It occurs to me that one of the problems I would like to solve as a user
of a search engine of any kind, the tune finder being just one of them -
is some indication (no matter how slight) of the quality of the result.

If even a small number of content providers consistently use the
Z field to provide their name and the date of transcription
, and the tune finder returns it as part of the search, I think users
will quickly do something like this:

1. search for 'Lady Ann Montgomery'
oops - 30 returns.   six are were transcribed by Henrik Norbeck, and
of those, three have a more recent date.  I know Henrik's been reliable
in the past, so I'll pick one of those.

2. search engine quietly adds to the ranking of the one picked. This could be
per tune, or could be per site, which makes for a smaller amount of processing.

3. next person comes along and asks for a tune, the higher ranked sites get
presented first.

Over time, the more often a transcriber's tunes get picked the more likely they
are to be presented at the top of the list to the subscriber.   Tunes which don't
use the Z field will still be presented, but will be unlikely to be picked unless there
are few others presenting the same content. It's not necessary to parse any given
field, simply present what is likely to already be there for tune collectors who've
been doing it for a while.  If there are a large number of duplicates, only the first
N could be presented without major harm.

Thoughts?

John Chambers wrote:

> | I use the Z: field to write I made the transcription, and give
> | then only the email adress. I found conveniant to write the Url
> | after it, so I don't use the F: field for this because I would
> | have been redundant. Do you think we should generalize the use of
> | the F: field for this purpose ?
>
> That depends.  Do  you  just  want  the  information  to  be  present
> somewhere  in  the headers?  Or would you like software to be able to
> use the information sensibly?
>
> If you only care that information be present, and don't care that  it
> be  usable, then it's fine to put it anywhere you like.  If you would
> like the information to be useful (so that not just a human, but also
> a piece of dumb software can understand it), then it's a good idea to
> consistently put the same information in the same place  in  all  the
> tunes.
>
> This is one of the reasons that I haven't  bothered  having  my  tune
> finder  keep  track  of a lot of the fields that people would like to
> ask about.  If you look at much online  abc,  you'll  soon  see  that
> people  are  supremely  sloppy about where they put what information.
> The C field is routinely used to hold place names, years, publishers,
> and  other  information.   The  O  field  is  used for just about any
> information you can imagine.  Titles contain composer names and  html
> <blink> tags.  And so on.
>
> In some cases, this is clearly intentional obfuscation.  For example,
> recently I saw a line like:
>    K:D % EDorian
> This made it clear that the transcriber  knew  the  correct  key  and
> mode, and was intentionally giving incorrect information.  Presumably
> this was to defeat people using software to look for things by key.
>
> When people do things like this, there's a real temptation  to  throw
> up your hands, and go play some tunes to cool off.
>
> To subscribe/unsubscribe, point your browser to: 
>http://www.tullochgorm.com/lists.html
To subscribe/unsubscribe, point your browser to: http://www.tullochgorm.com/lists.html

Reply via email to