Talk:PeopleFinder Interchange Format
From Katrina Help Info
| Table of contents |
|
1 Considerations for evolution of specs: |
Contrasting Design Goals
- The PFIF, People Finder Interchange Format, standard has different design goals from other data efforts. [1] (http://zesty.ca/pfif/1.1/#principles) The full PHIF spec (http://zesty.ca/pfif/1.1)
- The Katrina Data Project spec (http://www.katrinadataproject.com/data.aspx) is a single record with detailed information meant for detailed categorization.
Data duplication
PHIF allows repeats. Repeats are tolerated and better than purging records that shouldn't be nuked.
Data sources are critical when trying track down loved ones.
A separate record for notes aids in attribution. The importance of notes and sources was learned first hand after 9/11.
PFIF has multiple records for notes while KDP has a single note record.
Web or Desktop
The PFIF spec is designed for application use and hence the early adopting of XML Schema to allow for machine readable applications such as CRMs, RSS, and other internet based applications.
KDP is based on Comma Separated Values as the interchange format. This is a common standard for data interchange but does not support the web service, internet application designs that most of the application developers we are working with needed. For example, not John Plax's need for data to be ordered in the XML Schema for Salesforce support is not possible with a comma separated values as the interchange format.
Data sources and data creation
There is not a strong notion of data sources, only data creation in the KDP spec.
Privacy
The PFIF can learn from KDP in adding privacy support.
Considerations for evolution of specs:
Allow thoughtful reflection of what we can still improve.
More to document:
Q: Who is "the keeper" or the "prime developer" of the PFIF?
Q: Where did the PFIF start? Where, when. A little trivia and details are okay at the bottom of the page.
XML::PFIF perl module for generating PFIF files and outputting PFIF XML and RSS
- download from Ian (http://nessie.mcc.ac.uk/~ianb/projects/pfif/)
More documentation will be coming very soon, for now just have a look at testpfif and testrss, which output XML very similar to pfif-1.1-example.xml and rss-pfif-1.1-example.xml at http://zesty.ca/pfif/
I plan to make another release in a day or so with proper documentation and various small tweaks
to install:
gunzip -dc libxml-pfif-perl-0.1.tar.gz |tar xvf -
perl Makefile.PL
make
su -c "make install"
I will try very hard not to break the API.
Feedback and bug reports are very welcome to Ian.
