[DAS] [Fwd: Re: Writeback implementation]

[Andy Jenkinson]

Forgot to send this to the list...

-------- Original Message --------
Subject: Re: [DAS] Writeback implementation
Date: Wed, 29 Oct 2008 09:56:53 +0000
From: Andy Jenkinson <andy.jenkinson at ebi.ac.uk>
Organisation: European Bioinformatics Institute
To: Gustavo Salazar <gustavo at nbn.ac.za>
References: <49058C89.7050301 at nbn.ac.za>

Hi Gustavo,

The decentralised 'annotations of annotations' approach is a direction
that is likely to see most adoption in my opinion because it doesn't
require the original data provider to modify their source.

Were you planning on using the existing "features" command in order to
retrieve the annotations, or something else? I ask because it's feasible
to imagine a DAS source that does not support writeback but still
annotates another source's annotations. In fact the DASMI architecture
already does this with it's scoring servers.

Cheers,
Andy

Gustavo Salazar wrote:
> Hello everybody,
> 
> This is my first post in this list, therefore I'm going to start to 
> introduce myself. I'm Gustavo Salazar, I'm currently busy doing my  MSc 
> degree in computer science in the University of Cape Town - South Africa.
> The project that I'm working on is about the implementation of the 
> writeback capabilities in the DAS client Dasty.
> My original Idea was to use as a server the writeback implementation 
> created by Asia and Andreas. However i've been notice that this 
> implementation works as an extra server and Dazzle is kind of middleman 
> between the clients and the writeback (am I wrong?) which sound like a 
> good idea in terms of independence, but it looks to me that it will be 
> hard for a client to identify if a  feature is original or has been edited.
> That's why I decided to explore others alternatives and now I started to 
> work reimplementing the server DAS writeback capabilities not in Dazzle 
> but in MyDas.
> I thing the writeback server should works as a meta-annotation server, 
> which means that none of the modifications, additions or deletions will 
> be actually changing the original server. in such a way a DAS client 
> should see the information of the writeback as an extra layer, therefore 
> it should first query regular DAS servers, built in memory the graphic, 
> and at the end it will query the writeback server to modify this graphic 
> with the community information.
> In this way the user can choose to use the wb information or not.
> I will use the protocol as in 
> http://biodas.org/documents/das2/das2_writeback.html with the 
> modifications that appears in the Asia's Theses. which implies the use 
> of OpenId as the authorization system, I agree with the pros and and 
> cons of OpenId that Andy posted, therefore if the consensus is to use 
> another authorization system I will adapt my implementation.
> I will appreciate any comment or suggestions or if anybody wants more 
> details of my ideas please no hesitate in ask me.
> 
> Regards,
> 
> -- 
> Gustavo Salazar
> _______________________________________________
> DAS mailing list
> DAS at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/das