Re: [datatracker-rqmts] Slides for the face-to-face BoF next week

[Yaron Sheffer <[e-mail]>]

Sort of. I think Atom feeds should be the only primitive provided, and 
the user should take it from there. It can be as easy as using Google 
Reader, or you can use a much fancier reader application.

The server should have no notion of lists who are owned by somebody, 
only of anonymous feeds. The draft mentions that feeds should be based 
on lists, but I see no reason to have lists in the first place.

One drawback of this direction is that there will be overlaps (e.g. 
between "documents of the IPsecME WG" and "Documents that have 'ipsec' 
in their title"). Perhaps some feed readers can de-duplicate such cases, 
I haven't seen any that do. Alternatively, we may have a powerful enough 
feed creation mechanism, so that you can specify the union of these two 
feeds as a single feed. Then you can sort/filter the resulting feed on 
the client side if necessary.

Thanks,
	Yaron

On 11/03/2010 04:20 PM, Paul Hoffman wrote:
> At 4:09 PM +0200 11/3/10, Yaron Sheffer wrote:
>> Hi Paul,
>>
>> I can see two ways of dealing with this issue, there are probably more:
>>
>> - Make the Atom feeds not only per-draft, but more sophisticated. Example feed templates:
>>
>> * All drafts authored by X
>> * All drafts of the Y WG (which we already have, BTW)
>> * All drafts whose name/title includes the term Z.
>>
>> - Alternatively, we could publish and maintain OPML files for common collections, e.g. one per WG. [I know you dislike OPML, but it works for me]. But this in an inferior solution because people would have to update the OPML file (list of feeds) manually.
>
> OK, I'm glad I asked. It sounds like you are in favor of some of the list-creation parts of the draft, but not the "my list is stored on the IETF server" part, and not in favor of "I can see my list displayed on the Datatracker web site" part. Is that a fair summary?
>
> --Paul Hoffman, Director
> --VPN Consortium