[DAS2] Why use URIs for feature IDs?

[Andrew Dalke]

> I'm going to incur the possibility of pitchforks here.. :)

To mollify or intensify the pitchforks ...

Several people have said that "the id is the last component
of the URL" or "the URL is the base + '/' + the id".

That's what DAS1 did.  I don't like URL construction like
this. It makes the URL organization imposed by the specification
when it doesn't need to do so.  For example,

Allen prefers his URLs like this
    /feature?this=that    is the query interface
    /feature/F00001       is an identifier for the features

I might like it like this
    /feature_search.cgi?..   is the query interface
    /feature/F00001       is an identifier for the features

Still others as
    /features?this=that   is the query interface
    /feature/exon/A1      is an identifier for the features
    /feature/contig/A     is another identifier for the features

** NOTE: in this case the "last term of the URL" is not
sufficient as a unique short-form id  **

Or still others as
    /cgi-bin/fsearch.rb?this=that   is the query interface
    /data/F1                   is an identifier for the features
    /data/F2                   is another identifier

One advantage to hard-coding the URL organization into the
spec is the tradition from DAS1, and the general practice of
expecting one-off URL schemes during web scraping.

Another is that people understand it more easily.  It's
a lot easier to write out examples in one naming scheme than
it is to say "using the identifier from the record ..."

On the other hand, the programming is easier.

					Andrew
					dalke at dalkescientific.com