[DAS2] biopackages.net out of synch with spec?

Andrew Dalke dalke at dalkescientific.com
Wed Mar 15 14:32:26 UTC 2006

> I just ran through the source request on biopackages.net and it is  
> returning
> something that is very different from the current spec (CVS updated as  
> of
> this morning UK time).

The server isn't synched with any specific version of the spec. For
example, if I make a features request from


I get

<?xml version="1.0" standalone="no"?>

As from the discussion a few weeks ago we shouldn't be using the
since that says the document cannot be understood without consulting
the DTD, which doesn't exist.  And I don't want a DTD.

Also, the namespace needs to be  
(It's missing the 'genome') and the 'FEATURELIST' was replaced with
'FEATURES' a year ago.

In the types request

<?xml version="1.0" standalone="no"?>
<?xml-stylesheet type="text/xsl" href="/xsl/das-genome-type.xsl"?>
<!DOCTYPE DAS2TYPES SYSTEM "http://www.biodas.org/dtd/das2types.dtd">
   <TYPE id="SO:ARS" ontology="/das/ontology/obo/1/ontology/SO/0000436"  
name="ARS" definition="A sequence that can autonomously replicate, as a  
plasmid, when transformed into a bacterial host.">

the commented out namespace declaration needs to there, and the type
id 'SO:ARS' needs to be escaped as it's treated as an identifier  
with the "SO" protocol.  Plus, until yesterday I didn't know about the
'name' or 'definition' attributes.  These are now in the schema as
'title' and 'description'.

There are a few other differences, like problems in the taxid and
empty strings for timestamps.  I hand-updated examples from Allen's
server yesterday, in cvs under das/das2/draft3/ucla .  I found some
of these during the update, though others I pointed out about a
year ago.

Allen doesn't want to update the server until the spec is stable,
for two reasons.  First, he doesn't like the churn of doing work only
to have to make more changes.  Second, you're not the only one who says

>  It is much more fun for me to code to a working
> server because I have the opportunity to watch my code run.

and Allen's setup doesn't have the ability to implement two versions
at the same time.

					dalke at dalkescientific.com

More information about the DAS2 mailing list