[DAS2] Links to old DAS/2 grant and final progress report

[Gregg Helt]

On Wed, Oct 29, 2008 at 1:41 PM, Brian Gilman <
gilmanb at pantherinformatics.com> wrote:

> Hey Gregg,
>
>        I was wondering why DAS/DAS2 couldn't be represented using different
> wire protocols? For instance, why not das in json and das in xml etc. etc?


The DAS2.0 spec does allow server responses in alternative content formats.
Every capability in the "sources" doc can specify alternative content
formats via the <FORMAT> element, and also in the "types" doc alternative
formats for annotations can be listed on a per-type basis via <FORMAT>
elements.  Retrieving annotations in these alternative formats is done by
adding a "format=XYZ" to the retrieval URI's query parameters.  We
considered using standard HTTP content negotiation but felt that it would be
better to have the format embedded in the retrieval URIs rather than in the
HTTP headers, and that this approach also allowed more granularity without
additional request overhead.

This alternative content ability is used heavily by the Integrated Genome
Browser (IGB) DAS2.0 client and the Genometry DAS2.0 server to
request/deliver very efficient binary formats when possible.  If you take a
look at IGB's console output when loading annotations from the Affymetrix
Genometry server, you'll see that for many annotation types IGB sends
feature requests for custom binary formats like "bgn", "bps", or "brs".  In
more recent work I'm starting to use Google protocol buffers for alternative
content formats, and fiddling a little with json too.

One of the nice things about the Trellis DAS2 server framework I've been
developing is that the framework handles conversion of data model to data
format, at least when it understands the requested format.  So I hope to
generalize the conversion and add support for various formats into Trellis,
so they don't have to be reimplemented for every backend data source.

Regarding the discussion of SOAP and REST -- I agree that trying to compare
the two directly confuses the issue.  SOAP is a specific protocol whereas
REST is a set of principles for software design.  I try nowadays to compare
specific "RESTful" protocols to SOAP when needed.  However the grant
proposal was written for people who don't generally read web tech
specifications, back then whatever else SOAP was it definitely wasn't
RESTful, and there was no other handy term for an alternative to SOAP, so I
glossed over the category differences when comparing SOAP and REST.

      Gregg

Also, SOAP is a wire protocol while REST may not be... I think this confuses
> the issue?
>
> -B
> --
> Brian Gilman
> President Panther Informatics Inc.
> E-Mail: gilmanb at pantherinformatics.com
>        gilmanb at jforge.net
> AIM: gilmanb1
>
> <a href="www.scilink.com/people/gilmanb"><img src="
> http://www.scilink.com/images/buttons/scilink_viewmy_black.png"/></a>
>
>
>
> On Oct 29, 2008, at 5:14 PM, Gregg Helt wrote:
>
>  I've been reminded recently that when I say stuff like "well in the DAS/2
>> grant we proposed XYZ" few people can actually look at the old grant and
>> see
>> what I'm talking about.  So I've posted a copy in a more permanent
>> location:
>> http://biodas.s3.amazonaws.com/das2grant/DAS2+Grant+Proposal+Feb2003.doc
>>  .
>> I've also posted a copy of the final grant progress report:
>>
>> http://biodas.s3.amazonaws.com/das2grant/DAS2+Grant+Final+Progress+Report+Aug2008.doc
>> .
>> If you do take a look at the grant, keep in mind it was written over
>> five
>> years ago.  Back then for example REST was still a relatively new concept,
>> and SOAP hype was peaking, so there's a fair amount of text dedicated to
>> explaining the RESTful approach we wanted to take and why we weren't just
>> using SOAP.  Since then some of the specifics have definitely changed, but
>> I
>> think the general concepts hold up pretty well.  For instance the current
>> thread about meta-annotation had me looking back at the section on
>> "Feature
>> References" and finding it's still relevant.
>>
>> I've also added links on the BioDAS wiki DAS/2 pages to the grant and
>> progress report.
>>
>>     Gregg
>> _______________________________________________
>> DAS2 mailing list
>> DAS2 at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/das2
>>
>>
>