Ok, timing is important - but if tag authors don't find it usefull for
themselves they wont do it.
Elias Torres wrote:
After exciting vacations all around the blog Dave is back! Glad to have you.
Dave Levy wrote:
Don't know if anyone else has replied
Macros to present these as HTML buttons would be better in my opinion
I definitely don't mind having HTML (we have the beginning of these
macros already) but I do think that an API would be useful.
The top 200 and the others change all the time i.e. the set members , so
top 200 as RSS seems a nonsense to me. (Sorry rahul!) Also tags/user
tags/blog are a bit of an intersection. But
Agree. RSS/Atom for such a (sliding) set would be awkward. It's fine for
the first time you access it, but it would make no sense after that and
it would be crazy if a user were to subscribe to it. Atom is not the
answer to all data API/protocol needs!
I think the following queries with buttons would be good
Top 200 (with a parameter) ie #showMostRecentTagsUsed(Howmany)
This needs a time period as well so over the last 24 hours, last week?
#showRecentTagsUsed(Howmany Since)
and over what time period Most Recently Used ie #showRecentTags(Howmany,
Since boring)
The boring parameter sets a threshold like del.icio.us which states at
what point the questioner becomes interested in a tag. A site front page
might like a hot topics query
#showRecentTags(10 24h 5)
which might return the most used tags over 24 hours provided they're
over 5 uses
or an innovation query (Ellias's influencers)
#showRecentTags(30 24h 1)
So Boring=0 becomes special in that the sort order become time, whereas
in all the other queries the sort order is usage.
Is this a sufficient pragma?
(qty_displayed, time_horizion, interest_threshold)
All these suggestions for macros are great, but I'm worried we are
thinking we could end up with incredible tagging analysis and summaries
on the first pass at it.
Lets do it slowly then, but there must be some read functionality
I'm definitely not a tagging expert so I'm
trying to focus on making sure we capture the data and do something
efficient for the basic common UI paradigms we see today in tags.
Some of what I suggest above is an attempt to reflect your ideas of
measuing strong authors output.
Once
3.1 is released we can experiment/research other ways to show higher
quality nuggets of information from the dataset (plugins, models, etc).
I believe that the tag space is a bit understudied *and* overhyped, so
we should be careful how much we want to do for this release.
-Elias
i.e.
Rahul Jain wrote:
Not sure, if this capability is already there, but it will be nice to
get a feed for just the tags themselves, such as feed for:
- top 200 tags
- tags for a particular blog
- tags for a particular user
Either these feeds or a REST api for this data, may help if someone
wants to consolidate the tags from the blogging engine and some other
application that has tags in them. Thanks,
rahul
--
Dave
<http://www.sun.com> * David Levy *
*Sun Microsystems Ltd.*
55, King William St.,
London EC4R 9ND United Kingdom
Phone +44 (0) 20 7469 9908/x18308
Mobile +44 (0) 7710-360922
Blog http://blogs.sun.com/DaveLevy
Email [EMAIL PROTECTED]
Sun Proprietary & Confidential . This e-mail message is for the sole use
of the intended recipient(s) and may contain confidential and
privilidged information. Any unauthorised review, use, disclosure or
distribution is prohibited. If you are not the intended recepient,
please contact the sender by reply e-mail and destroy all copies of the
original message.
