[Tagdb] Nitins talk and future of tag systems

[Nitin Borwankar]

Hi guys,

Sorry for the delay in responding.  I have been in transition - 
finishing up with a previous assignment and spending some time at the 
Odeo offices.
I will probably do some work for them in the near future both tag and 
non-tag related database work.

The talk is frankly not recommended for a number of reasons but I do 
promise to get a blog article done real soon about a couple of new 
issues that I spoke about at the talk. 

One of them is a uniform notation for User-Tag-Item "related" queries of 
arbitrary depth and some issues that come up
when you do a query that is "too broad" or includes things in the "head' 
of the distribution vs the tail. Join vs Sub-query vs decoupled selects
to achieve the same result - which is more scalable.

The other topic is data 2.0 and privacy considerations - how should data 
be distributed across handheld storage (iPod/USB keychain/phone), 
desktop/laptop and web-based public server ( Yahoo/Google/MSN...) .

First one coming up over the weekend.

The talk was suboptimal because of a number of reasons - my laptop 
crashed  the same week and I had to move my environment  to a firewire 
drive.
Not all my app licenses worked after the move.
I spent my time getting my environment right and didn't work enough on 
the content of the presentation, a low key informal, "maybe 5 people", 
audience turned out to be about 20+ people crammed into a small room,   
I was somewhat disoriented by all this and rambled more than a bit .....
So audio is not the best medium to capture the ideas there - I'll make 
up for it with the blog posts as promised.

Nitin


Philipp Keller wrote:

>Hi everybody
>
>I just want to make a ping here.. :-)
>
>Nitin: How was your talk? Is there a mp3 or some script available?
>Or can I motivate you to write a blog post on your talk..? :-)
>
>I wrote a new article about the future of delicious and big scaled tag
>systems in general [1], maybe you can help me out with a few comments or
>something? 
>
>The article is about the performance problem this big scaled systems are
>facing. I believe the problem hasn't got too much attention from the
>croud so I hope I can stir up a little discussion. I'm sure that the
>whole tagging thing must go further now than three tables and a bit of
>query optimisation and cache..
>
>Did someone here set up a tagging system with Lucene or similar and are
>some performance tests available..?
>
>greets
>Philipp
>
>[1] http://www.pui.ch/phred/archives/2005/08/does-delicious-scale.html
>
>_______________________________________________
>Tagdb mailing list
>Tagdb at lists.tagschema.com
>http://lists.tagschema.com/mailman/listinfo/tagdb
>  
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: nitin.vcf
Type: text/x-vcard
Size: 150 bytes
Desc: not available
Url : http://nelson.textdrive.com/pipermail/tagdb/attachments/20050902/0c124dd7/nitin.vcf