[MOBY-dev] data by reference - a request for comments

[Paul Gordon]

I'll finally chime in...I'm in favour of data-by-reference at any level 
within a mobyData block.  Above that and you get weird issues with 
exception maintenance, and any restriction below that appears pretty 
artificial to me.

Edward Kawas wrote:
> Hi Martin,
>
> Don't worry about Taverna! Whatever we decide will most likely be the same
> amount of work. So let's just get it right.
>
> Eddie
>
> -----Original Message-----
> From: moby-dev-bounces at lists.open-bio.org
> [mailto:moby-dev-bounces at lists.open-bio.org] On Behalf Of Martin Senger
> Sent: July-24-08 5:06 PM
> To: Core developer announcements
> Subject: Re: [MOBY-dev] data by reference - a request for comments
>
> Thank you for all your comments. We need to move on, however. I am, almost
> impatiently, waiting for a conclusion because I need to implement the
> "data-by-reference" rather soon. Are there other people preparing/building
> their comments - so we should wait for them, or can I summarize what was
> said so far - and create an official "RFC" from it?
>
> My personal opinion is still on the side to have references only for
> primitives. I am not convinced by the Peter's vision of a huge collection
> that is too big even when it uses references. I am simply trying to find a
> solution that will be good for the most cases - and then deal with the
> marginal cases separately (like by designing the service differently).
>
> I would like to hear also Eddie's voice - because he knows how easy or hard
> it would be to make the way we decide to do the "data-by-reference" in
> Taverna (and I am still talking about the T1 which I expect to be supported
> for some time).
>
> Martin
>
>