Search

Discussion

Re:Apache (kris)

gatecrasher

Member
Posts: 14

gatecrasher @ 2004-04-03 10:32:53 UTC

I think Bert Weedon's version should show up as a cover and not only in a comment. If somebody's doing a search for the artist Bert Weedon now,

he won't find info on "Apache" in the database.


There are two ways to build a cover database:

- consider the first recorded version as the original

(-> originals.be, they list the first recording and performance (e.g. in movies))

- consider the first released version as the original

(-> your database and coverinfo.de [though they're wrong with "Apache", I guess they copied this bit from "The Originals"])


If I'm right, your approach is different from originals.be and you should be careful

when using infos from originals.be, as there are many old songs, for which the

first recorded and the first released version are not the same (obviously you are careful already).

This is such a case, and from your approach Bert Weedon's version must be

seen as a cover.

Of course in such cases I think you should add comments to the first released and also to the first recorded version in your database to make everything clear for visitors.


Everything IMHO.

I only want to help keeping everything consistent with the approach.


Keep up the good work, I love it that you're doing such good research!




Bastien

Manager
Posts: 35819

Bastien @ 2004-04-03 13:11:09 UTC

This is how we work: The original is the first released version, except if it's intuitively incorrect. In that case we take the first recorded version.


Now, concerning the Apache story and Weedon's version. The first released version is by The Shadows so that's the original. We also know Bert Weedon recorded a version earlier. However, we have no reason to take Weedon's version as original, as nobody heard this version and everybody based their version on the Shadows' version. We can thus keep The Shadows as original.

Bastien

Manager
Posts: 35819

Bastien @ 2004-04-03 13:13:45 UTC

There are two ways to build a cover database:

- consider the first recorded version as the original

(-> originals.be, they list the first recording and performance (e.g. in movies))

- consider the first released version as the original

(-> your database and coverinfo.de [though they're wrong with "Apache", I guess they copied this bit from "The Originals"])


In fact none of these two systems work for all songs in general. I could give lots of examples where rule one or two doesn't make sense. That's why we have a more general, less exact guideline, which is a combination of both. Classifying music is not an exact science Unhappy

Bastien

Manager
Posts: 35819

Bastien @ 2004-04-03 13:15:20 UTC

If somebody's doing a search for the artist Bert Weedon now,

he won't find info on "Apache" in the database.


Remains this very valid point. I have to admit I don't know what to do. It's not an original, as it has not been covered, and it's no cover either, as it was made before the Shadows version! That's why I suggested mentioning it in the comments only.

gatecrasher

Member
Posts: 14

gatecrasher @ 2004-04-03 20:30:54 UTC

However, we have no reason to take Weedon's version as original, as nobody heard this version and everybody based their version on the Shadows' version.


Nobody heard it? Weedon said in an interview that his version was promoted and was on the radio at the same time as The Shadows'.

An what makes you so sure that later cover versions are based on The Shadows' version and not on Bert Weedon's? Have you got info on that from all cover artists?


And do you really think an original can only be taken as an original if covering artists heard it and purposely based their cover versions on it?

If so, you will get into serious trouble. There are plenty of cover versions that are

based on an earlier cover version sometimes because the cover artist didn't even know the original.

Also it is nearly impossible to decide for every cover artist whether he knew/heard the original version and which version his cover version is based on.


I think every cover database should have their definitions and follow

them.

Please give your definition for the terms:

- cover version

- original version

(IMHO you should also put them on the site for every visitor to read,

the other sites state their rules, aims and definitions)


I don't see your problem with rules and definitions, they guarantee for

consistency which is very important for visitors that want to understand

and use the data.

For example, if you state as a rule "orginal = first released version" and

then always go by that rule, you will have a consistent database.

Of course you may see problems in some cases (but you also have the

problems with your "less exact guidelines", as you see), but the

problems that arise by following such a rule can be faced with comments

which explain the case to the visitor.


If you don't want to have exact rules it will be very difficult and you will have

to find other ways to make the data understandable for visitors.

But your method to define an original "by intuition" is highy debatable,

subjective and will often lead to no "accurate" conclusion because you

don't know all the facts and circumstances.


A visitor can only make use of the data in your database if he

knows its semantics. It doesn't help him if he doesn't know whether one

of the songs that you title as original is a first release or only a first recording.

He will also be confused about the year you put in brackets: is it the year of release, the year of recording, the year of performance...?

In a consistent database it is defined what the year means.

But if the semantics of every entry can change because you don't have exact rules and definitions (e.g. what is an original), you have to explain much more to the

visitor and write more comments.

Of course you could do that if you really wanted to continue following

your "less exact guidelines".


But at the moment your database seems inconsistent and that will confuse your

visitors.

Hope this discussion helps to make everything more understandable and

consistent.

Because all your efforts to be the most accurate database are pointless if

the visitor doesn't know how to interpret the data.



Sorry, my posting is way too long. :-)

Bastien

Manager
Posts: 35819

Bastien @ 2004-04-03 21:03:23 UTC

Nobody heard it? Weedon said in an interview that his version was promoted and was on the radio at the same time as The Shadows'.


Source?


An what makes you so sure that later cover versions are based on The Shadows' version and not on Bert Weedon's?


Because of the above. I thought no one knew Weedon's version.


And do you really think an original can only be taken as an original if covering artists heard it and purposely based their cover versions on it?


No, that's not my opinion. It's just that this case seems like a very special case so I was thinking in a broader context.


Also it is nearly impossible to decide for every cover artist whether he knew/heard the original version and which version his cover version is based on.


This is irrelevant for our guidelines so don't worry.


I think every cover database should have their definitions and follow

them.

Please give your definition for the terms:

- cover version

- original version

(IMHO you should also put them on the site for every visitor to read,

the other sites state their rules, aims and definitions)


We are using definitions for our concepts, but I just haven't written them out for the public yet.


I don't see your problem with rules and definitions, they guarantee for

consistency which is very important for visitors that want to understand

and use the data.

For example, if you state as a rule "orginal = first released version" and

then always go by that rule, you will have a consistent database.

Of course you may see problems in some cases (but you also have the

problems with your "less exact guidelines", as you see), but the

problems that arise by following such a rule can be faced with comments

which explain the case to the visitor.


I guess if you'd be working on the the data you'd understand. Everything seems quite straightforward until you try to classify the data. It's just impossible to fit in non-exact data in an exact system. To make it more concrete

  • You can not take the first release as the original because a lot of songs, which everybody considers as the original versions of the songs (think of old Dylan songs he performed in 60s but only released decades later) are released years after their covers.
  • You can not take the first performance as the original because this information is way too hard to find. It's even unclear what a first perfomance could be.

If you don't want to have exact rules it will be very difficult and you will have

to find other ways to make the data understandable for visitors.

But your method to define an original "by intuition" is highy debatable,

subjective and will often lead to no "accurate" conclusion because you

don't know all the facts and circumstances.


Exact rules are impossible to work with because they yield counterintuitive results. So yes, we choose to have subjective guidelines, which yield debatable results.



A visitor can only make use of the data in your database if he

knows its semantics. It doesn't help him if he doesn't know whether one

of the songs that you title as original is a first release or only a first recording.


Here you make an important point. The years filled in are release years, except were noted. Problem is we have not been consistintly adding comments in the latter case.


But if the semantics of every entry can change because you don't have exact rules and definitions (e.g. what is an original), you have to explain much more to the

visitor and write more comments.


And again a very valid point.


So what are the conclusions?

1) We choose to use subjective guidelines, because we don't think there is another way to do it. We accept the results are debatable, that's a logical result of subjective guidelines.

2) You are right I we should make our guidelines public, such that our visitors would understand are database better.

gatecrasher

Member
Posts: 14

gatecrasher @ 2004-04-04 05:05:06 UTC

The sources for the story of "Apache" have been given in this thread:

http://www.bertweedon.com/interview.htm

http://www.mcr26.freeserve.co.uk/shadows/Lordan/Default.htm


These sources also say that Weedon's version was recorded in early 1960

and that The Shadows recorded their version later on June 17th.

I've done more research for you about release dates and chart positions.

It seems clear that both versions were released in July 1960

(http://www.rocknroll.freeserve.co.uk/july.htm)

and that The Shadows got to #1 and Weedon got to #24 of the UK chart.

Alltogether it is likely that The Shadows' version was released ca. 1 week

earlier than Weedon's, though an "expert" in the newsgroup rec.music.rock-pop-r+b.1950s says otherwise:

http://groups.google.de/groups?selm=3a7d80ae.33415843%40news.dircon.co.uk


One source has 21.07.1960 for The Shadows and 28.07.1960 for Weedon (but these seem to be chart entry dates rather than release dates):

http://www.skidmore.edu/~gthompso/britrock/60brchro/60brch60.html

Another source has 23.07.1960 and 30.07.1960 as chart entry dates:

http://www.chartwatch.co.uk/tobias/CLUK5_S.HTM

http://www.chartwatch.co.uk/tobias/CLUK5_W.HTM


So we can say that both versions were around at the same time and were

"well-known". Also the guys from the newsgroups seem to "intuitively"

consider Weedon's version the original:

http://groups.google.de/groups?selm=3a674a9b.10631664%40news.dircon.co.uk

http://groups.google.de/groups?selm=3e127693.2996040%40news.blueyonder.co.uk


I think that's funny because you didn't even want to have Weedon in your

database. :-)


I guess if you'd be working on the the data you'd understand.

I think I do a little bit because I've seen those problems when contributing many

submissions and corrections to other databases (mainly coverinfo.de, also

the-breaks.com).


It's just impossible to fit in non-exact data in an exact system.

The problems depend on your definitions, if you define

"original = first release" then it will be easy, and for the few problem cases

you can put additional infos on previous recordings/performances in a

comment field, so you don't loose important information.

I can imagine why many music experts/journalists/radio people make

the assumption "original = first release", because:

- release dates are "hard facts" and easier to find than dates for

(pre)recordings and performances and thus you will have less

mistakes and more consistency in your results

- a song that hasn't been released (as sheet music or medium) virtually doesn't exist for them, it is not accessible to the public, it can't show up in charts

and often it is not even registered with a rights organization (e.g. ASCAP),

so it can have nearly the same significance as the songs (for example)

a teacher is "inventing" when improvising with a guitar in front of his class.


  • You can not take the first release as the original because a lot of songs, which everybody considers as the original versions of the songs (think of old Dylan songs he performed in 60s but only released decades later) are released years after their covers.
  • You can not take the first performance as the original because this information is way too hard to find. It's even unclear what a first perfomance could be.

You seem to contradict yourself.

At first you want to take Bob Dylan's (unreleased/not recorded) performances

as an original and then you say you can't take first performances as the original. :-)

I'm confused.

I can't see the problem with your first point (it depends on the definition

of original again), because you want to suggest that most people intuitively

define original version as first performance/recording and I don't have this

impression.

I totally agree with your second point. But then I wonder why you seem to

have first performances in your database, e.g. Todd Duncan performed "Unchained

Melody" in the movie "Unchained". But I don't know if you can call this

a recording as he only performs the song in the movie and there are no infos on a

recording with him as Al Hibbler sings "Unchained Melody" on the soundtrack

that was released. That's why Al Hibbler's version is commonly regarded as

the original.


So what are the conclusions?

Good analysis and conclusions. I'm very pleased to see that you're open-minded to arguments and aware of the problems.

But of course I know that there is no perfect way to do it.

No matter which guidelines you choose I will be happy to

help you with some consistency checks in the future as I have

a small list of problem cases to compare.


Bastien

Manager
Posts: 35819

Bastien @ 2004-04-04 09:20:21 UTC

I'm copying all your arguments about Apache into the Apache thread.


You seem to contradict yourself. At first you want to take Bob Dylan's (unreleased/not recorded) performances

as an original and then you say you can't take first performances as the original. :-)

I'm confused.

I can't see the problem with your first point (it depends on the definition

of original again), because you want to suggest that most people intuitively

define original version as first performance/recording and I don't have this

impression.


I'll focus on Bob Dylan so I think you'll get the picture better. Imagine he writes a song in 60s, and performs it a lot during the 60s, but never releases it. The song gets thus known as a Bob Dylan song. Now assume it gets covered in 70s by X, and that Bob Dylan himself releases it in 80s. Would the Bob Dylan version then be a cover of the X version? Intuitively that doesn't make sense to me. Therefore I think we should take Bob Dylan's version as the first version.


Good analysis and conclusions. I'm very pleased to see that you're open-minded to arguments and aware of the problems.

But of course I know that there is no perfect way to do it.

No matter which guidelines you choose I will be happy to

help you with some consistency checks in the future as I have

a small list of problem cases to compare.


Alright shoot them Smile


... and thanks for the thinking, I guess I see some stuff clearer now than before

gatecrasher

Member
Posts: 14

gatecrasher @ 2004-04-04 15:33:20 UTC

Imagine he writes a song in 60s, and performs it a lot during the 60s, but never releases it. The song gets thus known as a Bob Dylan song. Now assume it gets covered in 70s by X, and that Bob Dylan himself releases it in 80s. Would the Bob Dylan version then be a cover of the X version?


Depends on your definitions :-)

The people and magazines that follow the definition "original = first released"

would indeed say, Bob Dylan's version is a cover. Some would add

"technically" it's a cover but not "morally".

That's what would be in your comment if you also had this rule.


I don't think you doubt that you must bring up a definition and go by that.

The problem for you may be to bring up a defintion that really earns

the name "definition" and doesn't leave up everything to "intuition".

I'd like to help you in this case :-)

Just go by this and you've won (this time :-)):

cover version: a recording of a song that was first recorded or made

popular by somebody else

(from WordNet, e.g. http://cover.version.word.sytes.org/)


I think it's better to continue this discussion some time in the future (if necessary),

when you've written out your definitions and guidelines.


Just one last question to think about:

when you have a recording/famous performance that predates the first release,

but you don't have any info on the story of the song, connections between

artists, composer etc., what would you consider the original?

If you don't have a specific rule (e.g. original = first release), it will be

inevitable that your original may change with every bit of the song's story that

you come across from then on (and it will depend on who you ask as

artists and composers sometimes tend to have different views on things).

Conclusion:

It may be wise, to make your definitions and rules as specific as possible

and handle all additional info (category:"moral and intuitive stuff" :-))

in a comment.


The only thing I missed in your answer is your opinion on

the "Todd Duncan" case. If necessary you can make a new thread

and I can provide some sources on the case.