In the APML 0.5 draft,  Source Ranking is used to indicate the 'level of trust" that a user assigns to their "sources of information" which is an excellent primary cut but I think it leaves out an important piece of information -- that of "attention span"

While it looks like the 'profile" attribute will allow large shifts in perspective (home vs work vs hobby come to mind) I think something is needed to reflect different time spans of interest according to the source of informtion.   If I explictly set both CNN and National Geographics because I have an interest in natural history the stories I expect from each are quite different in nature.  eg.  while I expect CNN to tell me  (briefly) about a pod of dolphins beaching themselves this morning in Florida,  I expect National Geographics to give me a long article about communication in various of marine mammals. 

Each of these sources provide stories with different relevant time spans... I'm more likely to make a quick decision about selecting the detail or dismissing for the CNN story and leave the National Geographics story till a point where I have a longer time. But applications that 'drop' items according to age currently have no way to select between the two.

Hence I'd like to be able to say the National Geographics story should not be dissmissed from my application output (be it ticker tape or whatever) as a result of 'old age' nearly as quickly as the cnn  story.... the cnn story of today is unlikely to be of interest to me tomorrow... the national geographic  still will be  - I just haven't been able to get around to it.

Suggestion: Source Ranking be expanded to include:

Source Rankings

  Souce 

        Name

        GUID

         Rank

         Lifetime  (count in days{?} an item from this source is likely to be of interest for)

another example of liftetime or span could be sports - when its cricket season I am interested in cricket but outside the season I do not care as much about what is happening with cricket

As an Automated solution to this, we could correlate the length of an article with the length of time it stays relevent (I.e. the more content, the longer the decay). So short items are usually more alert tpye stuff (Someone logged into Tangler) where-as long stuff is more in-depth content you want to read any time over the next 7 days and should stay on the ticker longer.

I don't think that is quite it Chris...  there is not always a direct one -one-correspondence between length and longevity (between length and currency yes, but that is not the same thing at all).

I really think it is the nature of the media publisher rather than the article length in my example.

Hey Chris... even great minds can sometimes differ ... worse when one mind has an intimate knowledge of something and the other is simply poking around...

But I came to this forum for my post rather than to the Touchstone group  (yes I know,  when?) ..because I think it belongs in the base spec not just the vendor extensions area... a vendor still has the choice of implementing the full capability or not.... but I think this is something that I would want to port if I picked up my marbles and joined another game 

But then I'm just poking around.... and never claimed to be a great mind  ... but it isn't covered by the source ranking in any way...I rate both CNN feed and National Geographic high on the source ranking but want different results in terms of longevity.  I rank FOX News very low and don't expect to see it around at all ... here longevity is non-existant. and not a concern but it becomes a concern as my rank gets higher and higher from -4.9 all the way up to 5.0!

Re:lossy

Ok I'm familiar with the way the term is used in terms of jpeg etc.  but not sure how you are using it here...I'm  not at all sure why it would pertain to the explict parts of my profile... (implicit yes but explicit?  why would this be ... I put it in the profile because i want it there surely?) 

Haha

It looks like the major changes in the spec in 0.6 are the addition of "updated"
attribute to  deprecate nodes older than 30 days, and the application
identifier.

While I  think I follow the concept of casting out
old contributions so the profile doesn't become "stale"  (which i guess
pre-supposes that profiles are likely to be used by multiple apps as
the user's needs, desires etc change). -- I don't see how this "implicitly goverened attention span" adds a lot more value to the user on an onging basis than allowing the user to more readily define usefulness of the "explicitly defined attention span" as I suggest.   Perhaps it boils down to a matter of control... anarchist user that i am I want control!

so my last kick at the cat:

please consider my suggestion  ...

and
meanwhile I'll see how the file(s) i've generated so far meet my
needs;  either way I'll give feedback on whether it looks like v0.6 is
a small or great leap forward....

cheers, Dave