Why I NEED a fuzzy search right now

Veli Voipio said:

the user could hum a melody and the computer finds the correct hymn

There is software that does that. http://www.midomi.com/ 

MJ. Smith said:

I finally have my copy of Julian's Dictionary of Hymnology.
I have discovered that the hymns are not linked to their texts in Watts, et. al.

That is unfortunate!

MJ. Smith said:

Therefore, I need to do my own linking which thanks to L6 improvements is possible.

Wow, I can't imagine the time and dedication that would take, even if you had fuzzy search available! And this is the sort of thing Faithlife should be doing. Aren't you worried that you'd get halfway done with that monumental task only to find out they'd decided to ship an update of the Dictionary of Hymnology with better tagging included?

MJ. Smith said:

I've indexed hundreds of hymnals so, yes, I know what I'm getting into.

Oh wow, could I hire you to index Hymnal: A Worship Book (the hymnal my church uses, along with many other Mennonite churches in the US and Canada) and put its contents up on hymnary.org? I volunteered to do it a long time ago and have never gotten around to it. Too busy. But I would still love to see it up there. It's an important hymnal and a major lacuna in their collection. I could get you a copy of the book if you're interested.

MJ. Smith said:

However, I figure it like washing the car to make it rain. If I don't get started, Faithlife won't do it

Ha! You're right. That worked for me doing Martyr's Mirror. I got through about 1/8 of that massive 1000+ page book turning it into a PB when they announced it in pre-pub. The collection it's in still has a long way to go, though, so I probably should keep chipping away at my PB.

I'll admit, I've tried to 'divine' the 'mind of Logos' over the years.  In some cases, you can see they're building the base for later 'frosting'.  And in others, they're 'looking for Mr Goodbar' .... sales.

But their search engine remains a mystery. Clearly speed (indexing), stuff (broad groupings), converted stuff (topics, etc), and fancy stuff (languages, etc).

But the basic concept is still the same as their early days (and/or and hints of not), with shades of sophistication in the older Libronix fuzzy search.  

My impression is they're being passed by conceptually by very basic mobile apps.  

We're upgrading our 'nutrition' so I found an app that searches for why you're having a problem with XYZ.  Although the app is nutrition vs XYZ's (many to many), you can put in just about anything totally unrelated to food, habits, etc. and it does a fairly decent combinatorial analysis and presents the results very logically (complete with warnings about stupidity).

I always feel it's unfortunate for the Logosian Biblical researchers, hymnologists (you), and so forth. Still typing-in complex formulas, exacting text, and multiple combinations that could so easily be handled by algorythms available since the late 1980s.  And likely with expertise from the local community college, no less.  I'd give good odds, there's Logos staff that would love to deliver better results then 10,522 search hits with very poor user delivery.

Rayner said:

Fuzzy search is a good idea, but I'd be curious to know precisely how fuzzy it would be.

Fuzzy Wuzzy was a bear, 
Fuzzy Wuzzy had no hair,
Fuzzy Wuzzy wasn't fuzzy,
was he?

I was really tempted to try Fuzzy Wuzzy but here's an example from Libby.  Scan about half way down to see how it begins recombining words, etc and showing the percent match.

..

I tried fuzzy wuzzy on the library. Zippo.  However the phrase 'had no hair' was interesting, as the logic substitutes words (heir for hair, head for hair, etc).  Also, a fuzzy wuzzy search is fairly quick (not much to work with) while the no hair search was quite slow (many substitutions).

..