[Re-Proposal] Re: [Proposal] Goals for Item-Sharing vis-a-vis: Re: [Last call] Re: [Chandler-dev] Shipping ticketed-URLs around for individual items via email

[Grant Baillie]

Hi, Mimi

One random thought: If we are going to require emailed items to be on  
the server, then we don't need to send out the eimml at all. In other  
words, we could send an attachment that tells chandler how to get to  
the item on the server, but we don't need to send a local copy of the  
item, and therefore we wouldn't run into the whole email vs shared  
item conflict issue.

The downside of requiring the item to be shared means that there are  
now two servers we need to converse with before sending email, so  
twice the chance of an error. Also, I think the UI needs to reflect  
that the item is being "shared", not  "emailed" (i.e. so there is no  
confusion with what is expected of a "regular email client").

--Grant

On 12 Mar, 2008, at 10:14, Mimi Yin wrote:

> Here is what I understand:
>
> 1. If we want users to be able to click on something in an email and  
> subscribe to an item with Chandler Desktop (without going through  
> the Hub), the best way to do that is to keep the EIMML attachment  
> around so that Desktop users can click on it (like .ics attachments).
>
> 2. It also seems like we *do* want to keep around the ability to  
> download email directly into Chandler to facilitate item-sharing  
> scenarios so that Desktop users can receive invitations to subscribe  
> to items directly in Chandler Desktop.
>
> Both #1 and 2 are nice to haves however, they complete the item- 
> sharing workflow, but they're not things we have to have in the  
> short-term.
>
> On the other hand, we already do #s 1 and 2, so there's an argument  
> to be made that keeping them around isn't a big deal so long as any  
> bugs they cause are also not a big deal.
>
> My current feeling is that "spurious conflicts" caused by Edit/ 
> Update are not that common and something we can live with. Does  
> anyone else been experiencing a lot of 'email-related' bogus  
> conflicts?
>
> That being said, we *still* need to solve this problem where today  
> users can send out an item via email and then put it on the server  
> after-the-fact.
> - The original item is sent out without a read-write ticket.
> - The recipient could then put the item on the server and prevent  
> the sender of the item from gaining access to that item on the  
> server, or vice versa.
>
> I think this pretty much means that we need to *always* publish  
> items to the server when they're emailed. Meaning, item-sharing is  
> done via the server, not email, but email is used as a way to  
> address items to specific recipients and transport the item to  
> recipients' email clients.
>
> Proposed Workflow:
>
> 1. Create new message.
>
> 2. Hit Send
> - Item is automatically published to the server + read-write  
> ticketed URL
> - If user doesn't have sharing account, they are prompted to get  
> one ;)
>
> NICE-TO-HAVE
> - Ability to specify read-write versus read-only in the DV or a pop- 
> up or something...
>
> 3. Recipients receive email + ticketed sharing URL + EIMML  
> attachment in their email client.
> - Casual Collaborators click on ticketed sharing URL to access item  
> on Hub;
>
> NICE-TO-HAVES
> - CC can add item to their account
> - Desktop users can click on attachment to automatically subscribe  
> to item with Desktop, if they are not online when they click on the  
> attachment, they should still be able to see the item added to the  
> Desktop client
>
> I know we've gone back and forth on whether to keep the EIMML  
> attachment + inbound email functionality, but I think I'm finally  
> getting a better picture of what's required to make the item-sharing  
> scenarios work smoothly. So in other words, I think we're close!
>
> Still, I'm interested to hear if I'm still making assumptions that  
> aren't necessarily necessary ;) and/or if anyone else has  
> alternative workflow ideas that would simplify the way we do item- 
> sharing without compromising the user experience too much.
>
> Mimi
>
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
> Open Source Applications Foundation "chandler-dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/chandler-dev