collapse

* Member Info

 
 
Welcome, Guest. Please login or register.
Did you miss your activation email?
I'm gonna sing this verse, I said, I wasn't gonna sing no more - Last verse of Frank Stokes' last recording, Memphis Rounders Blues, September 30, 1929

Author Topic: Mother of all tables  (Read 6034 times)

0 Members and 1 Guest are viewing this topic.

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Mother of all tables
« on: March 07, 2009, 07:03:39 PM »
I was thinking how cool it might be to have a big sortable table of songs that included bar form, and position|tuning. We could extract them from the forum topics currently running on same.

So you could get, for ex., a list of 8 bar blues (probably with a "sub-type" qualifier, since there are two (three?) main types of bar forms of the 8 bars, you'd have to come up with tight nomenclature to convey exactly what you were talking about). Furthermore you could then sort those by position.

So if you had a happenin' version of, say, Crow Jane, you could find a bunch of other tunes like it, say Dryland Blues, Becky Deam. Likewise the various ragtime forms, waltz time and 16 bar blues. If you got real inclusive and included 12 bars / modified 12 bars it would be a big list but more useful and complete.

Then there are the totally unclassifiable left-field things, Wilkins, Robert Pete, you could classify them as Unusual, and easily find all the weird stuff. What other forms are there?

I'm thinking we could build a mother of all tables that extends what we're already doing with the song tables on the wiki, consolidating them all into one comprehensive view.

Here are some of the topics, and we could link back to the particular post that ref'd a particular tune I guess:

16 bar
Waltzes, 3/4
8 bar
rag blues & circle of 5ths

Other columns could be:
Chord content: With minor chords, with diminished chords (we haven't done that one yet), 9th chords, 6th chords, etc
Feel, time: Swing, straight time
Vocal phrasing: Long, short, both, also in this topic

You could further classify by other distinct columns like:
Ensemble type: solo, duo, jug band, string band etc
Instrument columns: guitar 6 string, guitar 12 string, guitar reso, piano, mando, fiddle, harp, etc. Might be too cumbersome, but really useful nonetheless.
Slide: Yes|No
Picks, string gauges, capo: -- nah! Maybe in a 'playing setup' column
Content: Murder ballad, disaster ballad, railroad song, etc
Folk tale group: St James Infirmary, John Henry, Casey Jones, Stagolee, The Dozens, etc
Dates: First recording
State: MS, TX, LA, etc
Style: Delta blues, Piedmont Blues -- Total can of worms but 'Unclassifiable', 'Ambiguous' and 'N/A' are always options
Notes: Anything else not covered

Might be getting a bit ambitious here but talk is cheap. I think it's a logical progression that would reflect the way things have evolved on the forum. In other words distill that work and thought into a single sortable table. Since we're concerned with historic recordings, once the work is done, and it would take a while, it's done. So the workload is finite.

It would be nice to get it all into a database someday but we could probably do 90% of what would be needed in a simple flat master spreadsheet, from which we could export wiki marked-up tables.
« Last Edit: March 07, 2009, 09:38:51 PM by Rivers »

Offline Richard

  • Member
  • Posts: 2416
  • Drove this for 25 years!
    • weekendblues
Re: Mother of all tables
« Reply #1 on: March 08, 2009, 01:45:07 AM »
Can you get help for database addiction  :-X
(That's enough of that. Ed)

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #2 on: March 08, 2009, 01:48:04 AM »
Rich, it's what I do for a living. :)

Not only that, it's just plain logic. Why have it split all over the place when simple analysis tells you you can get the complete picture. A song is a song, all that additional information is related to that song. Logic rules, you don't have to be a database expert to figure that out.
« Last Edit: March 08, 2009, 01:53:34 AM by Rivers »

Offline Richard

  • Member
  • Posts: 2416
  • Drove this for 25 years!
    • weekendblues
Re: Mother of all tables
« Reply #3 on: March 08, 2009, 01:16:35 PM »
I did once back in the days of hard coding Paradox and later in Delphi, but I have to say I remember little about it now - a sign of olde age!
(That's enough of that. Ed)

Offline Slack

  • Administrator
  • Member
  • Posts: 9213
Re: Mother of all tables
« Reply #4 on: March 08, 2009, 01:33:44 PM »
Sounds pretty interesting Mark -- also sounds like a lot of work, but maybe not - as you point out, we have a lot of info already, some data entry is needed to tie it all together and maybe if the effort was organized efficiently, it wouldn't be too bad.

Offline arlotone

  • Member
  • Posts: 126
    • arlotone.com
Re: Mother of all tables
« Reply #5 on: March 08, 2009, 04:45:17 PM »
I'm a PHP/MySQL developer and this would be pretty easy for me to build. In fact, I built a similar tool for the Old Town School of Folk Music -- an online index to their songbook, searchable by chords/keys/tempos/time signatures as well as user comments. So if you need a song in C that doesn't have an F chord and has something to do with the labor movement, you can find it (if it exists).

If there's a consensus this would be useful, and someone else can enter the existing data from the wiki format, and it can live on the weeniecampbell.com host, I'll volunteer to build it.

Offline Johnm

  • Global Moderator
  • Member
  • Posts: 13190
    • johnmillerguitar.com
Re: Mother of all tables
« Reply #6 on: March 08, 2009, 08:51:49 PM »
Hi Rivers,
Since my previous post was erased, I'll do a quick and dirty recap:
   *  I think the proposed table sounds like a "make work" item, and since the information targeted for inclusion on it is already on site, the time spent putting it together would be better spent conducting new research, doing new transcriptions, and engaging in new discussions.
   * One effect of mass categorizations of the type proposed is that they have the effect of flattening out the differences in the categories.  This is true even within the smaller categories with threads already running, like 8-bar blues.  John Hurt's "Sliding Delta", Ishmon Bracey's "Woman, Woman", Furry Lewis' "Dryland Blues" and Henry Johnson's "Crow Jane" could all be categorized as "8-Bar blues with Instrumental extensions at the end of the form".  If you are familiar with these pieces, though, you realize the extent to which the fact that they all share that category is little reflection of how very different they sound from each other.  With other categories, meaningful conclusions based upon a shared categorical status are even harder to come by:  What generalization can be made about Mississipi blues on the basis of the fact that Big Joe Williams, Hacksaw Harney, Bo Carter and Roosevelt Graves all came from that state?
   * To the extent that there are long-standing categorical threads like "Adventures in Spanish", it makes sense to keep adding new entries as appropriate.  I'm not sure what is to be gained by lumping together a great mass of categories from which correlations will be difficult to ascertain.
If you want to make the table, by all means make it, but I think more Leadbelly transcriptions or Willie McTell transcriptions would make more sense.
all best,
Johnm       

Offline dj

  • Global Moderator
  • Member
  • Posts: 2833
  • Howdy!
Re: Mother of all tables
« Reply #7 on: March 09, 2009, 04:37:00 AM »
As someone who's always been a list maker, I can see the benefit to a sortable table of the type proposed.  It would be interesting to sort on waltzes by region to see if any one area was more pron to producing artists who recorded waltzes.  Same with 8 bar, 16 bar, or unusual form blues - did they predominantly come from one region?  There are genuinely interesting possibilities for research there.   

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #8 on: March 09, 2009, 07:40:30 PM »
Yeah, sorry about erasing your post John, it was 5 AM, I meant to hit 'Quote' and hit 'Modify'. I've almost done that before in the past but always caught it before hitting save. Anyway sincere apologies, thanks for reposting, and I promise to never do it again.

I see it as a natural progression, like the wiki was a natural progression. I mean, probably when Dixon, Godrich & Rye proposed the first ed. of B&GR [cue the inspirational music] I daresay a lot of folks thought they were crazy. When JFK talked about going to the moon...  :P

Re. your point about correlations being difficult to ascertain, certainly. But much easier if you can play around with different groupings or views, which we cannot do at present. All your caveats are totally valid though, and could easily be stated in some form of preamble.

Arlotone, thanks for your kind offer of asistance, I'd certainly like to talk to you about leveraging database, but integrated within a mediawiki framework. Sure, we can do a PHP/mySql driven custom page, see the forum Tags index, for example, which is a PHP/mySQL script we integrated into the forum infrastructure using an extension to the SMF software. Key word is 'integrated', I think. Have you played around with PHP/mySQL in a mediawiki context? Me either. This could be our chance!

It would make a lot of sense to have a powerhouse central repository. Trick would be to make it really easy to query and present result sets on (a framework) page without having to have pro-level IT skills. Might be asking a bit much though, given today's technology. Or maybe not, if we think through the requirements up front and code it to meet them.

edit: corrected link to tag index
« Last Edit: November 16, 2021, 07:58:51 AM by Rivers »

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #9 on: March 09, 2009, 08:09:06 PM »
As someone who's always been a list maker, I can see the benefit to a sortable table of the type proposed.  It would be interesting to sort on waltzes by region to see if any one area was more pron to producing artists who recorded waltzes.  Same with 8 bar, 16 bar, or unusual form blues - did they predominantly come from one region?  There are genuinely interesting possibilities for research there.   

Pinging off that idea, likewise 12 string players not from GA, lap style players from Kathmandu, and so on. The database could be very functional with just a single initial table, at the song level.

Then you'd want to add an artist parent table, song session recording date/location children, and so on, like a real application. I've always wanted to do a country blues timeline database also, you could key into that.

Nobody else is gonna write it, so how about we do it? We can take our time here, it's not like it has to meet a commercial deadline or anything. Ain't no more country blues being recorded from the era we're talking about capturing. The information is out there, just nobody's brought it all together, and certainly no group with such active musical backgrounds, combined knowledge base and tech skills.

Of course the danger is we'd have nothing left to talk about on the forum. Personally I doubt that very much.  :)
« Last Edit: March 09, 2009, 08:29:31 PM by Rivers »

Offline CF

  • Member
  • Posts: 899
Re: Mother of all tables
« Reply #10 on: March 10, 2009, 04:52:34 AM »
I definitely understand the desire to reign in all that info out there & have it at your finger tips. My 'Master Table' wet dream is a song family tree that would incorporate just about everything relevant. I say go for it Rivers . . . although it may be a lonely venture. But what a resource! 
Stand By If You Wanna Hear It Again . . .

Offline Bunker Hill

  • Member
  • Posts: 2828
Re: Mother of all tables
« Reply #11 on: March 10, 2009, 06:14:02 AM »
It?s long been my intention to reign in all the books/articles/reviews mentioned and attempt to add to the rather overlooked ?weeniemedia? section of Wiki which somebody has valiantly started. Unfortunately the road to hell is paved with good intentions and my lack of progress being the classic example of it.
 

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #12 on: March 10, 2009, 03:25:53 PM »
Cheap, exactly. You could get it going fairly simply, and also phase-in features down the track, good planning and database design being the key.

BH, a good requirement would be for a system able to pull disparate, rich content together I think. The Tags on the forum fit articles that cover a lot of subjects, they would be perfect if they were just a bit more targeted, i.e. were at the post level rather than the whole topic.

In the meantime if you were to start dropping articles into the wiki we'd all much appreciate it! You can use wiki 'category' tags to link to artists, regions, publications, etc. If they don't exist we can easily create a category for them. I'd suggest, when you're ready, you kick off a new topic on this board and we can figure out the structure, details and generally assist. It's a lot of fun when you get past the learning curve. And really it's not like it has to be done tomorrow. A wise person once told me "If you want to eat an elephant, have an elephant sandwich every day for a year or two".

Other thoughts: We're kind of locked into the dual SMF <-> wiki presentation framework at present which is not a bad thing. But I'm wondering if there's some way we could make new developments reasonably platform-independent so if we ever wanted to move to 'the next big thing' we could migrate everything easily. A database would be great in this respect since it's independent and could be reused in any other content management system. Also on that note, clearly we would benefit from understanding the underlying wiki database better, we haven't had any reason to study it in detail yet, it just does its thing and does it pretty flawlessly so far as I can see.

I think it would be wise to keep in mind that technology changes really fast though, much as we all have come to like the current setup.
« Last Edit: March 10, 2009, 03:33:07 PM by Rivers »

Offline arlotone

  • Member
  • Posts: 126
    • arlotone.com
Re: Mother of all tables
« Reply #13 on: March 10, 2009, 07:52:20 PM »
Arlotone, thanks for your kind offer of asistance, I'd certainly like to talk to you about leveraging database, but integrated within a mediawiki framework.... Key word is 'integrated', I think.... Trick would be to make it really easy to query and present result sets on (a framework) page without having to have pro-level IT skills.

No, I haven't worked with the mediawiki software, but I'm wondering what level of integration you're talking about. Wrapping the page template around it so it's a seamless part of the site? I've done that kind of thing before. Exposing the database data to the wiki search form? That could probably be done without too much difficulty. What else?

In any case, what I'm picturing certainly wouldn't take any special skills to use. But I'm imagining more of a lookup tool (what tuning does this song use?) than an analysis tool (what were the most common tunings in Mississippi in the late 1920's?).

I'll PM you login information for my songbook index so you can get a better sense of where I'm coming from.

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #14 on: March 11, 2009, 05:13:24 AM »
Thanks Arlo, nice clean lookup on OTSFM. I see the BB is vBulletin and the index is on WordPress.

>> I'm imagining more of a lookup tool (what tuning does this song use?) than an analysis tool (what were the most common tunings in Mississippi in the late 1920's?)

Not sure why why you'd want to limit yourself to just a tuning column. If the database design contained other attributes they would enable additional sortable / selectable views, in various combinations with each other. I'm envisaging a fully relational database ultimately, with primary keys, foreign keys and relational constraints, repeating groups normalized out to child tables, join views, all the core database bells and whistles.

Designing and creating the database is the easy part. Initially loading it, maintaining it, and providing intuitive and functional access to it are the big challenges. A picture is worth a thousand words, I'll draw up a draft entity relationship diagram of what I'm proposing.
« Last Edit: March 11, 2009, 05:40:42 AM by Rivers »

Offline arlotone

  • Member
  • Posts: 126
    • arlotone.com
Re: Mother of all tables
« Reply #15 on: March 12, 2009, 08:51:57 AM »
I just mentioned tunings as one example. Of course there could be lots of different attributes of each record. And yes, it would be a relational database. In this example, the index is custom code, with the WordPress template wrapped around it for consistency with the other pages.

My offer is to build something comparable in scope to the OTSFM example, but with different attributes that this site's users would need. I think anyone could use an interface like that without difficulty. I can integrate it with the site if we can define "integrated." As with the wiki, it would be up to the site users to populate the tool with data over time.

I'll take a look at your diagram, but what I'm offering would not be hard to do ... maybe a day's work, which would be a reasonable contribution given all I've learned from this site.

Offline Rivers

  • Tech Support
  • Member
  • Posts: 7274
  • I like chicken pie
Re: Mother of all tables
« Reply #16 on: March 14, 2009, 08:14:04 PM »
Thanks Arlotone. Gimme a couple of weeks, I'm researching the possibilities on my home setup.

Tags:
 


SimplePortal 2.3.7 © 2008-2024, SimplePortal