.
Deletions are where a root number appeared in the previous consolidated version, but does not appear in the new one. This represents a sub-article that has been removed from the rules.
Changes are where a root number appears in both the previous and new consolidated version, but with different text. For the purposes of this website, this includes any changes to the text, identified using exact string matching. (This approach thus includes very minor editorial changes, such as changes in punctuation.) However, other methods are possible, and we would encourage users of the data to calculate alternative indicators of whether sub-articles have changed. One such alternative approach is presented here.
To distinguish between different iterations of our datasets, we employ a versioning system analogous to semantic versioning. Our version numbers have the following format: MAJOR.MINOR.PATCH. Our first release of a dataset will therefore be marked v1.0.0.
For parliament-specific information on citations, data coverage, and sources, see the README documents provided when downloading the data.